Web frameworks are churn-y because they are incredibly leaky abstractions covering really awkward impedance mismatches. This means that they are never quite satisfactory - and that just to use one, you need to be capable of building a new one yourself.
Think of a typical web app. Your data exists:
1. As rows in a database, accessed via SQL
2. As model objects on the server, accessed via method calls and attributes
3. As JSON, accessed via many HTTP endpoints with a limited set of verbs (GET/PUT/POST/DELETE)
4. As Javascript objects, accessed via (a different set of) method calls and attributes
5. As HTML tags, accessed via the DOM API
6. As pixels, styled by CSS.
--
Each time you translate from one layer to the next, there's a nasty impedance mismatch. This, in turn, attracts "magic": ORMs (DB<->Object); Angular Resources (REST<->JS Object); templating engines (JS Object<->DOM); etc. Each of these translation layers shares two characteristics:
(A) It is "magic": It abuses the semantics of one layer (eg DB model objects) in an attempt to interface with another (eg SQL).
(B) It's a terribly leaky abstraction.
This means that (a) every translation layer is prone to unintuitive failures, and (b) every advanced user of it needs to know enough to build one themselves. So when the impedance mismatch bites you on the ass, some fraction of users are going to flip the table, swear they could do better, and write their own. Which, of course, can't solve the underlying mismatch, and therefore won't be satisfactory...and so the cycle continues.
Of these nasty transitions, 4/5 are associated with the front end, so the front end gets the rap.
All of this is why, in the end, more and more... I do things "the hard way". If I'm using a dynamic language WTF do I need an ORM for, if I understand enough to write an SQL command, and use a library for that DB that does parameterized queries?
On the front end, I tend to lean towards abstractions that work together... I really like React and the material-ui library's switch to JSS. It's relatively clean, and useful. Even then, it's only a mild syntax adjustment, not a full on abstraction. React is more of an abstraction, but that comes with functional paradigms that aid in testing, and predictive behaviors.
It really depends though. One can always do just JS/HTML/CSS, and there's something to be said for that. There are lighter tools that are similar to JQuery to smooth over a few of the rough edges. There's really an ala cart of available options.
The problem is that people assume that the PFM (pure fucking magic) will solve it all for them. You can use the cleanest or simplest abstractions, and then still write layers of incomprehensible spaghetti in between.
> I do things "the hard way". If I'm using a dynamic language WTF do I need an ORM for
I think it's mostly premature optimization. People think writing DTOs is challenging, so they want an ORM. But since you end up needing DTOs anwyays, removing SQL capabilities from the app means writing SQL in not SQL, and things like joins suddenly become slow and problematic and result in really heavy systems that are harder to change. For the joy of a quicker startup the entire project moves slower.
ORMs have their place, but in the majority of the systems I've seen they were unnecessary, and in broad terms don't provide any particular productivity advantage over using "dumb" SQL-based mapping solutions (a la Dapper [https://github.com/StackExchange/Dapper]), that preserve the power of SQL.
I don't think this is true. Writing these objects isn't difficult, it's tedious and repetitive. That's why people keep trying to automate it!
The problem is that you can't quite automate it smoothly, because SQL doesn't work like objects. You avoid this interface issue by taking the hit for the tedious-and-repetitive stuff directly (and I agree that's often the right choice) - but that doesn't dissolve the problem.
I’m actually quite fond of SQL, but I disagree that ORMs are not a productivity boost. My experience is in Rails and I think ActiveRecord is a pretty clear win for simple queries. That being said, it is pretty common for less experienced developers to not understand what the ORM is actually doing.
On any project of sufficient size there will be fairly advanced reporting functionality that you generally will not be able to do using ORM so you will end up with a mix of ORM and direct SQL. Also ORM forces you to the lowest common denominator for supported RDBMs I generally do not want to be limited to the SQLite features if I am running PostgreSQL
I did qualify my statement with “for simple queries”.
For more complicated queries, a pattern I have become quite fond of is making database views and then using them as the backing table for an ORM model. In Rails, at least, this gives you the best of both worlds.
In my experience, migration scripts that include modifications to output JSON (as ALTER TABLE statements) are always tracked and programmed to be auto-executed on each version update.
About 17 years ago, I wrote a GUI-based code generator that allows me to generate the boilerplate JDBC cruft from SQL statements, with several options for common scenarios. The code it generates is extensible and provides helpers for extending.
To this day, I haven't found anything (including ORMs, Spring support, etc.) easier to use, more flexible, or more sensible.
I recently wrote one for C# / MSSQL. I've been using EF on side projects, but work was concerned with it, so I just wrote a CRUD sproc / Entity / Repository / Service / DTO generator. Connect to a database and select the tables you want to build for, and done. 5 layers of abstraction in under a second. It gets me to about 95% of what I need and I custom build the special circumstance stuff from the generated objects. It creates the "if exists / drop" stuff, exception handling and comments w/ dates in the sprocs that reference the objects in C#, and vice versa, what table was referenced, proper [Key] and [MaxLength] attributes, etc. Puts them in their proper folders for git too. It does exactly what I would have done had I done it manually.
Works great with no ORM abstraction and it was fun to write.
Yeah, very similar--with one primary enhancement: That is, I initially started generating objects-per-table, but found that it was a little limiting for a lot of use-cases I encountered. In particular, it didn't cover joins very well--particularly for queries that fed list views.
So, I expanded into supporting raw SQL SELECT queries that can include joins, which I parse and combine with DB metadata. I then generate the DTOs from there. So, a single DTO can have properties mapped to different tables, which I found much less redundant/limiting than entity-per-table designs.
In addition to the SELECT code, I can use simple checkboxes to also generate INSERT/DELETE/UPDATE/UPSERT code, which map the DTOs back to the underlying tables. It recognizes keys and includes multiple-table writes in a single transaction, etc. In addition to the DTOs and the DAO layer, it can also optionally generate a service interface.
Of the utilities I've written over the years, it is the one that most stands out as having paid me back incalculably.
If your application mostly reads from the DB then don't use a ORM. I think ORMs shine when you need to save/update a domain model to a relational DB.
In the CQRS world, a fairly typical choice when using a relational DB is to use a ORM for the write side and raw SQL for the read side.
I agree, except that I would argue it’s often the easier way. If I had to give a one sentence answer to the original question, it would be, “Front-end [web] development is so unstable because people introduce so much accidental complexity.”
For example, while I don’t disagree with Meredydd that there can be awkward mismatches between the layers he described, I also think several of those layers only exist if you presuppose an object model in your programming languages. Arguments about object-relational mismatch have been made as a criticism of OO for far longer than we’ve been building substantial front-ends for web apps.
If instead you stay closer to the real data, your architecture reduces to the more traditional persistence and presentation layers. Since you’re on the web you have a distributed system so you also need a protocol for the remote communication between those layers. However, there aren’t any inherent mismatches in that combination, any more than there are if you build native applications or distributed systems using something other than web technologies.
I think a lot of the disconnect can go further away as JSON data types (regardless of actual serialization such as BSON etc) and use are more well supported at the database layer. When I did more C#, and Entity Framework came out, I'd add an XML column to most of my tables, as well as a base class wrapper for my own use... that allowed by to write extension properties that wrapped around XML nodes under the covers. So I could extend with extra properties for things that didn't need to be queried on.
It worked pretty well, of course I actually started doing it because getting schema changes at my workplace was a painful endeavor.
I agree that a lot of the disconnect is induced by developers. It's also part of why I'm a pretty big proponent of a JS UI talking to a service written in JS. It allows for a lot less cognitive adjustment. I remember doing HTML/CSS/JS with Flash/Flex, with .Net, T-SQL, and VB6 in one workplace regularly. I swear every time I had to change from one to another, I was typing the wrong way for a good 15-20 minutes... answering questions at times took 2 minutes just to shake my brain out of whatever I was working in.
I started using less stored procedure code and the DB more as dumb storage, and embraced node pretty early on. Even if it is a "lesser" language, there's something to be said for one language to rule them all. (I do like modern JS though.)
Have you ever looked into Clojure/ClojureScript? The Clojure ecosystem seems to favor your approach. They embrace the dynamic nature of this kind of programming, are data-oriented, shun ORMs, and generally have solid principles (in my opinion). I found it was well worth working through the (somewhat steep) onboarding curve.
> The Clojure ecosystem seems to favor your approach. They embrace the dynamic nature of this kind of programming, are data-oriented, shun ORMs, and generally have solid principles (in my opinion).
I've recently started my Clojure journey (< 1 month in!), after stumbling across aphyr's very interesting work, and it is being driven entirely by this line of thought. It's taken me a long time, at least a decade, of moving deeper and deeper in to web development to start to really appreciate this perspective but it _feels_ like The Right Way at this point in my career. I'm hoping to, at the very least, be able to take those lessons from Clojure and apply them to the areas of my professional life.
> WTF do I need an ORM for, if I understand enough to write an SQL command, and use a library for that DB that does parameterized queries?
I am getting into frontend for a hobby project after spending a few years doing ML and applied stats, and I am currently asking myself this question. If your db interaction is simple, a query is almost as easy to write and maintain as an interaction with an ORM, and is significantly more flexible. If it is something more sophisticated, then your ORM quickly becomes more of a hindrance than a help. What am I missing here? Where is the virtue of an ORM beyond not having to use SQL?
In more static languages, you generally need to convert from the DB types to the Native types to your language. This means a lot of code (more than boilerplate ORMs) and prone to a lot of mistakes.
Some tooling that generates it for you helps a lot in some cases. I can see the appeal, but not in a dynamic language environment where there is less disconnect.
This is a great comment and an amazing insight. What's particularly interesting is that people have attempted to collapse (almost?) every stage of that abstraction hierarchy individually, but none of them have been so successful as to take over the world.
If you were writing a desktop application, you would still have at least three of the layers (serialized data on disk, in-memory data, and the rendering of the objects), but without the dramatic impedance mismatch that the web platform introduces everywhere.
Thanks! And funny you should mention that. We're challenging every layer of that heirarchy simultaneously, by building a development environment for the web and making it as integrated as Delphi or VB were on the desktop:
Even on the desktop those three layers involve impedance mismatch and much the same pathologies as
meredydd describes. But I guess three layers of it are better than 6.
It's data. There's a fix format for serializing it. In a lossless way. I don't know what meredydd talks about, there's no mismatch with regards to data.
You can get the same bits in your JS objects as you have in the DB. If not, that means your system is shit.
The problem with frameworks is not the hardship of funneling data up and down the stack.
The problem is that they are optimizing for different things. React optimizes for simplicity of making components. Angular optimizes for providing a full toolkit. And new versions then focus on different things. Server Side Rendering was hot, but now that Google just executes some JS and penalizes large downloads, it's the quest for less bytes on the wire. And tree-shake-ability. And faster time to first paint.
And as browsers and the web changes, so do frameworks. And frameworks try to target, at the same time, both the future, and the very present problems, they try to provide instant gratification, yet try to optimize for the future.
So they usually look half-assed useless pieces of autogenerated-by-MS-Word code all the time. But they work, nevertheless, and power a lot of sites.
> I don't know what meredydd talks about, there's no mismatch with regards to data.
He's talking about different services each having their own preferred way to structure the data. When the layout differs, it cannot simply be a memcpy, and so you get tools to try to ease the tedium of translating one structure's layout into to another. They get the job done most of the time, but run into edge cases that return the developer back to manual tedium. Since developers do not like tedious work, some set out to find a new solution that solves for those edge cases, but they end up leaving many more on the table for the next intrepid developer.
Absolutely. But user/business data? That doesn't matter. When you design the system/stack you pick the right components/tools (right data structures) that can losslessly represent the input/output of the neighboring/adjacent layers. If you want to store 500 byte long fields, then make your DB column 500 byte wide, make sure the backend accepts 500 byte long input, but rejects longer ones, make sure your HTML input has a maxlen=500 (and account for Unicode code point surrogate / multibyte fuckery if applicable)
There's mismatch, of course, but as I've detailed in a sibling comment [0], it's because of difference in purpose and function. A DB is different from a HTML/CSS layout rendering engine, because they have a very (set) of purpose(s), hence different interfaces, and so on. And frameworks are glue between these functions (and the layers as we allocate them to).
> Since developers do not like tedious work, some set out to find a new solution that solves for those edge cases [...]
Yes, perfectly agreed. And since we concentrate on different edge-cases each time, we move from trade-off to trade-off with each new framework, and browsing trend/fad (mobile, tablet, SSR, ultra-tree-shakable-gzip-able, "native" [mobile] compilable, etc).
>You can get the same bits in your JS objects as you have in the DB. If not, that means your system is shit.
I'm not sure what exactly you mean by that. But one thing is absolutely clear. You cannot automatically derive a logical layer from the layer above or below. If you could, there would be no reason to have seperate (logical) layers in the first place.
That's why you get an "impedance mismatch" that has to be bridged by providing some additional information, which often has consequences for performance, debuggabilty and clarity.
Maybe I misunderstand the gist of your comment though.
Maybe I misunderstood the original comment, but the claim was that frameworks are leaky (this I wholeheartedly agree with, and this of course leads to impedance mismatches, after all, pixels on the screen are very different from an SQL DB, but that's why we have the libraries and frameworks, to help us do this translation from one layer to the other, to glue together very different functional components of systems). But then follows it up with talk about data. How JSON and SQL is not a great match. Which is nonsense. You can losslessly represent the same data in both JSON and SQL, you can engineer a perfect system for handling data (you use the same field and column types, lengths, constraints, validation, and so on on both the front- and the backend, and it works), the mismatch is not around data.
The problems are about development trade offs (TypeScript vs JS, small library - few features, complexity - code modularization + chunked lazy loading, optimization - script load time vs development time, and throw in cross browser compatibility; supported features vs complexity - HTTP/2 is nice, and fast, but it's more complex plus you need HTTP1.1 too for old clients, and maybe your API somewhere doesn't support prefetch, or you can't hint your backend to push that to the client, or you can't access the raw request after the framework extracted the request attributes, blablabla), visual communication (current/modern components vs old jQuery sprinkled DOM result in different sites; mobile first, mobile browsers, React Native and Ionic). And these trade offs are different over time. So we get different frameworks over time. And since the change in browsing is very fast, and the effort to start a new framework is small, we get a lot of new unstable frameworks. (And since those frameworks rarely mature really, we get a lot of new ones, because there's not really a "sunk cost" for developers when abandoning the old ones. And it's easy and hip to pick up new skills, and try them out on a new project, etc.)
