I use Sublime Text every day. I love it. I paid for it. I even develop plugins for it.[1] But I don't recommend it to peers. Why? Because Sublime Text has a low bus factor.[2]
6 years ago, I bought TextMate. Like Sublime Text, I enjoyed using it. Sadly, TextMate became abandonware. The sole developer suffered burnout. Bugs were never fixed. Promised features never showed up. Users wanted to fix these issues themselves. They begged the developer to release the source code, but it never happened. TM 2.0 was open-sourced, but TM 1.5 remains proprietary.
Developers marry their editors. They spend years learning every aspect of them. They customize their configuration. They install and develop plugins and extensions. If an editor becomes abandonware, the users of that editor have wasted an immense amount of time. All of their knowledge and work surrounding their editor becomes useless. They're forced to start over with something else.
All that said, I would recommend Sublime Text to others if the developer made an anti-abandonware pledge. The pledge would be to release the source code if no new version had been released in more than a year. I realize he has to make a living. Heck, I don't mind if he gets filthy rich from selling Sublime Text. I'm simply worried that despite his good intentions, he'll end up hurting his users in the long run. With an anti-abandonware pledge, users would be able to avoid the fate of TextMate users.
1. Sublime Text 2 (and also Textmate 1.5) are STILL good editors. And they will continue to be good editors for the foreseeable future. Just like Vim and Emacs. You don't need a newer version every year. Even though Vim and Emacs DO release newer versions, I have no idea what is in the new releases. Why? Because Emacs 22 and Vim 6.3 do well enough. Textmate 1.5 does well enough. Sublime Text 2 does just fine.
2. I get the bus factor ideal, but it doesn't always turn out that way. Sure it's a chance, but look at the Jack Slocum, ExtJS situation. He was a one man show. He was a machine that could output beautiful amazing work. Then he actually let Sench Labs take over. He seemed to disappear for a time. They had a few stumbles. Now he's back. Same could happen. The same could happen for Sublime.
Overall. I don't see it as a hinderance. Watch Railscasts or Peepcode. They don't seem to have a problem using a 5 year old editor. Heck, I still use a table saw from 1982.
Actually, Emacs 24 was quite an important release that improved many things. For example, it introduced an integrated package control and lexical binding.
If everyone involved with Sublime disappeared tomorrow, Sublime would still work -- and likely continue to be usable for at least a few years.
I don't understand what the concern is here.
As you say, developers marry their editors. I certainly would be frustrated if Sublime were abandoned tomorrow, but I'd still rather use it today over any other editor.
Like we aren't used to this too. Textmate and many other good editors were left eating dust and no one died. If Sublime Text dies I'm sure another will take its place.
I'm also a Sublime Text 2 paying customer who doesn't actively recommend the editor, though I find a lot to like about it.
For me, the reason is a bit more day-to-day practical in that I find myself increasingly doing development on Linux systems running on ARM chips and Sublime Text has no support for ARM builds. It is something often asked for in the forums and on the userecho site but is always shot down by the developer.
Editors are an odd case for me as a developer in that I'd rather use an editor I like slightly less that works on every platform I use in a consistent manner than use one that I like more but has limited platform support.
If it had something like tramp in Emacs, that wouldn't be an issue. Heck, I sometimes log in to remote VPS where I have an Emacs configured to suit all my needs, open Emacs there and edit file on my local machine using Emacs tramp.
There is an excellent sftp sync plugin for sublime text. It's not free but has a similar nag trial option. I don't have the name or link off hand though.
>I use Sublime Text every day. I love it. I paid for it. I even develop plugins for it.[1] But I don't recommend it to peers. Why? Because Sublime Text has a low bus factor.
Life has a "low bus factor".
This bus might hit YOU, after all.
Enjoy what you can, while you can. Who cares if you have to move to something else in 5 years?
Do you feel like the years you used TextMate were wasted? Haven't you wrote stuff with it?
I've seen tons of promising OSS projects becoming abandon-ware when a few core developers left -- or taking a turn and getting rewritten in a way I couldn't stand.
Case in point, TextMate becoming Open Source (isn't that the case, IIRC?) didn't really help it overcome obsoleteness. It got some new builds, and a few guys tinkering, but nothing worthy to use as a 2013 editor.
Tcl seemed to have a lot of the wind taken out of its sails when Ousterhout moved on. The language was - and is - still developed by very competent people, but it seemed to suffer some from the lack of a visible leader.
