I feel like I'm perpetually stuck between what the author describes as "beginner" and "early". I understand what programming is, I can write a bash script that does what I want it to (granted, I have to read a ton of man pages to make sure I understand what it is I want to accomplish), I can write simple programs in Visual Basic or Python or Javascript that do simple tasks. I understand program flow, logic, and all the basics of high-school level algebra.
The problem is, I can't wrap my head around many of the concepts I read about here in the HN comments and elsewhere on programming blogs and such. No matter how much I try to understand it (and by understand it, I mean fully grasp what the person is talking about without having to look up every other word or phrase), I can't seem to put it all together. Things like inverted trees, functional programming (I've heard of Haskell and I'd love to learn it, but I have no head for mathematics at that level), polymorphism, and so on.
Maybe I need to just practice more; maybe I need to pick something interesting from Github and dive into the code to try to understand it better (preferably something well documented of course). Or maybe I need to just stop, and accept that I can whip out a script or simple web thingy if I really need to, and stick to being a hardware guy, which I'm actually good at.
You don't have to understand everything. I've been programming for 30 years and it took me 3 years to understand monads. (At least I think I understand them /grin.) I also have no idea what "inverted trees" mean - if you're talking about the recent thread about the guy who got rejected by Google, I believe it just meant trees with the left and right nodes reversed.
...but it makes you feel stupid if you don't. Better to use "Real World OCaml" if you're more interested in the ideas themselves than in their formalizations or related nomenclature.
Exactly. That's where I'm at right now; I know what Haskell is, I love the idea of it, I've enjoyed some of the fruits of it (XMonad). But it was when I tried to dig deeper into it that I felt lost, and yes, stupid. I've never been a math whiz; I am great at visualizing concepts but truly grasping the theory behind them is where I get lost. Based on my junior high school testing, I was placed in Advanced Algebra in my first year of high school. I nearly failed the class because it took me all year to grok the distributive property. I look back on that and I feel ashamed, because once I understood it, it seemed so damn simple! And so it is when I try to advance beyond my current level of programming skill; I hit brick walls and I feel like I left my sledgehammer at home. My pocketknife, even though I know every millimeter of it, won't cut through those walls.
I don't think the concepts are hard to understand, I think that - in Haskell - they're just being presented in a way that is incompatible with my way of thinking.
Having found Haskell materials as simply not suited for me I decided - quite a few years back - to learn Haskell (or the concepts behind Haskell, at least) my own way: by learning first Erlang (it sounded cool), then Scheme (mainly to be able to read many, many papers that use it), then OCaml and Scala (because the type systems and pragmatism) and finally Clean (to fill the last gaps in my knowledge). I progressed from dynamic to static typing and from eager to non-eager evaluation. It took me I think about 2 years to do all this and, of course, it wasn't that easy, but somewhat surprisingly it worked. I never wrote - and I'm not sure I ever will, but that's a completely different matter - any non-trivial Haskell code, yet I'm able to read and enjoy Haskell-related papers.
It's important to realise that there is always more than one way to learn things. You should know yourself well enough to see when the "normal" way simply isn't for you; this way you can go search for alternative ways. I guarantee that you'll find them, if you search hard enough :)
I'd put you at early. You're ready to learn more. Don't pick something interesting up from Github. Reading the source code from a complete project is tough as hell for an experienced programmer.
The stuff you don't know all needs to be studied. People don't happen on functional programming or build trees a propos of nothing; they studied it, either in college or on their own.
Find a good book and dig into it. If you want to learn stuff like functional programming, read The Little Schemer. If you want to learn about trees and other data structures and algorithms, make a few pots of coffee and work your way through the Algorithms bible by CLRS.
Polymorphism is actually pretty easy once you understand the ins and outs of OOP. It just sounds scary. I've taught a first-year high school beginner programming class about polymorphism by the end of the year.
Thanks. I really just pulled those concepts out at random, things that I'm interested in but I know I'm not ready to tackle yet.
I did study programming in college (well, tech school, so that's likely part of my problem), but it was SQL/400 on IBM AS/400 machines and was geared towards direct employment at one specific company. In other words, it bored me to death, and I switched to web design halfway through. I really didn't learn anything there that I hadn't already taught myself, and I was surprised to find that I was the only person in the class (including the instructor!) who knew what Linux was. We had an AIX server that had been donated by some company a few years before, and it was sitting there unused, like a monolith from another era, until the instructor allowed me to work on it in my spare time. I managed to get it up and running, and set it up as a local webserver so we could practice server-side scripting with more control over the environment, and learn a bit more about how web servers do what they do.
And that's the core of it I think; my passion is for tinkering and fixing things, and while there's a lot of that kind of thing in the programming world at large, I'm really more of a hands-on, direct kind of person rather than an abstract thinker. My sister is a website and graphic designer in her spare time, and she and I have often talked about starting our own hosting service geared towards creative professionals, with me handling the back end and her doing the front end and marketing. Maybe I should pursue that instead of wasting time trying to learn advanced programming concepts when I'm not actually seeking employment as a programmer.
Or maybe you need to set aside the shell scripting and start learning a different language? "Things like inverted trees, functional programming, [and] polymorphism" really have no analogue in a shell language. Start working in C++ or Java, for example, and I suspect these things will start making more sense.
Pick a medium sized project, and sketch out all the steps and sub-goals (and update/redo this sketch as you make progress and understand what the final implementation may look like). Then slowly conquer those sub goals one by one, googling furiously and reading what you find very carefully the whole way.
You can decrease/increase the scope of the project part way through depending on how smoothly things are progressing. Fully executing on a simplified initial version is much better than half way executing on a grander vision.
Don't be afraid to ask for help if you get stuck. You will be amazed at everything that you have learned by the end.
The problem is, I can't wrap my head around many of the concepts I read about here in the HN comments and elsewhere on programming blogs and such. No matter how much I try to understand it (and by understand it, I mean fully grasp what the person is talking about without having to look up every other word or phrase), I can't seem to put it all together. Things like inverted trees, functional programming (I've heard of Haskell and I'd love to learn it, but I have no head for mathematics at that level), polymorphism, and so on.
Maybe I need to just practice more; maybe I need to pick something interesting from Github and dive into the code to try to understand it better (preferably something well documented of course). Or maybe I need to just stop, and accept that I can whip out a script or simple web thingy if I really need to, and stick to being a hardware guy, which I'm actually good at.