Also, you can put the "business logic" into one place, and represent it and then push that representation to the client. (GWT, Scala.js, or crude autogenerated forms, point and click website/workflow builders, and so on). Of course, if you want to change the system, then it might be a big pain in the ass to represent something very different than what it was designed for, so these kinds of entombed vertical complexity barriers lead to a metastable state - when you hack something quick on the layer most accessible for you with respect to the task, instead of properly implement it in the whole vertical stack, and these hacks grow and the elegant single source representation of business logic goes out of the window. (Or if you implement everything in one place, you are destined to implement a very powerful - or verbose - DSL to describe the "front end logic" - which should be CSS, and the DB optimization logic
- which should be SQL, and so on.)
That's good, in each cycle, sooner or later you get improvement: webpack over grunt, react over jQuery and npm over vendoring your jQuery plugins.
OP forgot how life looked when your web app project was handcrafted HTML page with manually inserted scripts tags. When your form submission was multi-level backend API in PHP. jQuery plugins with 20+ options published randomly on the internet.
I feel like this is a false dichotomy. The choice doesn't have to be roasting squirrels over an open flame/handcrafting PHP pages vs. shiny futurism/Node+React. I have been getting along just fine with Rails, HTML, and a sprinkling of JS for over 10 years.
There is an immense productivity gain to be found in mastering a set of tools. After a while of dogmatic tooling changes you begin to analyze more critically whether the new shiny thing is going to provide any real value. I would argue that for the huge majority of websites, tried and tested tooling is perfectly fine and definitely more robust and supported.
Every tool you introduce is hours of troubleshooting just waiting to happen.
That stuff works fine if your client-side needs are simple, but if you actually want a single-page application, or just an application with a lot of rich JS functionality, it quickly becomes unwieldy.
That stuff works fine if your client-side needs are simple
To be fair, just about anything works fine if your client-side needs are simple. However, I have reached the opposite conclusion to you: the more customised and complicated and large-scale and long-lived the software becomes, the less value I see in a lot of the popular but ever-changing web technologies and the more I am likely to favour building on the standard foundations with minimal dependencies in between and usually a relatively small but high-value set of libraries.
The benefits of quickly fetching many tiny packages with a package manager or of building on top of all-encompassing frameworks or automation tools are mostly found in two situations, in my experience: getting started quickly (including rapid prototyping exercises) and ongoing development if (and only if) you are staying almost entirely within the bounds of what your chosen technologies already do well.
However, if your requirements start to evolve and diversify in a longer-lasting project, it’s all too easy for those numerous tiny dependencies to become a liability or for that framework or tool you built everything around to become a straitjacket. The relatively short lifetimes of many of these technologies can also become an expensive problem if the community drifts away and the security and compatibility work slows down or stops entirely but your project still depends on them as much as ever.
Monolithic frameworks can also corner you into edge cases where you end up having to write shitty work arounds because their "opinionated" framework didn't have an opinion based in reality.
That was my point, that nothing solves every problem. Monolithic frameworks aren't inherently better than micro libraries, they are just different paths to achieving the same goal. I'm still going to choose micro libraries, because flexibility is more valuable to me than batteries included, which is just as good an argument as batteries included is better than flexibility.
and when ios made the app sides unusable due to a varying notch, having a whole set of unstandard properties you have to handle to manage correctly being into safari on a iphone x
if we had companies following standard decently and a linear, planned grow instead of the organic mess we're into, it'd be far easier to produce building blocks that work in a stable manner over time.
we're better today than in netscape days, but marginally. as complexity increase the cost of this constant churn does too and the saving from better framework are not quite enough to offset the constant fads that come and go.
What is curious is that the tech industry has been very conservative about rethinking these 6 things that you just listed. Why rows in a relational database? Why objects? Why Javascript? And why HTML? We'd surely be in a better place if we got rid of these things and rethought our approach from first principles.
I've written about these issues many times before.
Regarding the problem with objects, I wrote "Object Oriented Programming Is An Expensive Disaster Which Must End":
To answer the question "Why Is Front-End Development So Unstable?" the answer is surely, in part, the fact that we refuse to build technologies that are designed to be great front-end technologies.
There have been decades of attempts at databases with a different model than relational and we have seen what happened.
The reality is that a relational model is very well suited for general purpose databases. For specific needs you can use timeseries databases or key/value store, but at this point I seriously doubt that the Nth attempt of killing the relational model will succeed.
And honestly I prefer an enforced relational schema rather than an ungodly mess of schemaless documents.
HTML is a good compromise between procedural / event-based UI frameworks (such Java Swing or Apache Wicket) and visual UI design tools (such as RAD Studio, Apple's Interface Builder or Adobe Dreamweaver) that allow you to implement the most common patterns fairly easily while often making the design of more custom UIs much more difficult.
There have been attempts at rethinking it though; noSQL was the buzzword of a couple of years ago, and even nowdays there's mature tools like Firebase that allow you to store and retrieve data much more directly than e.g. SQL. The challenge in nosql storage is of course data migrations and whatnot. But yeah, in theory you can just open up a MongoDB instance to your front-end and not have to bother with SQL or much of a back-end.
And your last comment doesn't make much sense tbf; the big frameworks, most notably Angular and React (and its ecosystem) were both designed to be great front-end technologies.
Angular and React both rely on Javascript and HTML, so you can't describe them as "designed to be great front-end technologies". HTML was designed for document exchange, it is a descendent of SGML. Javascript was initially meant to be a light weight scripting language that allowed dynamic elements in HTML. It's gotten better over the years, but it is still far from what you would expect if you were trying to build a great programming environment for the front end. As to the limits of HTML, just consider forms. In the last 20 years, there have been very few new form elements added. Compare the form elements available in HTML in 2018 to what VisualBasic 6 had achieved by 1999, or what Netbeans/Swing offered by 2003.
> HTML was designed for document exchange, it is a descendent of SGML. Javascript was initially meant to be a light weight scripting language that allowed dynamic elements in HTML
I always see people saying this, but why does it matter? Electricity was originally piped into homes for lighting, but we don't need an alternate way to power all the electric devices in our home. Unix was designed for computers that are quite different from the ones we use today. And so on.
I've been saying the same for years. As long as we are working directly with HTML, we will have these impedance problems.
The component/event model a la Swing et. al. is a far more elegant match for modern Web development, which is now essentially the same as building window-based native applications.
OTOH, HTML was designed for static content delivery. Even if a framework must generate some HTML to remain compatible with browsers, there's no reason we have to work or think in HTML as our central interface model.
Give me a canvas, let me lay out (and style) components, then let me respond to events.
HTML has lots of problems, but it's there, sitting on every computer and mobile phone that you can think of. There's not really another viable cross-platform alternative. Your only option would be to render to HTML5 canvas and create some alternative rendering model from the DOM (not unlike Flash). There are tools that exist to do that today (CreateJS, for example), but they're not mainstream. You pay a huge penalty by going against the standard platform. All the interop, tooling, and libraries out in the world work on HTML/JS. Re-inventing that from scratch and coming up with your own hacks/solutions for accessibility, responsive design, style sheets, components, etc, is expensive and unlikely to succeed. I agree that OOP is an expensive disaster...but thankfully JS is flexible enough to code using functional patterns.
I'd guess that a large portion of new frameworks start simply because everyone who spent enough time with the old one gets sick of the bad or non-existent documentation for edge cases, etc.
Rinse and repeat.
Vue is no better, after reading the rest of the comments. In many ways worse.
Even with all the improvement is web technologies in recent past, the browser is still not close to native widget.
And finally the users themselves. No matter how neat you managed to be in the underlying layers, the UI is messy. It can change fast, meaning either massive cost rebuilding an entire application, or breaking those neat layer.
Users also wants everything connected to everything. It does not matter if those connections are explicitely done via ugly spagetthi code or implicitely through clever abstractions, they effectively exist and that's how the requirements, user experience feedback, bugs and testing will be based on.
Some good points... but I think you're too pessimistic. Some tiny subset of those fraction of users you mention manage to do something different / better. React is one example. GraphQL is another.
This is a really well-written and well-thought out piece. The author touches on a number of points but never gets polemical. This piece resonated with me particularly well:
> Be wary of self-promotion
> Over the last few years I’ve seen much more aggressive self-marketing in the JavaScript world, possibly due to the rise of paid online training materials and the employment/consulting advantage of being a Github ‘celebrity’. I’ve no problem with people being incentivised for good content, but increasingly I feel I see dishonest tactics: self-citation; invented, proprietary terminology (so searching takes you back to the author’s materials), and name-squatting (e.g. ‘Standard.js’)
This is becoming a big problem in JS world, particularly over the last 3 to 4 years. Like the author, I have no problem with people taking credit and being respected, but we as a community are becoming much too centered around people/personality than technology. I've been guilty of this in the past too (love you Pete Hunt ;-) but it's not healthy IMHO.
I'm finding the growing celebrity culture of programming to be troubling. I find that programming is becoming less and less about code or architecture and more about evangelism and marketing, but I'm not sure what changed in the incentive structure for this to happen, and I would like to find out.
Aside: This has made me reconsider being a career programmer.
Is this actually a new phenomena? I remember when I started programming c. 2000, the big-name programming gurus were folks like Kent Beck, Ward Cunningham, Uncle Bob, Ron Jeffries, Dave Thomas, Joel Spolsky, Jeff Atwood, Martin Fowler, et al. While some of them mostly knew their stuff, some were outright hacks (Peter Norvig's vs. Ron Jeffries' attempts to create a sudoku solver is one of my favorite examples [1]). Most of their names wouldn't be recognized by someone just getting into programming now.
As I've gotten older, I've learned that there's often an inverse correlation between how good a programmer is and how likely you are to have heard of them, because the folks who get good at programming are actually programming and not writing blog articles about programming. Most of the latter are really content marketing, trying to drum up business for consultancies, and so you should treat them as advertisement rather than advice.
I think this is a very good point. But it's also important to remember that not all programmers/tinkerers that are creating blog posts/videos are looking for fame - some of them are just proud of something they built, and that's fine.
Inspiration is a great enabler, and taking a few shortcuts to make something work may be just the thing that allows someone to actually finish one of their projects. The problem is that if you lack the knowledge to see that something is a hack - you learn things the wrong way. Still, without this inspiration - some people wouldn't start to learning at all.
People that are actually doing stuff have no time to talk about it, and people that are talking about it are wasting time by not doing it.
So beware of gurus. No one gets that popular without spending a large proportion of their time promoting. In the decade I’ve been programming I’ve never had time to write even one damn article.
You're at the right place to work out what started it.
The aggressive personal branding is clearly spillover of startup hype/marketing/branding styles and techniques. If your whole career has been at startups then that is just called "marketing" to you.
Also: it works. "No one ever got fired for going with IBM" is now applied down to the micro-level. Few choose a project because of a code/architecture review, most choose one because it has a good reputation. The industry has fantastic expertise at manufacturing reputations by now.
It's an extremely large and diverse field. There are plenty of us who aren't celebrities and don't care about celebrities (except to the extent that it is caused by deep expertise). Just like you ignore reality TV celebrities, ignore these guys. It's just noise. Turn off Twitter and read "Growing Object-Oriented Software, Guided by Tests" instead.
No, it was a quip, basically intended to mean "forget about minute to minute popularity and read a book that is well respected by practitioners in the field and has stood up to the test of time". And I didn't mention the authors, it's about content, not personalities.
Man, I couldn't get through it! It seemed like its own kind of hype-y proselytizing and I didn't find its examples spoke to the kind of issues I actually have. Maybe I didn't give it enough of a chance, but I don't know, I had been hearing for years that it's so great and I was really excited to start reading it, and it just didn't speak to me at all. Anyway, YMMV I guess!
It's not a religious text or anything (I suspect that the proselytising tone comes from the fact that at the time it was written, it's recommendations were still a bit controversial, especially in the kind of organisations likely to have these kinds of codebases). If it's not useful to you, then that's that. The point is, other books written ten years ago by people with deep expertise (rather than a hypey blog post by someone who just discovered some principle five minutes ago) probably will be.
This is a different topic than my comment that you replied to but the "ten years ago" thing stuck out to me: I think books (and blog posts!) written very recently by people with deep expertise are likely useful as well. Figuring out which ones those are is the tough part.
Ha! I'm not sure how serious this comment is in terms of how you choose to spend your time at work, but a person with that attitude isn't going to last long working for me. I expect them to deliver.
I'm not a tyrant, and I avoid setting hard deadlines as much as reasonably possible, but I do expect concrete evidence of output that delivers customer and/or business value.
Still, in many organisations I think you're right. So much energy is expended on marketing to/currying favour with managers that not much is actually delivered. It's how you end up with 100+ people taking more than 6 months to deliver two pages that allow customers to register for a service (names omitted to protect the guilty).
Someone with that attitude isn't going to last long working for you anyway: as soon as they're hired, they'll put your company on their resume, blog about everything they've learned, trump up their first project, and start looking for their next gig, usually getting a big pay increase in the process.
The minimum ability someone needs to have for me to consider them a programmer is the being able to write a parser using yacc or something equivalent. Anything a person with that ability does is programming because they know enough to realise when switching to putting legos together it more efficient.
If not programmers, what are these people who spend their days programming, but who may not necessarily have ever written a parser using yacc because none of the jobs they've ever had have required that?
I am a software engineer that loves to help others. Nothing to do with self-marketing, at least not intentionally, but a welcomed skill in the team. So the question makes sense to me.
I have tried it a couple of times (I do have a couple of posts on my github pages). I just don't have the time / motivation combination to do it on a regular basis.
I find it quite time consuming writing in a way that makes it clear to others what I am doing. I'll come over to your computer and explain the same concepts in ten minutes that would take me an hour or two to write up and format in a manner that I would want to display publicly.
(I find answering questions on Stack Overflow a lot less time consuming to get a useful point across).
And because you don't have time / motivation everyone that has is a self marketer?
I agree it's silly to expect every developer to have a blog. I strongly disagree with the only motivation for it being self marketing, or it being a somehow unreasonable question if you have one.