And it's fabulous, I switched from Sublime Text 2 last week and have no interest in Sublime Text 3 anymore. I was disturbed that Sublime Text 2 had several irritating bugs that I encountered most days and yet all dev was going into the next version that would require a new license. Also I am a sucker for OS X native feel and ST2 clearly had effort put in there, but still fell short.
I use TM2 regularly and really like it for these reasons. My only real gripe, and why I might switch to ST3 at some point, is auto complete. TM technically has code completion[1], but I've never found it particularly useful compared to the popup style that ST and XCode use.
I keep trying the alpha every now and again but there's some little annoyances that to me make it a step backwards in usability from TextMate 1, let alone Sublime Text. e.g.
* Double clicking a directory will make that directory the top level directory in the project side bar.
* Having to double-click to open a file; you can single click but only if you click the small file icon instead.
* Silly fading animation whenever you open a file.
I know most of the time I use the file search anyways but those first impressions really put me off using it.
For me it was the switch from ATSUI to CoreText for the code view rendering. It completely changed the way the text was antialiased, and removed the faux styling for fonts that lacked variants. As someone who liked Monoco, this killed it for me.
At the time TM2 was still private beta and not yet open sourced. I figured if I was gonna change to a CoreText editor, ST2 had the better support.
It would also be nice if my $70 license included free upgrades to future versions. There's just not enough new for me to spend $70 every year for a slightly better text editor. I'll stick with whatever version it is that I have now.
Emacs is old, complex and beautifully configurable. Vim is elegant and fast until you see Vimscript. Sublime is modern and great of out-of-the-box, but developed by a very smart guy that won't talk to his users on the forums).
No. The Python API is being significantly improved in Vim 7.4, but there has been no notion of moving away from Vimscript or even dropping support for the other languages (Ruby, TCL, Lua, MzScheme and Perl (i'm probably missing 1, but meh)).
I don't buy into that. These guys have spent their time and money building something that provides real value. Opening the source will directly compromise their ability to make that money.
Nothing is stopping YOU from building an editor in your own time and making the source available.
Sublime was able to change it from a "buy once, always use" to paid upgrades. This makes it a more sustainable business model. This is something that TextMate never had.
The trouble is, being rich already is a sustainable business model. If you have a tool where the only developer(s) working on it have made enough money that they don't need to continue running it commercially and have no personal interest in building more than they already have, paid upgrades are no guarantee of anything from a user's point of view.
As a counterpoint, as long as the version you already have is useful, and as long as it's a permanent copy that is not going to just stop working if the supplier takes their ball and goes home (I'm looking at you, Adobe), all you're risking is not having improvements that you were never guaranteed anyway. If the current version is better for you than whatever else you could have used instead, it's probably still worth buying it even with no future-proofing, IMHO.
> Developers marry their editors. They spend years learning every aspect of them. They customize their configuration. They install and develop plugins and extensions. If an editor becomes abandonware, the users of that editor have wasted an immense amount of time. All of their knowledge and work surrounding their editor becomes useless. They're forced to start over with something else.
This is like saying you won't marry that beautiful girl because she can die tomorrow and you will not know what to do with her dresses. Isn't more logical to enjoy beautiful girls and pleasant software while they are here?
> This is like saying you won't marry that beautiful girl because she can die tomorrow and you will not know what to do with her dresses. Isn't more logical to enjoy beautiful girls and pleasant software while they are here?
Yeah, because it's fair to make an analogy between a Girl and a Text Editor. A text editor is a tool, I want the hammer to be there for me tomorrow, you can't assure that with a closed source editor (After Text Mate I swore I wasn't going to make the same mistake again, ever.)
Precisely my point: a tool is a tool, you don't marry a hammer and you don't marry your text editor. You can have several hammers in your toolbox, the old hammer will never be jealous from new hammers.
You may invest some time/money on custom accessories for some editor, but if the author disappears it is not like a divorce, you can still use the tool until it breaks: if you can't get spare parts to fix your broken hammer because the maker went out of business, you just go to the hardware store and get a new hammer from other supplier.
>The bus factor is the total number of key developers who would need to be incapacitated (as by getting hit by a bus/truck) to send the project into such disarray that it would not be able to proceed.
So, higher is better ? I was wondering too the other day and settled on "augmenting the bus factor", in a good way.
I don't think that's fair. As others have said, ST2 is a great editor and will be for a while.
I'm probably an aberration, but I don't customize my text editors beyond the basics, and I definitely don't write plugins for them. It would probably make me a better developer, but I'm just not motivated to do that.
You didn't mention what you do recommend to peers?
You're not an aberration; defaults are defaults for a reason. Perhaps around this community it's more common to stray from the vanilla editor but in most cases people will adapt to whatever's given straight to them.