I agree there are talkers and there are do-ers, and very few who are good at both. They do exist but they are rare. This is true for many professions, not just programmers.
Sure. Luckily the blogs written by the two categories are fairly easy to tell apart. Even checking if the domain is medium.com or something else is a good first estimate, but to be sure you should look at the content.
I'll come over to your computer and explain the same concepts in ten minutes that would take me an hour or two to write up and format in a manner that I would want to display publicly.
That’s right, but multiplied by the number of people who read it…?
If you ever used any F/OSS software, or made use of a blog post when solving some problem, you owe to the community a blog or something similar. That can be answering SE questions etc. occasionally or an anonymous blog, not contributing knowledge to the community that saves you many years of work every day is outright selfish.
Furhtermore, I don't see what makes self marketing a bad thing given it is of the honest kind. The CV you send to them in order to get the interview is self marketing too, and so are thw clean clothes you wore to that occasion.
I cant count how many times a random blog post has saved me blinking weeks. You are being very rude with that statement.
If you want to write a blog by all means do so. I have written a handful of blog posts but I don't have the time / motivation to do that in my spare time on a regular basis.
I think its silly to assume that all software engineers want to and are good at writing blog posts. Could you explain what is rude about that?
> you owe to the community a blog or something similar
Do I?
Maybe that should be put in the terms and conditions if that is the case.
I have a decent score on Stack Overflow from answering questions (~379k people reached apparently), is that good enough for you?
I'm not going to answer the second part, but for the rudeness, you are being rude because you reduce blogging about one's experiences w/ programming and related stuff to self marketing.
Nobody owes you anything. He gets to have whatever perspective he wants and it's not rude because you disagree with it. You need to take a step back and pretend for a moment that your perspective isn't the only one.
No, we don't. The FOSS community operates on the idea that people that want to contribute to it (financially, programming, etc) can do so. Nothing about open source obligates that behavior, outside of certain licenses. Nobody "owes" FOSS anything either, that's up to them.
It is really shocking people can think this way when also making use of hundreds of people's free labour, and when the very existence of most of programming careers are a direct product of OSS innovation. Hopefuly this attitude is not the widespread one.
Those people are volunteering their time. I'm one of them. But that doesn't mean owed something as a result, that's absurd. I made OSS because I wanted to.
> Any why don't you want to answer the second part?
Because I think the answer is obvious. BTW I did not see the Stack Owerflow part, that is the exact kind of compensation for tge greater community I was talking about. Thank you for that.
For the other paragraph, well, that's how I read you comment in its context, and I maintain my interpretation. Maybe I'm misunderstanding.
I answer Stack Overflow out of mainly self interest. I don't have a portfolio of work to show people at job interviews - my work is mainly in house. I realized it isn't that hard to get a decent score on SO if you know something well.
Also because it annoys me seeing things done poorly having worked in enough maintenance programming jobs and worked with crap that could have been done so much better (this is probably a bigger motivation with certain topics).
I don't do it because I feel I owe anyone anything.
It does feel good to contribute somewhat, but that is far from my primary motivation.
Yikes! As someone who can't stand being the center of attention and also can't stand talking about themselves (yes I see the irony here), this is terrifying.
I think that this was inevitable given the rise of social media and the Internet. It’s still entirely possible to be a great developer without having a Twitter or Facebook account, but it’s tempting to feel like I’m missing out on information from not subscribing to the latest fight on Twitter.
Once the market is large enough, you're going to have some people at the top who excel at marketing. Look at almost any large market, it's not the "best" products that win, it's the best marketers.
Similar boat. I’ve been presenting passion projects to user groups / unconferences for a decade. Now, I wasn’t aware at the time but around 2013 seems like the turning point. I was getting tingles then. I never wanted to be a celebrity but it seems my aptitude has sent the wrong signals to some. I refuse to tow company lines, but acknowledge funding of open work or course. Starting to explore other outlets for sohpistic endeavors eg. live streaming (shameless plug, building an open source platform, search: ispooge live). People subscribe and keep coming back and chatting too, didn’t expect that but it’s pretty cool. I feel it can lead to a good career in some way, involving good natural collaboration. Maybe even money ;).
It's gone mainstream, and just like the internet going mainstream, it isn't what it used to be. These days I only enjoy the programming I do on the weekend.
It's a problem for other communities, like Python. I read a blog post recently from a well known python dev where he talks about explicitly cultivating "fame" around his open source tools. My experience, as one who has tried to promote my own tools, is that literally nothing good will come of it for the community as a whole. It's a bad practice in my opinion.
What incentives would a someone have to invest a considerable amount of time in releasing and maintaining an open source project if they can't even use that to promote themselves?
Maybe they have a genuine interest in the problem they're addressing, and in solving it. I work on some open source projects that gain little attention from others but have great utility to me, personally.
It's a nice sentiment but that's how you end up with abandoned, low quality projects that no one uses. Or even worse - abandoned, low quality projects that half the Internet relies on.
> It's a nice sentiment but that's how you end up with abandoned, low quality projects that no one uses.
How is it a bad thing that a piece of software that no one uses is abandoned by its developers? Or should i put it like this: how is it bad that no one uses a piece of software that has been abandoned by its developers? Whichever way you prefer to put it.
> Or even worse - abandoned, low quality projects that half the Internet relies on.
It's not my responsibility to make sure that my software is used for things that it is useful for. In fact, I typically use a license where I deny responsibility for fitness for a particular purpose, like MIT or GPL, exactly because it isn't and shouldn't be my problem.
I think this should be accounted for when deciding whether to use open source software in your project. Of course, with Javascript projects I often end up looking at shitty dependencies because some popular library depends on some slightly less popular library which depends on a brain fart someone put on github 10 years ago and never touched since despite stale, open issues. This is a problem with the community's attitude towards dependencies, not with what software I make available for others to use.
Besides, a low quality project that half of the internet relies on doesn't really need any additional promotion or advertisement, does it?
Calling out Vue.js as a possible solution is hilarious. Selling a trending technology as a silver bullet is exactly the problem he complains about in the article.
Comparing the development methodology of Facebook - the company behind React - and this mindset is illuminating. Facebook famously rewrote PHP and added extensions like XHP rather than start their codebase from scratch. React follows a similar philosphy. It can be incrementally introduced into a codebase without radical rewrites.
These lessons are largely ignored in the open source community, where everyone will gladly tell you how your tech is all wrong. This is actually really easy to avoid. Just focus on solving the problems you have and only adopt technology that is relevant to you.
It's also much easier to adopt if you have a lot of people who aren't familiar with React or even SPAs in general. Templates feel more familiar to them than JSX. And if you're building something complex enough to need it, Vuex is a hell of a lot simpler to learn than Redux. In general, you don't have to know anything about functional programming to use Vue and it's core libraries, and that can be pretty nice if your team isn't familiar with it.
I will say that Vue feels like an evolution like a cousin to Angular 1, without trying to do so much in the box. I still prefer React, but can see the appeal. Similarly, I can see the appeal of WebComponents.
I think the appeal of Vue is that it's "Angular done correct" and "Angular without the bloat you never use". The core concept of Angular was never bad, it's just that it had tons of pitfalls and bugs and it wanted to be everything at the same time, not just the view-layer.
At some point, when modules get "micro" enough, the effort of managing them and learning their usage outweighs the effort of implementing the thing yourself. Most JavaScript projects seem to walk that line quite closely.
I'm a React developer at work, but I recently gave Vue a try at home. This is one of the things that most stood out to me about it. React makes a selling point out of the fact that it's "just rendering". It doesn't try to do everything for you, it tries to do one thing really well. That sounds great, but it leaves out a factor. Having your primary framework be "simple" and "lean" doesn't translate to your project being simple and lean, because your project still needs all that other stuff. You just now have to get it from somewhere else, which often adds more complexity than it's worth.
When I downloaded Vue all I had to do was add a script tag to my HTML document and start writing a modern reactive UI. No other modules or frameworks. No build system. No special JS-targeted languages. Vue does everything you need, and it does it all pretty well, and sometimes that's more valuable.
I learned React at home, then learned Vue. I fell in love with Vue because i clicked on the getting started, threw a bit of code in a html file, and ran it, and it worked, then I played around and changed things, the moment i ran into something like "oh how do i do this?" the documentation is second to none for client side frameworks, I found what I was looking for quickly.
Then I started to find other stuff like vetur, the vue cli, and I felt like I was just building stuff, and it worked, and I wasn't pulling in other plugins just to get stuff done.
The barrier to entry was much lower IMO.
Just wish I could switch from Angular X to Vue at work :(
Seriously, I feel the same way. I tried Angular1 5 years (?) ago and had to do ember for a gig. They were way too complicated, bloated, and terrible.
Vue just works. One of the book suggest that if you like React JSX you can do it too in VueJS. I've never done React but the thought of mixing javascript with template sounds horrifying. VueJS default template system make sense and it is self contain to a file or component.
The view and the presentational code that builds it are one, they belong to the same presentational layer and always have been. To artificially separate them was a major mistake of the old paradigm. Older approaches, like Knockout, Angular, Vue, etc., have to support presentational code as well, and obviously it is mixed into your html, but in a way that is severely limited and causes complications. Think of it, to be able to do your v-if, v-for, v-else, etc. an entire templating language had to be built, your view isn't part of the scope that that can access the data-model, so dependency injection had to be utilized, components in that model aren't components but directives that have to be registered to re-use one another. The complexity is actually mind-boggling.
These are the exact reasons why React came about - it removes that complexity entirely. It is tiring having to learn frameworks and arbitrary extensions, googling for framework-specific solutions for every trouble that surfaces.
My small goto has been preact with director plus a small custom redux implementation (which contains thunk, action arrays, event emitter, etc). At less than 5KB zipped, it's amazing how well it performs.
If I remember correctly, first versions of React were similar in that respect: you could just add a script tag to your page and you that’s it (save for the mount-point thing).
2) Since things like decent state management don't come in the box, you're almost forced to start including other modules, which pushes you back towards a build system.
The script tag approach is nice and lowers the threshold for the curious, but looking back, I'd recommend starting with vue-cli right away, because as your app grows you'll run into limitations of that setup, esp. with more advanced babel features, runtime transforms, ... The warning about using the cli in the vue docs is imo a bit exaggerated, you get some nice prompts and an equally smooth experience :)
Sorry but I find font-end development pretty stable.
Having to learn a new library every 3-4 years (because of a new job, generally) is it really a big deal?
In my career I also had to learn Java, PHP, C#, Node.js, bash, batch, Python… Build apps with Spring, CakePHP, Symfony 1 & 2, ASP.NET MVC 3 & 4… Query database with Hibernate, Entity Framework, Linq SQL… Handle dependencies with Maven, Nugget… Store data in Postgres, MySQL, SQL Server, Mongo, Redis, Redshift… Host on OVH, Digital Ocean, Heroku, AWS… Learn REST and GraphQL, SVN and Git, Linux and Windows Server, Vagrant and Docker… And according to HN, what's really hype right now is Rust, Go, FoundationDB, Kubernetes, and so on…
Any idea why back-end development is so unstable? :)
> Having to learn a new library every 3-4 years (because of a new job, generally) is it really a big deal?
_If_ frontend development is all you do, then no, it isn't.
One of the good things about the internet is that it's possible for a lone wolf to build and ship something amazing. One person can understand enough of the full stack, from HTML/CSS through JS to a serverside language and basic server administration, to build something great, unfunded.
The constant churn in frontend development makes that increasingly difficult. If you find a useful library for your backend code, great, just 'gem install' it (or pip or whatever). Useful JS library for your frontend code? You've got to learn the packaging technology du jour just to be able to install it. In two years' time this'll be different, and you'll have to get to grips with another complex web of tiny little components that fit together in ways you don't fully understand. Whereas your backend (for me, Ruby+Postgres) is still essentially the same it's always been, and just as great.
Indie full-stack developers can no longer keep up with frontend development, and I think that's a great loss.
> Indie full-stack developers can no longer keep up with frontend development, and I think that's a great loss.
Who exactly? Did you ever hear a front-end developer say that or are you building a strawman?
There's nothing stopping anyone from building a simple jQuery powered website. There's nothing stopping anyone from building an app with Angular 1 either. Hell, you can even download .zip files on GitHub, unarchive them in a folder and start using them in your projects. Complex build system are only a way to streamline development when you get to a certain level of projects output.
I was a full-stack developer 5 years ago. I am not any longer due to the chaos in frontend development technologies, best practices, etc. It was too much for me to keep up with, so I opted out and only do backend stuff now.
I could not quite articulate what I was thinking, but you have captured my thoughts very well.
I used to be a 'full-stack' developer - back when 'installing' JS libraries meant dropping the file(s) into your `static/js/` directory and running with it. I find frontend technology so chaotic (not complicated, just chaotic) that combined with the rapid improvements of browser/css/html feature support I was left in the dust and relegated to be a backend engineer.
I think you really nailed it with the packaging concerns. It doesn't help that every little framework (and mini-framework) seems to think that the problem can be fixed by just shipping their own little CLI wrapper.
Not having a decent build and package management system is what makes the frontend ecosystem so hard, IMO. I'm hoping this finally changes in the next couple of years.
I've recently gone back to do some Java development after idk, five years of front-end and mobile, and the tooling and frameworks are still the same. A bit more polished maybe, but, Java, Spring, Gradle, bit of Hibernate (urgh), bit of Jackson magic, not much has changed since then, so it was pretty easy to get back up to speed in that area.
Did you read the article? Here's a snippet from the beginning:
> Do front end technologies actually change that quickly?
> In the sense of major view technologies, probably not... Consider this list of the highest ‘starred’ JavaScript front-end technologies on Github:
+------------------------------------------------------------+
| Library | Stars | Released | Age |
|------------------------------------------------------------+
| React | 96986 | March 2015 | 3 years |
| Vue | 95727 | October 2015 | 2.5 years |
| Angular (1) | 58531 | October 2010 | 7.5 years |
| jQuery | 49061 | August 2006 | 11 years |
| Angular (2+) | 36665 | December 2015 | 2.5 years |
| Backbone | 27194 | October 2010 | 7.5 years |
| Polymer | 19668 | May 2015 | 3 years |
| Ember | 19003 | December 2011 | 6.5 years |
| Aurelia | 10506 | June 2016 | 2 years |
| Knockout | 8894 | July 2010 | 8 years |
+------------------------------------------------------------+
> 2.5 years for the youngest isn’t that old in the scheme of things - it’s less than half the support lifespan of your typical desktop OS, for example - but it’s still a ways off our caricature. So what is causing this perception of rapid, even unsustainable change?
Small correction on the chart, React has actually been open sourced for ~5 years, React Native is ~3 years old.
As far as how fast frontend technologies are changing, in my ~10 year career, I've only had to use jQuery (~2008-present), Backbone (~2012-present) and React (2015-present) professionally. These libraries weren't swapped out on a whim, and I still use all 3 in different projects on a weekly basis. While I've dabbled with most of the others, I haven't selected them or encountered them professionally.
To some extent, I think library churn can be attributed to keeping things interesting when the problems you're solving are less interesting.
Most frontend engineers could probably benefit from rolling their own SPA library (not for production use, but to learn how the different pieces come together). There's probably room for a few more "From Scratch" tutorials in this space akin to Destroy All Software's series of screencasts. Working through something like this would make it clear that the differences between the libraries above are smaller than it might otherwise seem.
I never understood this argument. You are comparing frameworks. All of those are nothing but fancy ways to play with the DOM. There's nobody telling anyone to jump from one to the other.
Do you migrate your PHP or .NET projects every time a new framework comes out? No. You shouldn't with front-end either.
It's as if you printed a table that goes "CakePHP; Yii; CodeIgniter; Laravel; Symfony; Phalcon; Zend." and then complained about the PHP ecosystem.
I think the title is referring to the ecosystem and not the technology.
• How many frameworks for Java, Python, Bash, etc did you have to learn?
• Did you switch from Redis to X because the former was deprecated N months later?
• Are OVH, Heroku, AWS updates getting in the way of your development process?
Learning a technology by itself like Node.js is not a problem, the problem is that front-end development nowadays requires so many small and fragile tools just to build a web application compared to the number of tools that you would need in another language like C# where the most recognized libraries are several times more stable and change with much less frequency. Mentioning things like database engines, cloud platforms and version control tools is out of the scope of the discussion because they are encapsulated tools, and at least the ones that you mentioned (before you updated your comment) have a very low release rate, and to my opinion are very-very stable in comparison with the JavaScript ecosystem.
> Having to learn a new library every 3-4 years (usually when starting a new job) is it really a big deal?
Yes, it is.
> In my career I also had to learn …
Not for the same project and/or development cycle, but surprise me if I that's the case.
> And according to HN, what's really hype right now is Rust, Go, FoundationDB, Kubernetes, and so on
People stick to them and that's it, I don't know any serious company changing their tech stack every N months just because they are trendy on HN. Several years later they may switch to a new tech stack because the project requirements have changed but not because the author of one of the thousand libraries that you are importing decided to carelessly modify a feature that another dozen libraries use, consequently breaking not just your code but also many others. Or, in another scenario, your package manager doesn't betray you by destroying all your file system during a careless upgrade [1].
> Learn REST and GraphQL, SVN and Git, Linux and Windows Server, Vagrant and Docker
You are missing the point entirely… We are talking about how "unstable" front-end development is, not how many tools you have to learn. If I could learn 10-20 tools for front-end development and be certain that they will not drastically change in a year or so, or at least get deprecated, then — and only then — I would not have to worry so much about available updates via the main and (unfortunately) only package manager that the JavaScript community have.
I've only seen one issue in React that affected me.. when they started requiring PascalCase for component names with JSX. Everything else has been really nice. I glommed onto Redux and thunks pretty early on, and have been wrapping around fetch about as long. I used Bootstrap for a long time, and progressed to material-ui with very little disconnect. Those have been my personal preferences for several years, and they haven't changed THAT much in terms of usage.
Now that's nothing like I've seen with say Angular, Angular2-6 (including the long, painful Angular 2 beta cycle). I've also seen tooling evolve, and some of that painful... Webpack and Babel have changed a lot. create-react-app doesn't solve it all, but helps abstract some of it.
Node has had growing pains and is finally fairly stable for those developing/adapting native code to run in node.
I haven't really had to re-learn anything for a project as it's progressed any more than I did 15 years ago on the backend.
The JSX transform has, as far as I know, _always_ required PascalCase for component names. That's how it distinguishes between "this tag name is actually a variable pointing to a component" and "this tag name should be turned into a string".
Nope.. early on, you could use any local variable... there was a whitelist of html elements... they did a cleanup that required the PascalCase and defaulted to html element render for anything else. It was around 0.10 or so, iirc.
If all you do is developing applications that don't need to be maintained for 10 years or longer (as many business applications do), then you are fine.
Otherwise you are horribly screwed if you try to buy in into the current HTML-based fronted development.
> 2.5 years for the youngest isn’t that old in the scheme of things - it’s less than half the support lifespan of your typical desktop OS, for example - but it’s still a ways off our caricature.
Let's compare it with some server-side technologies:
Python/Django: 2005
Python/Flask: 2010
Nginx: 2004
Perl/Catalyst: 2006
Maybe I'm biased towards older software, but the difference to the popular frontend stuff is striking.
The only thing in the frontend with a comparable history (that I can think of) is jQuery -- and I still use it today, but it seems many developers frown upon it these days. And it's a library, not a framework.
The difference is that the backend model has been around for 20 years if not more. You're locating lots of old-ish and stable libraries because that's when that stuff was made.
Maybe you don't remember the days of Zope, Subway, Turbogears, Paste, Pylons, Werkzeug etc just within Python, but that happened. React is a relatively young model for the web. Wait 10 years and it'll look the same.
I personally don't feel it's that unstable. Pick a library, stick with it, stop jumping to the next framework and the churn stops. It's part of being a mature dev. Things always look worse to outsiders than they are.
The front-end world is rediscovering lessons learned in the days of desktop thick clients from the 90s, if not earlier. I rage far too often that I have to jump through all these hoops to slam out some UI that I could have done in high school with VB6.
It’s a real stretch to compare a reactive library like vue with vb6. Building uis in vue is a pleasure and you almost entirely avoid ui state bugs since you aren’t directly manipulating the ui controls (dom)
> VB6 apps I've ever seen didn't even handle, or enable a window resize properly.
I think the complaint is that GUI builders from the 90s could have been adopted and improved (responsive for one) on the web in the last 20 years. Flash had a GUI builder, animation timeline, etc. ten years ago.
Well, then take Qt Designer if you must - automatic layout management to gracefully handle window resize.
The amount of crap you have to go through to make that work properly (and even then it won't work right in all cases) using the common web frameworks is just astounding ...
I'm referring to CSS media queries in HTML/CSS, which don't have a nearly as clean equivalent for applications. Even then, I've done better with React. If I had to build native apps today, I'd look for a React Native adapter for whatever platform or reach for electron first.
> Pick a library, stick with it, stop jumping to the next framework and the churn stops.
I wish it were that easy. At work, we've picked angular 1, and now there's angular 5, which is a totally different framework (and the migration path only works well if you did everything juuuust right), and no clear policy on how long angular 1 will be maintained.
So either we churn, or we have no idea how long there will be any support.
This is the risk in the authors advice to just pick a monolithic framework.
We’ve been building a React framework. We’ve been careful about how we choose components to use, so there has been very little churn. And because we control the implementation we’ve been able to build things we could never have done with a library like Angular.
They made a fair number of mistakes with Angular.js (1), so the rewrite was a sensible, if painful change for early adopters. Since version 2, the model has remained relatively stable with changes being improvements rather than completely breaking.
These are some very fair points. React also really does seem to be becoming a "standard" more so than anything I've seen in years. Time will tell of course, but I think you're on to something.
> but it seems many developers frown upon it these days
This is sadly very true. jQuery is almost used as an insult these days. Having code be called "jQuery Spaghetti" is about the worst epithet that can be hurled at a javascript codebase these days.
jQuery is great for small snippets of code, but encourages a quick and dirty style that collapses for larger applications. Most jQuery applications couple their business logic to the mechanics of working with the DOM.
It was also intended to smooth over browser inconsistencies and gaps in APIs. This is less relevant these days now that browsers are more mature.
> jQuery is great for small snippets of code, but encourages a quick and dirty style that collapses for larger applications.
I keep hearing (well, reading on HN) criticisms of this sort about jQuery and they are absolutely not supported by the years of experience I have using jQuery for front-ends that have in many cases been quite complex. I use it because it greatly simplifies DOM manipulation and event handling. (e.g., no need to remove event handlers for a subtree removed by the .remove() method.)
In my experience the app tends to evolve around jQuery selectors and event handlers, rather than having a well defined structure. This is not a fault of jQuery, as it never intended to solve those problems. Like you said, it is purely meant to solve the DOM issue.
Bad jQuery apps follow the pattern: When “.x” is clicked, show “.x .y”, add className “xyz” to “.z”, and fire off an XHR to “/api”.
This defines what the developer wants to happen, but is brittle and hard to test. Frameworks would typically break this into actions or methods that modify state and a UI that updates when the state changes, so that the parts can be effectively unit tested and the UI can be changed without touching the rest of the logic.
This is possible to do with jQuery too, but the library doesn’t do anything to help you. This is one of the primary complaints levelled at React, too.
I’ve seen 10,000 line single jquery files handed off to other teams who were asked to merge them into their 4,000 loc customized quasi consumer of the 10,000 loc file. This was at a well respected and iconic tech company with very smart engineers.
Even good people and teams can suck at architecture.
I’ve taken those codebases and refactored them to be testable so I know it’s possible. But why not just use something sane like react to begin with, considering it’s 2018 now. At that point, backbone was only a year old.
> Bad jQuery apps follow the pattern: When “.x” is clicked, show “.x .y”, add className “xyz” to “.z”, and fire off an XHR to “/api”.
This defines what the developer wants to happen,
Yes, it does. All in one place, in about 10 easy to understand, easy to debug lines of code.
>but is brittle and hard to test.
I don't see how it's brittle, or hard to test. You click the button, and verify that it does what it's supposed to. Having done it more times than I can possibly count, and ending up with robust code, shipped on time, I can tell you that it's not hard, if you know how to do it.
> Frameworks would typically break this into actions or methods that modify state and a UI that updates when the state changes,
Which, IMO, makes something simple to understand and dead simple to debug into something painfully, ridiculously complex. And people complain about "jQuery spaghetti code".
>so that the parts can be effectively unit tested and the UI can be changed without touching the rest of the logic.
You cannot effectively unit test UI code, if by effective, you mean that it can replace manual testing. It can't replace manual testing. Someone will have to click that button under all of the likely scenarios and verify that it works. I'm sure it's not to you, but to me, writing unit tests for UI code would be a massive waste of time.
If someone thinks they can be more productive using JS frameworks, and I don't have any personal financial stake in their productivity, then they should use them. But I don't see web development using a few simple tools like jQuery as difficult or mysterious or time-consuming.
Last year I interviewed a few recent boot camp grads and they all sent me a link to their copy of the same React-based project. When I asked them questions about how things worked, in generic terms, for example, "what do you think makes this picture slide down slightly and expand in size when I mouse over it", they had no clue. I guess it was just a React component they dropped in, following the steps in the tutorial. I'm not faulting those people -- they paid a lot of money and did exactly what they were told they needed to do to land a sweet high-paying programming job. But they're simply of no use to me. I need the one who looks at it and knows right away that it would take about 4 lines of CSS to accomplish, even if they had to consult a CSS reference to find out exactly which properties/values to use.
Scale that to a VERY large application. There are lots of applications where class names and even hierarchies get re-used by other teams working on another portion of an application. In practice, CSS blows up, other portions of the app stop working correctly. Yes, you CAN use discipline in order to create good applications with jquery and others.
In the end, I'll take my single state tree (source of truth) and one-way rendering path, and find it much more sane, without weird state and interaction bugs that show up.
>In practice, CSS blows up, other portions of the app stop working correctly.
Perhaps in your practice, tracker1. But not in mine.
That's what everyone seems to be missing here. This is all a bunch of hand waving to me, because I have a lot of experience shipping a lot of robust, maintainable code, and none of it has been true for me.
If you need all of this stuff to ship robust, maintainable code, then you would be foolish not to use it. But I would be foolish to use some complex framework that I do not need to ship robust, maintainable code.
I think the facade is starting to crack with many of these JS frameworks. More and more people are writing articles like the OP and saying that this particular emperor has no clothes.
> because I have a lot of experience shipping a lot of robust, maintainable code, and none of it has been true for me
Ave you ever had to work on an application that's more than 5 years old, with an active dev team of 30+ (just the web developers) that have had over 200 hands in the pie including contractors then? I have, and it was a nightmare. Frameworks and modern tools help to take care in these situations.
The application above was around 2007-2008 IIRC... different parts made by different teams, cobbled together. Layers of backend cruft as well. I did help start a new project, that had much more consistent/clean structure. But when you have too many hands in a pie, and no automated testing in place, you wind up with that eventually.
I've been developing web based applications since 1996. I've lived through the eras without the browsers and abilities we have today... the growth of the DOM and the JS language from a few interfaces for forms, to being able to do so very much.
You don't have to use anything to ship your code... but you have to do something to get a few dozen devs working on something cohesive.
>Ave you ever had to work on an application that's more than 5 years old, with an active dev team of 30+ (just the web developers) that have had over 200 hands in the pie including contractors then? I have, and it was a nightmare.
I was a rank and file developer on a similar shit show years ago. The main difference was that it was a greenfield project at a startup. The people calling the shots were dead set on using the latest fads -- at the time it was the Rational Unified Process, with all the attendant documentation, and EJBs. I knew that Entity Beans were a monumentally stupid idea when I first read the O'Reilly book about them. And I said so, to anyone who would listen, to no avail. I think the team reached 40 developers at its peak, of which maybe half were totally incompetent. (By my standards.) The schedule slipped rapidly, we were put on mandatory 6 day work weeks, I bailed out, easily finding another job, and eventually the whole project cratered.
I'm not going to argue with your experience, tracker1, but all of my experience tells me that the ability of the developers is the best predictor of the outcome of a software development project. And our profession, unfortunately, is awash with incompetent people. For example, people who have worked for over 5 years as a Java developer, who are unable to write a Hello World program in Java from scratch in a plain text editor, and compile it and run it from the command line.
I came into my example above 5 years in... it was "Enterprise" .Net and not Java, but a lot of the same techniques at play... I hated it all. I'll take today's JS/NPM ecosystem over those days. I rally against ORM, and DI/IOC frameworks to the end, they aren't needed in JS.
That said, I'm not saying no frameworks/libraries/tools, and am okay minimizing. But I'd rather use Vue, React, Redux and other libraries/patterns than not in most cases, and find them better overall than ad-hoc jQuery. And don't get me wrong, I've written a lot of ad-hoc and organized code without modern tooling. I'll take today's module systems, builders and bundlers.
As to Java hello world, frankly every time I've had to touch a Java project, it takes 2-3 days to get a build environment running on a local dev machine... it's nightmarish. I've never even taken to learning Java from scratch my exposure has been so bad. I did learn C# from the command line compiler and a book early on, I didn't have VS to hold my hand. I later did learn VS etc, and that was nicer still.
Getting the pieces together with node/js has been difficult, and painful and a slow process even keeping up with node since it was first announced in 2009. It's taken effort. But anything more than a quick demo, I'd rather have it. I'll leave TypeScript/flow and similar alone though, I don't think they bring more than they take most of the time.
>That said, I'm not saying no frameworks/libraries/tools, and am okay minimizing. But I'd rather use Vue, React, Redux and other libraries/patterns than not in most cases, and find them better overall than ad-hoc jQuery. And don't get me wrong, I've written a lot of ad-hoc and organized code without modern tooling. I'll take today's module systems, builders and bundlers.
I realize I'm outside the mainstream schools of thought. My approach to developing software absolutely depends on having a rare and special kind of developer doing the work. The emphasis is on deep expertise in the core technologies - JS, HTML and CSS -- that will stay around while fads come and go. And quite frankly, my approach does not scale well. It's not that I couldn't keep 40 developers of the caliber I require productive, it's just nigh on impossible for a non-Google-class company to hire that many in one place. (Considering all you've heard about Google's hiring process, imagine how much it costs them to hire a single developer -- even before the first paycheck is cut.)
>As to Java hello world, frankly every time I've had to touch a Java project, it takes 2-3 days to get a build environment running on a local dev machine... it's nightmarish. I've never even taken to learning Java from scratch my exposure has been so bad.
Maybe you're reversing cause and effect. Maybe your exposure has been so bad because you've never made the effort to learn it from scratch.
When I adopt a technology for use, I go really deep. I'll get a book, start on page 1, and work through it to the end. (Sometimes I might deem the last few chapters skippable.) After a few months I'll re-read parts of it as a refresher. (That has proven super-helpful.)
That's why I think Reactjs and its like are a lot of fuss and bother to do something that I can already do easily with a lot fewer moving parts. And also, if I decide to use React, I, and (perhaps to a slightly lesser extent) my developers, will go deep into it, and that takes a lot of time and effort. So I have to be very judicious in what new shiny thing I go chasing after. I need to see an obvious, significant return on that type of investment. (I don't care about my resume having the latest buzzwords.) And I just don't see it with any of these frameworks.
You can test conditional logic, state mutation, and any function without side effects. It takes very little time and definitely improves the codebase. Testing is very dependent on the culture of your team and company though.
>You can test conditional logic, state mutation, and any function without side effects. It takes very little time and definitely improves the codebase. Testing is very dependent on the culture of your team and company though.
For sure. We have a team of people dedicated to testing. Compared to developers, they're easier and cheaper to hire. Our developers, who are very brilliant and very expensive, mainly write code for production, and leave testing to the testers. They write code designed to run efficiently and be maintainable, not to be more amenable to unit-testing.
This has proven to work extremely well for us. If you need unit testing to ship robust code in a short timeframe, it would be foolish for you to not write the unit tests. But our team culture is such that we don't do things that we don't need to do to be productive and successful, and unit testing is one of those things.
functional code is VERY maintainable... If you built it to be thrown away, or replaced, that tends to make it more maintainable. Making something testable, even if you don't write tests, leads to better code.
CSS selectors are highly dependent on the structure of your HTML. If you refactor your HTML then you’ll be required to rewrite many of your CSS selectors. None of that is surprising at all, but jQuery makes it hard to abstract the use of CSS selectors away from the rest of your UI code. Compare with Angular2+ where the logical separation of Components and the HTML of the template is rigidly enforced making refactoring and “scaling” considerably easier.
jQuery is library that provides DOM shorthand - it isn’t a true application framework.
Now evolve that to say React, where your components aren't separated by a DSL from your component code. ;-)
Not to dig on Angular too much, I think it's fine, but I'd much rather use Vue if I was going for small, or React if I'm going for larger interactions/components/applications. Angular does have a lot in the box, but when it takes 4 guys to lift that box, there's a lot less value than 4 smaller boxes.
>If you refactor your HTML then you’ll be required to rewrite many of your CSS selectors.
I haven't found that to be a major issue. And I don't think the term "refactor" is the best choice here.
jQuery is library that provides DOM shorthand - it isn’t a true application framework.
Of course not. I don't want a third-party application framework. I discovered years ago that they're not worth the extra effort or the technical debt one incurs with them. (At least not to me. I'm not telling anyone what they should think or that they should design their applications like I do.)
My approach for me and the developers on my team is to develop a deep level of understanding of HTML, CSS and JS. Once you have that, the prospect of doing things for yourself that frameworks would otherwise provide does not seem all that daunting. To me, frameworks like React, Angular, etc. don't reduce complexity, they increase it.
And I hasten to note that I realize that many bright, capable people hold a different view , and I'm not saying they're wrong.
Something like React isn't even an application framework, though, it's a view library of compositional functions feeding into a tree diff.
How does functional, side-effect-free code increase complexity? What makes your side-effecting code more testable and verifiable than unambiguous in/out functional transforms?
Personally for me it's a trade off and it does increase complexity.
With jQuery and such you just include a script and back then there were no package management system, glup/grunt, webpack, yeoman/brunch, etc..
Also the client side rendering make SEO hard.
jQuery just get stuff done but at the same time the organization of your code is up to you and you do sacrifice some reusability but in general the trade off is complexity.
VueJS as much as I love learning this framework, it is complex with webpack, cli, browser plugin, and etc...
So I'm gonna be honest: I don't have much time or patience for "but webpack is complex" complaints. Webpack is complex, sure. It has many moving parts. It is not complicated, and that means its complexity is more or less "remember a few nouns". Each of the parts those nouns encapsulate is simple and obvious with straightforward inputs and outputs. The interactions may not always be so straightforward--but, in practice, they are. I am comfortable asserting that you shouldn't be surprised by anything webpack or gulp or whatever does. If you are, you haven't internalized what you're working on to a sufficient degree. Because it is just. not. that. big. a. deal.
"Just get things done" when the way you're doing so is less clear, less testable, and less reliable is not, to me, the hallmark of a developer I would trust with anything I cared about.
I've heard about webpack and build tools a ton and I don't understand the issue.
If your project is simple, webpack setup is copy/paste (v3) or no config at all (v4). If your project is doing complex stuff, webpack is more in-depth, but a breeze compared to the make files and XML config of other languages. More importantly, webpack has a very clear tutorial and extensive documentation. You have to go incredibly far off the normal path to run into something that isn't covered. Off topic, but augmenting webpack dev server express instance made dev work many times easier on my current project.
Having worked with large codebase written in jQuery here are some of my thoughts,
There were a ton of times where the complexity of the situation made it so hard to know how to debug something. I mean you had to keep track in your head in a given piece of code what the UI state was, what classes or event handlers were toggled on or off, what the value of various variables were etc. In order to keep my sanity I would have to create functions that basically do what React or other frameworks give you out of the box so it would be easy to reason about everything.
There was so much code like this:
var $snippet = $(<div></div>).append(...).addClass().on('click', handler => {
if (alienState) {
$snippet.off('click')
} else {
$snippet.on('click', handler2)
}
})
$(".parentClass > ").remove() /clear any dom framents in there //before a fresh injection! fingers crossed!*/
$(".parentClass").append($snippet)...
etc etc
I'm sure there is a way to do disciplined excellent jquery code. I'm always quick to say its the coder(s) that make a piece of software code good or bad. But if you were to make a well made jquery application you would need to implement some kind of design pattern, and exercise some kind of conventions and discipline that you could really get for free with a framework.
That sounds like an issue caused by an indecent debugger to me. I havent had to debug JS, but indeed the browser debuggers can use some sophistication and customisability, IMHO.
JS debuggers are generally very good. The level of interactiveness of JS also puts them ahead of most non-lisp languages.
Debugging async code isn't like debugging other code. Once you cross that async barrier, all stack traces disappear (I understand Google's doing some work on making that better).
In this particular case, debugging is harder because the UI is also holding all the data, so in order to reset to known good data, you have to reset the page itself. One of the biggest benefits to moving to React (or similar) is that you start keeping all that data in a central place.
The fact that you're keeping track of your actual DOM and event handling at all? That's what I got sick of. Going from jQuery to React was a huge breath of fresh air.
I am one of those people. It isn't that I don't understand it's purpose. What I hate is when people like my boss, who haven't read a text book on the subject for 13 years pull Jquery out automatically and force it into everyones face before any actual purpose for doing that is established. It really winds me up.
We can definitely agree on that. I tend to think about what the actual JS needs will be. If it's going to be relatively simple then I can hack it together in lighting speed in jquery and it stays easy to work on. If the complexity is going to be higher or needs to do more than a few fetches or xhr calls, it's time to consider something else like React.
Personally I'd rather use Grape over Sinatra. I also wrote Modern[0] to provide a really nice, functionally-oriented OpenAPI server for Ruby that is very much Not Rails. (Looking at porting it to TypeScript soon to provide a similar-feeling API.)
I think more developers should use jquery to prototype an app and move torward a framework when it is needed. The same should be said when adopting redux when using react . Also webpack for building and moving beyond SQLite for data .
> I think more developers should use jquery to prototype an app and move torward a framework when it is needed.
They really shouldn't. With VueJS it's not that difficult to get setup and running with the CLI and components is such a game changer.
Don't get me wrong. I love jQuery, there's nothing better than taking someone else's library and then extending it. Here's an example [0].
However, one thing that I have learned. It's much simpler and easier to develop code in Vue compared to jQuery. It took me a day to develop a wizard exactly the same as [1]. Of which it's now a component which I can drop in anywhere in my codebase (which I do numerous times) and in future Vue projects.
So FuelUX code: 463 LOC vs my Wizard: 180 LOC.
One last thing to leave with. I have a page builder much like Elementor. In jQuery I'm doing a lot with the Dom and it's such a pain in the ass. It's about 20k LOC alone with hundreds of functions.
Today, with Vue. I prototyped a JSX/JSON backed version, where a watcher listens to the store. The UI commits any changes to the store and the watcher re-renders the HTML. I did this with just ~90 LOC.
Completely unbelievable from my perspective and I'm still only a newbie when it comes to Vue, Webpack, Babel, etc.
jQuery/jQuery UI. It was great, but it's time to move on.
> One last thing to leave with. I have a page builder much like Elementor. In jQuery I'm doing a lot with the Dom and it's such a pain in the ass. It's about 20k LOC alone with hundreds of functions.
> Today, with Vue. I prototyped a JSX/JSON backed version, where a watcher listens to the store. The UI commits any changes to the store and the watcher re-renders the HTML. I did this with just ~90 LOC.
The community with the first decent (open-source or paid) page builder gets my vote. I could use that in so many projects (and tell people "Yes it's WordPress" when the underlying tech is something else).
I was expecting it to be React, with each component having a frontend view and an admin view, and then I'd do server-side-rendering, but happy if it's VueJS...
They really shouldn't. Modern browsers have lots of the selector magic built-in, and frameworks that are very, very small can handle the rest without the spaghetti.
jQuery alone isn't enough because it solves a completely different problem. It doesn't have a built in templating system which means you need a seperate plugin at which point it makes more sense to use Vue or React instead.
IMO VueJs is new JQuery (evolotion). JSX(React) is more revolutionary way to build UI and not only web specific. If you need web specific or progressive enhancement or you just don't know modern JS development tools it's better to start with Vue. My friend who can barely code does quite nice pages in Vue. It's a big success of this framework. Although I prefer React as it's gives me more control/freedom.
I used it before then, but could see why a lot wouldn't touch it until well after 2013-2015 or so. Despite following and using pretty early on... it wasn't until 0.8 that a lot of the core issues were really resolved and it stabilized a lot at that point. The ensuing iojs split, joyent handoff, and node org was a lot of turmoil.
Add to that, the fact that NPM is its' own company, and not part of the Node Foundation could be mildly concerning.
Don't get me wrong, I love node/npm but it wasn't really a catalyst until a little later than 2009.
I'd actually say that its less about instability and more about rapid experimentation.
There's two main catalysts - the youth of single page web apps and the openness of the modern web.
The complexity of your everyday website is increasing as they move from static sites to web apps. The pressure to be as responsive and fluid as a mobile app is high, and that requires solutions for data management and synchronicity. There is no standardized way to solve these yet.
When you tie this increasing complexity with the free-form structure of the web (there is no recommended way to build a web app unlike mobile apps or to a lesser extent desktop apps), and then add the growth of JavaScript from a toy language to a full-blown complex ecosystem, you get even fewer established practices.
This is not to say that there will never be best practices but rather that what you're seeing is a language scaling with its needs without having a central dictatorial entity at its helm. You're seeing unprecedented growth and that can easily be confused with chaos and instability.
It's also very unstable and young because requirements keep moving, and developers keep figuring out more 'auto-magic' and simply better ways of implementing UIs.
UI used to be simple, a page, a few links, click through to other pages. Then we needed templated pages, because we had a lot of pages.
Then full blown 'applications' were ported to the web, and life was a nightmare of double-headed dragons and dark holes with pikes. Jquery provided some help here but it was more in terms of nuts and bolts, than as an architecture.
Then came all the MVC libs which really helped enforcing a structure for large apps.
Then came Angular and all of its madness, but a positive step ahead, giving us usable SPA.
React brought some sanity back and now Vue (in my opinion, don't shoot) brings the sanest/cleanest way forward.
My guess is what we do on the web will change once again, in a few years, and we will throw all that out and start again with new framework. Whether it will be in AR or still on the web, there are still many problems to solve.
To expand on your point, somewhere along the line, the web went from a content delivery mechanism to a full GUI + content delivery mehanism.
The allure and advantages to developing a full GUI on the web are many, including portability and availability.
But the downside is that you are suddenly developing to the lowest common denominator - the browser - which not only has to support the "old" web but complicated GUIs (something it was never designed for), which in turn means all sorts of "magic", as you put it, to make it work (i.e. crazy DOM manipulations).
And because browsers were never designed for this, there is no One True Way to develop a simple user action. Search for "drag and drop in javascript" for example. So what happens is, everyone invents their own way of doing stuff, which turned into a race of who is more popular/clever, resulting in a gajillion ways to do the same thing.
There are drag/drop APIs defined... but people make abstractions... This happens even in desktop UI development. There are LOTS of front-end libraries.. but it takes a lot of effort to make them, so there are still far fewer.
The web is easily flexible, and configurable in terms of how you deliver your UX... you aren't hamstrung into creating something that looks like it was made in 1996.
Delivery and flexibility are the greatest pros. But while the delivery mechanism is great, the UI has to be delivered and interpreted every time, through an unknown engine that theoretically follows a spec. This results in a) slower execution times and b) inconsistencies in how the UI is interpreted. Javascript browser applications are simply slower and less stable for than their native counterparts for those reasons.
All that on top of a language which is on par with PHP as far as consistency, and often times worse.
With regards to your 1996 comment, again I beg to differ. Try writing Photoshop or GarageBand as a web app. I shudder to think what that would look like, and the cost each would incur.
There is a reason that "mobile web apps" have lost the battle to native, essentially - the friction to deliver an app is much smaller (app stores), every platform is more-less consistent and the apps are smoother and faster instead of choppy and half-baked for web apps. I would much rather, for example, use a Home Depot app than their website. Or Coinbase vs. GDAX (even though GDAX is more feature-rich). I can't even think of a webapp that I'd rather use (on a regular basis) if an app is available instead.
But maybe I'm just a graybeard who can't accept change. Or maybe I'm simply old enough to remember what 1996 was like for UI, and that the web today feels like those times, just with prettier icons. Likely a combination of both.
I take umbrage with your PHP reference... JavaScript the language is pretty damned consistent, and nice. The browser APIs, that's a different story.
As to PhotoShop or GarageBand, I think there's some more flushing out to do as far as APIs, and frankly that kind of stuff might be workable with WebAssembly. It would take a lot of effort to convert everything, also, not sure if you can use WASM in WebWorkers, but that may be more ideal. As to the cost, are you meaning overhead, or programming cost?
I'd guess a not insignificant number of "native" apps for mobile are at least partly framed web apps. Also, I would probably just reach for React Native and re-use as much as possible, which would still leverage JS for most of it.
In the end it depends... you get a LOT from web applications today. I see it as a bigger boost for internal/corporate/enterprise applications. Because you don't have to manage wide infrastructure and updates in nearly the same ways, you can simply do server deploys and be done.
Also the 1996 apps reference was in that most UI toolkits default to looks that haven't changed much since 1990's and aren't skinnable in the least. I remember the late 90's and early 2000's, litestep and winamp were two of my favorite things. I know that was user customizations of apps, over developer/designer. All the same, maybe I want my site/app to look like Material Design over Bootstrap, or something else.
It's because of all the icky data-sync for the client/service round trips. An async, slow, failure-prone concern wired throughout the entire codebase. I/O hurts everyone in subtly different ways and it's hard to solve so you see a lot of churn.
service/database has this problem too (the ORM flamewars of the 00s) but to a much less degree because it is in a happier place on the latency chart: https://gist.github.com/jboner/2841832
Continued: Churn is further compounded by the Lisp Curse; an individual hacker can scratch his itch in Javascript all by himself and then stick it on github, the other three 1995 langs (Java/Ruby/Python) never quite reached this level of empowerment of individuals. That's great but also means an explosion of activity from smart inexperienced hackers who don't quite have a full understanding of the problem they are trying to solve because they are just one person. And when the problem is I/O that's a recipe for going in circles. http://winestockwebdesign.com/Essays/Lisp_Curse.html
Not really - those things exist in native frontend as well but it's more stable than js. Writing GUI is not trivial, people take stuff like data binding/templates/MVC etc. for granted nowadays, go check out early desktop UI libraries like say win32/MFC/winforms/gtk+. Took a long time to get to stuff like WPF, Qt quick, etc.
With browser there were several constraints :
ES5 is extremely error prone and scales very poorly to larger codebases
slow JS VMs
missingt for sane layout like flexbox
1. Got better with npm, we pack, es6 transpilers, typescript/flow and tooling built along the way
2./3. Got solved by finally depricating everything < IE9 (IE in general if you're lucky) and tooling
I'd say angular is relatively stable now and decently productive once you get a hang of their patterns, I've been using it for different frontend projects since early betas, written ionic apps for mobile with it - have very few complaints about the framework, some of the stuff didn't always work along the way (like ahead of time compilation, i18n) but recently (5+) I've had 0 issues with the framework and I'm quite productive with it. I think the alternatives will mature as well and things will settle.
Native frontend on desktop and mobile have in-memory database and sync out of band. Web browser tab is too resource constrained to run a in-memory database, and less secure as well (you don't have root on your iphone so Facebook can restrict your data access patterns, prevent scrapers, prevent mass-delete of all those old posts, etc)
Those things don't remove the sync issues, blocking the UI thread, they just enable a slightly better use experience.
And dealing with asynchronous code is not the only issue in writing frontend code, GUI apps can be really complex beasts, managing that complexity effectively requires tools that didn't exist in js world untill recently.
They kind of do though, the complexity of data sync can now be dealt with entirely separately from the complexity of the app logic itself (which is not true of, say REST architecture)
I agree with this. It also doesn't help that frontend code usually has to run on several versions of at 3 different runtimes. With backend code you can freely either stick with an older platform for stability, and adopt a new platform for new features. With frontend you don't have that choice in the same way.
Related to this is the idea that people believe front-end software development is easy. In reality, it's one of the most challenging, chaotic and difficult parts of the development chain, especially when talking about web development.
> "Put yourself in the shoes of a junior-to-mid-level JavaScript developer, writing a new application for the first time.
> "It starts innocently enough. You have a completely clean slate and want to keep things simple. You are a devout Agilist and YAGNI is your watchword. So you begin with a ‘simple, barbones framework’. That sounds good, doesn’t it? (Even if it did not, that’s often the only choice you’ve got).
> "Being barebones it does little, so the task falls on your shoulders to choose some helper libraries. If you are doing frontend work, it might be helpers for Redux for forms and API requests. If backend, it might be middlewares for Express."
Why is the default answer to the problem of "it does little" to "choose some helper libraries"? Shouldn't it be: "I'll write some code to make it do more"?
Why does this first-time app developer need Redux? Why would she need a helper for Redux to do API requests? These days you can just call fetch() — it really doesn't get any easier.
All the APIs that one gets in the browser or in Node.js are already very high-level and easily understandable by entry-level developers. If anything, piling more leaky abstractions on top will just make things more complex. How did it happen that junior JavaScript developers feel so insecure about using the standard stuff?
Maybe it's because more senior developers have spent two decades crapping on the standard stuff. Therefore, as a self-appointed spokesperson for the veterans, I hereby grant all junior JavaScript developers full permission to just go ahead and program whatever they want without having to look for a single library or creaky build tool — and they shouldn't feel bad about it in the least.
> How did it happen that junior JavaScript developers feel so insecure about using the standard stuff?
I know a lot of "developers" that are learning frameworks and not javascript. When you start in a black box then you look for other black boxes to fill in the gaps you need.
I am not even opposed to that. I strongly believe in never re-inventing (unless its academic) what can be used and has been tested, but when the developers are not javascript developers but react, vue, angular developers. Then you get this gap where the inexperienced are afraid to outside of what others have built.
One has to learn somehow, and reinventing is an important part of that.
Aspiring artists will paint nude figures and flower arrangements and white cubes on draperies, even though these subjects were done to death centuries ago. Junior programmers should feel the same way about doing things that have already been done: someone probably has done it better, but I still have to do it for myself, maybe a few times over to get the hang of what's the point.
The junior developer has it easy. Imagine being the tech lead for a startup that has to pick the framework that'll be used to build an application that needs to e maintained and that they'll need to hire developers for to maintain.
I was this person, I made a choice and now every single person is casting doubt on it. The choice wouldn't have even mattered, It's near impossible to find experienced front-end developers of any kind right now and everybody will always ask "why not [this framework I'm familiar with]?"
Asking this question without looking at the underlying platform is ignoring a huge source of insight into the problem and its solutions.
Growing up from a document viewer to an app platform in the particular way it did left the web as the most important and ubiquitous GUI app deployment and runtime platform, but without a dev-accessible native GUI component system. No other major modern UI platform was _ever_ in this state.
This left all the work of creating the most basic widget system to userland, while at the same time the platform's features available to that userland were relatively paltry to begin with. As those features of JavaScript and DOM grew, frameworks were released frequently because newer ones were designed to take advantage of newer features. This lead to a huge amount of fragmentation.
Which brings us to now. Outside of threading, the runtime features are much more suitable to an app platform now, and finally the web has a native UI component system in Web Components. I think there's one or two more big shifts ahead of us toward Web Components and better utilizing Workers, and then things will settle down because we'll finally all be using the same platform features. Even if we still have opinionated and helpful layers on top, the foundation will be more stable and standardized.
> Growing up from a document viewer to an app platform in the particular way it did left the web as the most important and ubiquitous GUI app deployment and runtime platform, but without a dev-accessible native GUI component system.
Exactly this. Web technologies cannot be understood without understanding the history of the web itself.
You could just as easily imagine, say (random example), Internet + Gopher + Lisp playing the role of Internet + HTTP/HTML + JS.
It was somewhat inevitable chain of events, plus "right time, right place"for particular choices of tech.
The question that intrigues me is why Tcl/Tk didn't take off around 1995. We could have had internet apps about a decade before web apps were feasible.
The only answer I can think of is that no one in 1995 would have contemplated running a Tcl script that their computer had just downloaded from a random host, but Javascript boiled that frog gradually. I'd be keen to hear other ideas.
> What is to be done? ... Consider non-microlib architectures
What about what happened with Angular (1), the 3rd library on your list? That whole framework was deprecated in favor of Angular (2). Additionally, the list of front-end libraries given doesn't really reflect the landscape front-end devs have traversed.
Years ago, I remember working on an app that heavily relied on YUI around the time that library was shutdown. I also heard stories about how ExtJS 4 wasn't backwards compatible. I'm not sure I buy into the author's thesis about the reason for front-end fatigue.
I feel like Angular was reaching escape velocity and becoming the one thing to learn and just as it was happening they announced Angular 2, which was originally going to have no upgrade path and no two-way binding (arguably the killer feature), out of performance concerns that weren't a problem for the vast majority of use cases. Even after they backtracked on a lot of this it was too late; they had forever tarnished their own product.
I really loved Angular 1 precisely because it had everything you needed, more or less, built-in.
Additionally, there was no initial update path between Angular 1 and Angular 2 until the community cried foul. It's been a while, but after they finally released an Angular 1 to Angular 2 bridge, I read several tutorials and books that discussed the upgrade path but said to absolutely not use it in production. It would kill the performance of your app. That sort of made the value of an Angular 1 to Angular 2 bridge pretty weak. If one were to migrate to Angular 2, it would seem more time efficient to just to start fresh rather than spend time on a wonky bridge between the two frameworks.
At AngularConnect last year I now migration of large applications was a topic and if you are interested I think it should be available online.
Also: angular.js isn't dead, it is just stabilized. Yes, no more features (from Google) but it should be safe to use and also I think they signalled quite clearly that they would be happy to let the community pick it up.
Front end has to "look good" and be "good enough to use". Little finnicky errors are often tolerated as a user can be expected to work around a certain issues. Since issues can be quickly spotted code hygiene and rigorous software quality practices aren't as important. Also, performance and resource utilisation aren't as important, to a degree anyway.
Back-end, unsupervised, or high-performance code on the other hand is largely invisible to the end user. It has to run for long times without error and if something goes wrong can be tricky to get at, diagnose, and repair.
These requirements demand a higher level of code-hygene. Coding standards, logging, documentation, automated testing etc. etc. which makes developing this kind of code much more expensive than you would expect for something that you ultimately "can't see" (unless something goes wrong).
I've seen organisations struggle with this dichotomy where they're trying to do both without appreciating these distinctions. Product managers getting annoyed that it takes so long to get GUI changes in, while engineers getting annoyed that their (necessary) tech process overhead often gets shorted because "it doesn't add value".
You can get away with taking short-cuts on the front-end, that you can't get away with elsewhere but ultimately this leads to a messy, unwieldy codebase with mountains of tech-debt and once you get to a certain point it makes sense to just throw it out and start it again.
Though "starting again" is considered to be something you should never do [0] with front-end stuff it's a little easier to get away with because, like I say it's all visual, and a good bit easier to test and play with and tease out the subtleties.
I disagree. Rigorous code standards are just as necessary on front-end code as back-end. In my experience, no code review, poor design or documentation, and lax standards lead to jittery experiences. For example, people write complex blocking code that causes the page to freeze for moments while that processing is happening. People write event handlers that execute too often and slow user interaction.
My point was though, that such issues can be quickly picked up just by using the application, and you can usually elicit the causal factors by prodding at it. Often times such issues in non-UI code can go for years without it being noticed.
But broadly speaking yes I agree, but one or two of my line managers in the past would not.
I work with elm now for frontend. It has often been criticized for being too conservative and slow paced in its development. But the result is that code is easier to maintain, libraries have a great quality and most of the time there is only one, obvious, way of doing something.
1. By chance, after a long and tiring typescript project, I looked for something else.
2. No, my first functional language was clojure script, which I tried before elm, I liked the concept but just couldn't bare the syntax.
3. Phoenix elixir, I do maintain it. (in fact I'm more of a backend programmer)
4. Now I work mostly with style-element (the alpha, called stylish elephant) which is the only "framework like" library I use. Otherwise I use different utility libraries for date, color manipulation (hsluv which I wrote the elm implementation), data structure manipulation (many in elm-community, like list.extra)... I use way less libraries that I would with JS, and that's good.
I wrote this in Elm: https://ixberg.com. (feel free to sign up, use it, share it). I learned Elm by going to an intro course at Lambdaconf, and watching this https://frontendmasters.com/courses/elm/ and coding in it. Elm was my first PURE functional language. I've done some Scheme before. Backend is Firebase. No special libraries in particular. No regrets.
Having worked as a frontend dev for a while, my observation is this - because the UI is closer to the user, and closer to the product side (for consumer-facing apps), it has faster iteration cycles, doubling or tripling the rate of technology decisions that are made.
You'd never rewrite an API twice in a year, but maybe you'd rebuild your site or interface that often, if you're responding to users.
With increased demand for tools comes more supply - more libraries and frameworks with more stars and follows.
The reason we don't see this in the mobile world is because it's a much more closed ecosystem - there's no open spec for mobile like there is for browsers.
Yeah, there's a perfect storm that happens on the frontend.
(1) Users are more empowered and expected to continually change requirements all the time.
(2) The underlying technology is broken. Specifically, there is no component model that makes it easy to compose existing components. Every frontend app reinvents not just the wheel but the tools to cut the stone.
(3) There is no UI component model provided by the browser. Every webapp looks and behaves completely different from every other webapp. There are no common metaphors, common components, common fonts, common icons or even best practices. Every company writes their own UI kit and interface guidelines.
(4) Thanks to the highly intelligent fuckers at Google/Facebook/Yahoo everybody has very high expectations when it comes to Web UIs. These guys are also driving rapid innovation in the space. Unlike Windows apps -- which haven't changed much in 25 years -- Webapps seem to become dramatically different every 3-5 years.
The result is the wild west meets post-apocalyptic disaster where everything is constantly changing and the developers don't have the tools to actually manage or control this change.
I agree. Throw into the mix non functional aesthetic changes. Even if the actual requirements/functionality didn't change new design trends would mean you'd be at minimum re-skinning/themeing your site/web app every 18 months or so just so it doesn't look "dated".
Makes you wonder why there isn't something like Rack or WSGI for front-end components. If there was one interface that all frameworks adhered to, you could, in theory, use components written in one framework from within another framework. Elm sort of does this. Not sure about others.
Yes, no one takes the time to research. Most popular does not equate to good.
I took the time to do the research back in 2013 and found that Ember basically solves all the problems OP talks about. Tom Dale, one of the creators of Ember, blogged over a year ago [1] about how Ember has been around since the days of Backbone. As the years rolled on, other frameworks became obsolete, but Ember evolved (with sane upgrade paths) and is still relevant.
I think I'd add to this comment that you can't solve a lot of these problems with more code because in the browser code size is a very significant factor. On the desktop, or even on mobile, code size isn't very important. So we can have big libraries that do a lot of work.
well, ever since the proliferation of mobile touch devices (it's been less than a decade) there's been this somewhat perverse history of "mobile-first" really meaning "native-first." tons of mobile websites are still essentially ads for their counterparts in app stores. this is for lots of reasons (mostly to do with financial incentives for both app makers and app stores), but ultimately the web still isn't as good as native. the chaos of the front-end world is a reflection of web developers trying to solve that, rather than those working on operating systems, web browsers, standards, etc.
the other reason is that vue hasn't be around very long ;)
It's unstable because there are millions of developers working with front end technologies and that community is rapidly finding better ways of doing things.
And when better ways are found, because development is so time consuming and difficult, people move to embrace those new and better ways.
I'm glad of the rapid change because there is alot more improvement needed at the front end.
Consider for example the vast amount of crushingly complex configuration that is required for many front end development tasks - this is yet to be addressed, although giant leaps have been made - such as create-react-app - thank god.
JavaScript has to become "zero configuration" to the greatest extent practical, and JavaScript programmer should dump configuration heavy tools as soon as they possibly can for zero configuration tools.
Compared to how much code size and complexity for the tooling increased the benefits for the end user are tiny, if there are any at all. We are just getting SPAs which do not work with JS off where a simple static page would do perfectly.
I know that the star is a de-facto way to calculate the popularity of particular repo. But, let me put aside this calculation for a moment, then I myself have stats around 250 repos and more than 40% of it, I have starred just because to refer it sometime later, not because I like it or love it, just for future ref., when I have to work on something that might need help or ref. from that particular repo.
So, if there is something that have a way bookmarks the repo then chances are hight these repos have lesser stars. Also, there is popular trend nowadays I see that 100 other developers have starred one repo for some particular reason that chances are hight that other 5 developers might star it because they just follow the developer, who have starred.
FWIW, the GitHub age isn't entirely accurate because both React and Vue had major rewrites and the current default branch doesn't hold the entire development history. Vue's first public release was Feb of 2014 and React is just celebrating its 5th birthday.
I do wonder how much of the instability relates to the strong opinions that the JS community often has (I'm guilty of this too). Any broad stereotype is going to be wrong a lot, but it does feel like JS devs on average are a lot more opinionated about the way things ought to be. They are also more disgusted and unhappy working in codebases that don't conform to their standards.
I truly don't mean this as a bad thing, in fact I put myself in this category. Obviously it can be a bad thing if it causes you to be unable to work with others, but it can also make for elegant, consistent, and hackable codebases that boost productivity and minimize bugs.
Programming industry is the only one more fashion-driven than the fashion industry. This is a known issue going back to the dawn time.
The big secret is that for most of us this is "works as intended". We like to program. Imagine being a mountain climber who loves climbing and gets paid (well!) to climb mountains. With real mountains, eventually you get to the top and you have to stop. With programming, the act of climbing the mountain creates more mountain above you. As long as you can find people to pay for it, you can keep climbing.
If you like programming please shoosh so the marks don't get wise.
If you just want to get shit done use Elm-lang and get on with your life.
Nice idea, but some hipster developer at a company we query decided it would be A Good Idea (TM) to switch from using DNS to query their data to a REST API over HTTPS that returns JSON. Normally, this would be a "whatever" type thing, but see, the code I work on is in a critical, real-time path (as a phone call is being made). We were promised that Java client code would be provided, but see, we don't use Java.
I always find it curious that almost no one is talking about not using a framework to begin with. I have been writing complex UI with vanilla js for some time now, never had any issues or took longer than I expected to get a job done.
You have to ask yourself whether, in the process of building that complex UI, you haven't invented another framework.
After the first few web apps, one necessarily sees the repetitive elements and wants to capture that in libraries and abstractions. Routing URLs? Reuse that widget? This is just coding instinct.
I think it is possible to have a stream of projects where each new one gets its framework copied and pasted from the previous. That does not mean it's not a framework, just one without name and documentation.
Then there are projects of a certain size that you wouldn't want to do like this. Stakes too high, more than 1-2 developers. Then you can either brush up your framework and give it a name and documentation and contribute to the 'problem' of too many frameworks.
Or use something that others use and is just as suitable.
The real complaint right now should be Babel v7 -- or more specifically, the community around it. Babel v7 has been in development for quite a while. It's still changing and isn't even a RC yet.
I understand the need to test new compiler builds, but production isn't the place to do that (especially not until a RC is released). A beta designation is meaningless if its used as though it were just a normal production release.
Using a beta compiler in production tools, libraries, etc shows EXTREME inexperience. I have a number of dependencies that I cannot update because the new versions rely on various v7 builds (ranging from ancient alphas to bleeding edge betas).
I suspect the issue is that most JS dev teams are small with projects small enough that edge cases are uncommon and easier to find and fix.
Stackoverflow dev survey shows that the average US dev is 28.7 (down from 31.6 in 2015) with an average of 10 years of experience despite 58% having less than 5 years of experience.
That means that on average, you have a very young team. Further, the vast majority of teams most likely do not have ANYONE with more than 5 years behind them. It's not surprising that such things aren't recognized for the disaster-in-the-making that they are.
Web technologies have settled down and stabilized dramatically over the past 3 years. I suspect that as more web devs gain more experience, we'll see this continue as rashness is replaced with a more conservative approach.
To me this attributes more towards general evolution of the browsers which naturally leads to more complex requirements for the modern web-app. I remember that in 2010, when we started re-writing our desktop apps during our transition to web, the initial requirements were such that it was still possible (although very hacky) to use the latest version of ASP.NET.
Then 6 months later we received a new set of requirements which rendered this _latest_ stack useless. We literally had people sleeping under the desks pulling all-nighters trying to transfer a session state between browser and server without dropping an egg. That's was around time when backbone.js came out.
If we compare the things that browsers can do _now_ with what was possible 5y or 10y ago, no wonder why we have so many new frameworks.
Browsers are trying to catch up with (or even replace) an OS as a main container for running an application and because it's hard to release all features _at once_ without agreeing on standards, we have a situation when people have to choose a new library each time a new major API/feature is available. And this what I see is the main reason why `front-end development is so unstable`
WebAsm is coming which will again lead to yet another huge re-write.
(1) The closer you get to the front end (especially GUIs), the further you get from algorithms and other mathematical certainties and the closer you get to something that's more a matter of design and fashion. People get tired of GUIs looking and feeling a certain way and they change, and that often requires a code re-work because changes in GUI look and feel can break previous abstractions.
(2) Form factors and display technologies change. Mobile completely annihilated 30 years of GUI work for example. To a lesser extent high-DPI screens have done a number on some older GUI systems. The stuff under the hood changes too but more slowly and in less fundamental ways e.g. an SSD is mostly a "disk" from an abstraction point of view even though the tech is wildly different. Command lines are fundamentally unchanged since the 1970s. VT emulations have become more advanced but the basic paradigm is incredibly stable.
(3) GUIs look deceptively easy but in reality they're incredibly hard. A modern GUI layer is at the same level of complexity as a modern game engine like Unreal. The list of edge cases and layouts and modes and languages and interaction metaphors and input devices and so on and so on you have to support is very long. Software engineers routinely underestimate this difficulty and try to re-invent GUIs and GUI abstractions in a quest to be clean and small and simple. They're then forced to shoehorn everything a modern GUI requires into these clean and simple abstractions, breaking them and turning them into overloaded monstrosities. Simple clean systems get ugly when you try to extend them too far. A more elegant design could be achieved by starting with the premise that the problem is big and hairy and developing heavier and more flexible abstractions accordingly.
There are a set of converging interests causing pain in the ecosystem. First a company promoting a project, then people hoping to provide services, consulting or training, and third people hoping to get a job or padding their current resumes.
None of these people have a particular interest in the technology or its merit, its anything they can make money or benefit from and then forums like HN and others are used to aggressively push out these technologies.
By the time people start catching on about flaws, over-engineering or needless complexity the latter 2 groups have already moved on to something new. Rinse and repeat and HN is very much a part of this.
There is very little tolerance for skepticism and proper scrutiny of anything 'new' and in the hype train. These are usually dismissed with ad hominems about 'grey beards' or people 'scared' of change. Since everything is so 'new and cutting edge' there is no respect for any authority or experience and scrutiny becomes impossible. And since self interest is involved the arguments become needlessly charged.
I assume that observation applies to more than just front-end development. At least I've personally seen the effect more generally.
I think another self-interested actor is the "nobody ever got fired for buying IBM" manager, except that, today, it's not IBM. IBM has, of course, been replaced by whichever technology is currently in high repute. They need never gain technical knowledge themselves.
Granted, in some cases, that's actually a case of legitimate delegation to their engineers (whose conflicting interests you've addressed), but, since they have the same access to the Internet as everyone else, that can't be assumed.
> And since self interest is involved the arguments become needlessly charged.
I hadn't fully considered that until now.
That isn't to say that I'm not a cynic, but, rather, that, for example, the average HN commenter may not want to come to terms with having, even to a slight degree, such self interest. That can lead a form of needlessly charged argument, defensiveness, that wouldn't necessarily be obvious otherwise.
In a professional setting, it can lead to seemingly-irrational decisions in the face of overwhelming evidence that the other way is better.
I think this is a piece by someone who has not really been in front end web development as a career (or does not specialize in it) for very many years. The author's thesis is that 'poor choice architecture' is what drives churn in front end. I disagree. The churn in JS tech has reduced over the past few years.
Some of the churn was in part due to limitations of the web platform. ES5 was limited in features. The web has evolved a lot, and now we don't need to depend on every framework re-inventing things like modules.
React has been around for 5 years now, and it's still the dominant choice for front end architecture. This is not by accident. React solves some of the biggest challenges in front end development with a clean, minimalistic component API, a virtual DOM abstraction to minimize re-renders, JSX (still superior to underpowered template languages), and one-way data flow (which makes refactoring and state management easier). Also functional component model has implications that it can render to any target.
React is not just a web development innovation, it's a UI innovation. There has never been a desktop UI framework with this much flexibility and power. You can target web, native mobile, native desktop, create custom canvas-rendering if you like, render to touch bar, VR, etc...no limits.
The author sort of implies that a one-size-fits-all framework or some standard way of doing things across all projects would lead to less frustration and less churn in the JS community.
We've seen many attempts at this (Ember, Angular, etc). The thing is, not all projects have the same needs. The web is a big place. There are simple web pages with mostly text content (and little need for JS) all the way to rich apps with interactive video, real time chat, dynamic graphs and complex customized data feeds for each user. There is no one-size-fits-all comprehensive framework that will ever meet those demands. For the easy projects, they're overkill and for the complex, long-lived projects, you're likely to beat your head against the limitations of the framework.
The innovation that has come in state management, style management, and component-based design in the past few years is unrivaled by any other UI community. I think the article is a surface level grasp at the situation. It implies that the JS community is spinning its wheels instead of producing solid long-lasting paradigms for application development. I disagree.
It nicely shows HN bias when comment like this is downvoted into oblivion. "JS framework churn" is a thing of the past, with only a few I-hate-JS types trying to keep it alive.
and I'm not sure it even applied to people actually working with JS even in the past.
> Before we get carried away, it’s worth validating whether the meme really has basis in reality. Do front end technologies actually change that quickly?
> In the sense of major view technologies, probably not. Consider this list of the highest ‘starred’ JavaScript front-end technologies on Github:
Seriously? So you're basing front end technology changes based on stars on Github?
Also I think it's wrong to say that front end technologies have "changed." It's just people have discovered better or just different ways of building things and that should be okay.
I agree with the portion how it can be overwhelming to a junior developer, but it is also easier in some ways. In truth, if you are a junior developer (or hell any developer really), your focus should be less on choosing the right technology and more on understanding what technologies fit what needs.
I think the stars were mainly to get some kind of data, given how it's hard to get real live data.
The article itself resonates with a non-data-powered feeling I've had for years. The front-end change meme is from a few years ago, during the highly unstable period shortly after Angular and (with it) single-page applications rose up, then suddenly Angular 2 was announced (which made everyone feel like their investment and commitment to Angular 1 was a wasted effort), React climbed up and got a few alternative implementations, React state management got a couple of alternatives (and a lot of thought went into those; I see that era as more of a philosophical era), and there was some drama on both the Angular and React side, causing people to split off and do things like create a new programming languages (Atscript), Angular the good parts (Aurelia), and (a bit late to the party), Vue.
"Unstable", "Issue", "Problem". The opposite of the front-end world is not something to desire. Change means people are fixing problems and giving away their solutions for free. To desire less change means wishing fewer people gave away their work.
Change is not necessary a good thing. Bad change is worse then no change at all. However, when it comes to programming frameworks, rapid, unstable, change is generally not very desirable. Take a look at the most popular open-source OS/kernel in the world: Linux.
It's a Unix-clone using a monolithic kernel made in a time when Lisp machines existed and microkernels were all the rage. It used ANSI C which, to be fair, was new at the time, but stuck with it, in the face of C99 and C++.
And, as a windowing system, it still uses X11, a system designed in the mid '80s. There has been a very slow migration towards Wayland (which is now already 10 years old).
As it's scripting language it uses bash (1989), which inherits a lot of it's characteristics from the Bourne shell (1977).
Yes, in the case of Linux may be a little bit biased as OS development back than was still more mature than front-end is now, but still, the principle still holds: "slow and steady wins the race".
I am ok with that. If more people giving away their work ends up with "left-pad" situation. Also most of git hub repos are crap to be abandoned after couple months.
As in article if people share code to be "ninja", "rockstar" devs that is wrong approach and I don't want that code. Because solution is not honest, not reliable and probably will be dumped as soon as "rockstar" will find another flashy term where he can gain more "ninja" points.
The article is right about the frontend side, but I think it doesn't look to good on the backend either. The actual code side (frameworks, libraries and languages) is relatively stable, but mostly because most of the complexity nowadays goes away from monoliths and into services, of which many are off the shelf, and plagued with the same issues as all the trendy javascript frameworks. Nosql-database du jour that solves some problems allegedly very efficiently, messaging systems, thousand ways to script infrastructure and virtualization.. the hype is not in the code and libraries, it's in the infrastructure components, but that doesn't make it any better.
I do find a bit of the argument hyperbolic... My preference today is Koa + about 3-5 modules (including 1-2 custom) on the backend and React, Redux, Redux-thunks, redux-first-router and fetch (in most browsers) on the front end. For UI, will generally suggest bootstrap or material-ui (the library). create-react-app does take a lot of the guesswork out of the front-end.
In the end, front/back-end tech grows at an incredible pace. Look at go, rust, .Net Core and a half dozen other options that sprung up in the past 6-8 years and all the frameworks around them. It isn't new. Last generation(s) it was Python, Ruby, C#, Java and on, and on... this isn't new.
I believe classifying libs/frameworks by number of stars is not enough ; "when" these frameworks got stars, or how many stars they got in the past 6 month, is probably important. Not sure jQuery or angular 1 got many stars lately...
Agreed. The quality of the stars are important. What percentage of people star repo's where they've read the Readme but never used the code or used the code and decided it was bad quality and never unstarred the repo
This problem is way deeper than web front ends. There were/are dozens of frameworks for Windows and Unix front ends too, before browsers were invented. Every year or two some new tool became popular which made all the previous ones obsolete.
Not really. On Unix it was Motif vs OpenLook. On Windows it was MFC vs OWL. On Mac it was PowerPlant vs MacApp. And that was it really, and you could have stuck with one for over a decade and never been short of work.
I built a system for filtering my HN feed and one of the first rules I put in was "no links to sites that have pop-ins" and tedium matched that. (Not like I have an automated way to do that but it is not too hard to add sites like that to a block list.)
I give the author credit for thoughtfully trying, but not much of this seems to actually explain why it is Javascript that has this problem, more than Python or Ruby or Java or C#. The reason for Javascript's excessive churn must, I think, be related to its most typical use case as a browser client-side language, and in particular to its use in single-page apps. The rest of it are all good points, but I'm not sure they explain why Python, etc. don't have this problem (as bad, anyway).
So here’s the thing... browsers are generally excellent at maintaining backwards compatibility (see “SmooshGate” for christsake). Old frontend frameworks still work just fine, therefore when you complain about frontend development be “unstable” you’re complaining about other people creating or adopting new frameworks/tools. If these things really aren’t worth learning you are free to continue using the old ones.
- not a universal framework (you cannot create an app that has a static `/about` page but a dynamic `/search` page. With Next your app is either all static or all dynamic.)
This offers a good explanation for why the most popular options are generally the ones that developers choose, but not why the thing that is currently popular shifts as fast as it does.
Why did React take the limelight while Angular fell out of favor? What caused that shift in what is popular?
My personal experience with this was that Angular 2 was announced, which was completely non backwards compatible. So there were a bunch of people that
1) Needed to rewrite their code
2) Distrusted Angular to be stable
React had a philosophy of small reusable components and incremental upgrade. This is a particularly attractive idea for people experiencing the above. I think this is probably part of the shift in popularity.
> Put yourself in the shoes of a junior-to-mid-level JavaScript developer, writing a new application for the first time.
> But how could you do better, Junior Developer? Who was there to guide you? The Senior Developers, too, are learning as they go. We’re caught in this avalanche too, just trying to keep up to date and remain employable.
No, this is exactly the problem: junior developers making major architecture decisions unassisted. The "senior" developer who is "caught in the avalanche" isn't senior by any reasonable definition of the word.
Development in general, and, it seems, front end web development in particular, is both easy and hard. Somewhere between batteries-included frameworks, blogs, YouTube channels and MOOCs, it's very easy to get from zero to something good looking and more or less functional pretty quickly. But this can easily hide the complexity that pretty quickly creeps in, and reasoning about complexity is one of the crucial things that a senior developer brings to the table.
Yes exactly. Building something fast without understanding the complexity behind can be both a blessing and a curse.
It's a blessing for the one who needs to deliver.
It's a curse for the one who needs to learn.
To be efficient in the long term, and not get lost among the framework wars, one must use what was learned in Engineering School, University or from seniors : write specs, set priorities, plan ahead. Use or build the tool that suits your need, do not follow trends.
> But this can easily hide the complexity that pretty quickly creeps in, and reasoning about complexity is one of the crucial things that a senior developer brings to the table.
Yes- so true. Taking a web app / whatever from the beginning stages to handle proper complexity, anything from scaling to ... the business wanting cough forcing cough a "quick fix".
We're missing the context here. For many (most?) web apps I feel like things have kind of gone wrong if the front end requires "major architectural decisions". That's not a knock against the greatness of React, etc. so much as it's my observation that most web apps are small and from a technical standpoint quite boring, a smattering of static pages, a couple web forms, etc.
I think the instability is a byproduct of what it is.
Today we have https://caniuse.com but this sort of tool in one form or another has existed for a long time - I think that mobile, and native mobile have only compounded the problem.
Frameworks, and ecosystems (Node, and what it spawned) are practically a requirement to get things done. But abstracting issues away from front end developers, hiding them behind hacks and fixes only serves to make things worse, and harder.
The foundation of the front end isn't just broken, it is a fuckstorm... and we keep piling more stuff on top of that foundation trying to make it make sense, make it rational. With each layer we add creating new sets of issues that only get worse over time. Then someone "smart" enough to build something a bit better does just that, and we get this: https://xkcd.com/927/
(Note the above links are just my favorite "examples" of these issues not even a good indicator of the real problems therein, everyone has their own grievances)
And this IS the heart of the issue. Any one sane, any one with a high level of skill isn't going to want to work in an environment where they aren't in control where someone can yank the carpet out from under them at any given moment. No one is going to stay long when new browsers and missing packages create more work, or work that has to be re-done to deal with some edge case or issue that literally came out of nowhere. Front end people see those of us working on server side code where we have a HIGH degree of control (orders of magnitude) and realize how bad it is for themselves. Simply put any one who is good invests the time and energy to get OUT of that line of work. Tallent will quickly migrate away from the broken foundation of the front end. This only serves to exacerbate the situation, as a new "smart" person onboards thinking "I can do this better" and being right to some degree.
Because on average, the skill level in the whole webdev community is low. Below that of "a dev with average schooling, and average industry experience"
Webdev has close to no barriers for entry, and is under the strongest influence from the fact that the evaluation of deliverables is not done by another tech professional out of all software development jobs
The combo above makes things that are unheard of in other software development niches possible, like meeting borderline or de-facto frauds running webdev businesses selling "ultrasophisticated" corporate websites on Wordpress to F500 types for few $k USD per hour of dev time, or well entrenched in-house "developers" in tech giants who made it to six digits on technobabling to non-tech managers and copy and paste.
All of this is evidently projecting onto to the tech solutions used in the trade, and the popular image of the webdev development process. And over the time, popular stereotypes are becoming self fulfilling prophesies: Angular - was an okish framework at near 1.0, but marketing messaging made it look like an "enterprise stuff," and their devs eventually turned it into it - purposelessly overengineered monster filled with SOAisms; jQuery - got bad fame for unusable, animation rich websites made by least talented part of the dev community, and this infamity has both sealed its further development, and was responsible for attracting even more unskilled devs into making websites with crawling slow animation;
And like this for few pages, as well as for the notion of webdev world being "unstable." Many people here who are not novices to webdev can probably call few sites that had a "single piece of JS code continuously maintained over 10 years" or more. I myself knew people who were making yandex.ru homepage in nineties, they say that a very sophisticated ajax autocomplete code there was developed and maintained continuously since 2004
I don't necessarily disagree with you, but I don't think that has to do with FE development in general. Some of the sharpest/smartest devs I know are FE JS guys. Seriously.
On the backend I have seen probably more "low skill" and even "incompetent" Java devs than any other language. Why? My guess is the low barriers to entry (as you say) and the ubiquity. I've also worked with some amazingly sharp Java devs too.
You're definitely right about the "borderline or de-facto frauds running webdev businesses selling 'ultrasophisticated' corporate websites on Wordpress to F500 types for few $k USD per hour of dev time." That's a rough problem and makes legitimate big-time web development bids induce sticker shock.
How does something like regular desktop development have any more barriers? I picked up Python at 14 and that had absolutely no barriers for entry other than installing IDLE on a laptop. If anything, webdev has more barriers for entry because you need a webserver and a backend of some kind.
I would say: 1 loose portfolio of demos without any history of signed off successful projects will still be getting you a decent job in webdev, but for sure not in commercial desktop software industry
Well, I can say that I am just coming out of few years long miserable period of my life that came after I was kicked out of Canada thanks to "LMIA disaster" after spending 6 years of my life securing my future in the country. Right now, I am easily passing under the points system, but nah, after such shameless bait and switch with immigration policy, there is no way I am going back.
Good things, I am out of webdev as an industry, and finally working in a more serious function in an engineering consulting now - something I was aiming at from my teen years, when I was still not in "money over job satisfaction" mode of thinking.
Knockout is incredibly lightweight and easy to get into. However, it's not the "in" framework like Angular and kin, so it gets a lot of bad reactions.
The biggest reason for this, I believe, separate to any technical benefits/advantages, is that the "big" and "in" frameworks have a gigantic knowledge-base hosted on the entire internet. Because of the sheer amount of people on that ship, they've all collectively encountered and tried/solved/figured-out a large amount of usages and combinations of the library. What that means is that mediocre and/or "shallow"-knowledged developers can easily leverage that to solve problems they might encounter by copy-pasting solutions. Rather than solving it for themselves.
Now, that doesn't necessarily mean they're not capable of doing it, it simply means that they have to spend more time figuring out the problem and/or learning the framework in order to be able to do so. Most people would rather opt out to doing a quick search, copy-pasting, adapting to their specific context, and moving along to the next problem. So, when they have to deal with something like KO, they know what it's going to entail: Hours of learning, or digging into internals and figuring out the quirks of the framework.
It's because we keep trying to use something designed for displaying hypertext documents as platform for interactive applications.
It's dumb and successful.
Think of a typical web app. Your data exists:
1. As rows in a database, accessed via SQL
2. As model objects on the server, accessed via method calls and attributes
3. As JSON, accessed via many HTTP endpoints with a limited set of verbs (GET/PUT/POST/DELETE)
4. As Javascript objects, accessed via (a different set of) method calls and attributes
5. As HTML tags, accessed via the DOM API
6. As pixels, styled by CSS.
--
Each time you translate from one layer to the next, there's a nasty impedance mismatch. This, in turn, attracts "magic": ORMs (DB<->Object); Angular Resources (REST<->JS Object); templating engines (JS Object<->DOM); etc. Each of these translation layers shares two characteristics:
(A) It is "magic": It abuses the semantics of one layer (eg DB model objects) in an attempt to interface with another (eg SQL).
(B) It's a terribly leaky abstraction.
This means that (a) every translation layer is prone to unintuitive failures, and (b) every advanced user of it needs to know enough to build one themselves. So when the impedance mismatch bites you on the ass, some fraction of users are going to flip the table, swear they could do better, and write their own. Which, of course, can't solve the underlying mismatch, and therefore won't be satisfactory...and so the cycle continues.
Of these nasty transitions, 4/5 are associated with the front end, so the front end gets the rap.
(I gave a lightning talk at PyCon two weeks ago, about exactly this - stacking this up against the "Zen of Python" and talking about some of the ways we avoid this in Anvil: https://anvil.works/blog/pycon18-making-the-web-more-pythoni...)