In the case of ST2, it was pretty damn solid out of the box. There's tremendous power in the customization, but it's in no way required.
Precisely. I salivate for Sublime Text, but I restrain myself from even trying because of that. Vim is very powerful and will probably run on my contact lens in 10 years.
I've been married to Photoshop for almost 20 years (which already ties me to either OS X or Windows) and our relationship has recently gone sour financially. Open Source gives you a peace of mind that's very reassuring when it's your daily, almost irreplaceable tool.
Your suggestion is good, I think we haven't yet reached the sweet spot between user freedom and developer paycheck.
I love Sublime Text. For a year or two I tried and tried to get behind vim, to no avail. I thought there must be something wrong with me, perhaps I'm not as smart as those other developers who sit there staring at several adjacent vim windows hitting cryptic key sequences as if the essence of vim were an extension of their mind (maybe it is?). I couldn't do it.
Then I made the realization that the editor you use has no bearing on your skill as a developer (well, as long as you're not using Notepad). I picked up Sublime, and my productivity shot up by an order of magnitude. It's fun, supports every single language I've written in, and has a plugin out there for my favorite color scheme of all time, Tomorrow Night (previously it was Solarized Dark). With sublime I can code circles around my former vim-imprisoned self any day.
Your editor does have a bearing on your skill as a developer. When I interview candidates, the question I ask is "what is your favourite editor?" I don't actually care (much) what the answer is, but I do care that the developer periodically examines the tools they use a lot, and see if they can improve their productivity.
Editors do make a difference in your productivity - that is why people have favourites and why they have flamewars. That productivity matters.
You don't need scientific research to know that a good extensible, programmable editor that can become an extension of your mind and hands is going to make you more productive than Notepad. It's like playing a musical instrument. The finger movements and thought processes move into the background over time as you master them with proper practice and you start operating at a higher level. A world-class musician isn't thinking about which notes to play when and where to put their fingers but rather how to put more emotion into a performance.
Similarly, a top flight programmer who has mastered emacs or vim or Sublime Text won't be operating at a conscious level but will be controlling the environment in front of them at a far higher level, especially if the conceptual model of the text buffer is the same no matter what they're doing with it.
An extreme comparison would be to sit two people of the same ability down in front of the Linux kernel source tree, one armed with emacs, cscope, syntax highlighting, org-mode and cc-mode and the other with nano and see who is faster at some non-trivial task.
You don't need scientific research to know that a good extensible, programmable editor that can become an extension of your mind and hands is going to make you more productive than Notepad.
Sure you don't. You can feel the difference. The problem occurs when people treat their editors as an extension of themselves and can't tell the difference between "not liking you" and "not liking your editor", eventually getting defensive. That's a "cognitive level" I don't want to get myself into.
There will always be flamewars and defensiveness about technology choices. Some people are insecure, some aren't. Not everyone becomes a dick just because they've mastered an editor. My best friend for many years used vim while I used emacs and we used to constantly rib each other about our choices. But we still worked very well together. Maybe you need to work where people don't care what you use to get stuff done as long as it gets done?
It doesn't have to be measured to 7 significant digits in a scientific way. But if you spent a week trying out various editors every now and then for the kind of work you do, you should at least be able to make a more informed choice about which editor to use. Or you may discover functionality in another and see if you can get it in your current favourite.
I've interviewed people who were exposed to Nano at school, and have never tried out any other editor. They are missing out on productivity improvements available in many other editors. More middling of the scale I've also interviewed someone who used Eclipse as their first editor and has proceeded to use it ever since despite not working on Java projects any more. That isn't bad per se, but I rank someone higher if they at least try out others every now and then.
> Editors do make a difference in your productivity - that is why people have favourites and why they have flamewars
The fact that people feel strongly about something does not mean it is important. Lots of people have very strong feelings about the New Jersey Devils or the Miami Heat. That doesn't mean that sports or teams matter.
I'm not saying your conclusion is wrong, but your comment does not constitute a valid argument in favor of that conclusion.
As a vim user, I feel as if the essence of vim is an extension of my mind. Much like riding a bicycle, I instinctively think "select this line and the four lines below it and copy it" as "V4jy" with little cognitive thought. Kind of cool from a cognitive science standpoint.
Don't worry. I find vim completely dated, arcane, and esoteric. Sure, it can do almost 25% of the stuff sublime can do but every feature has it's own learning curve.
If I want to select four lines of text in any text editor that isn't vim, I just hold down the shift key and use any of the cursor navigation available to me. Not so in vim.
Upon leaving vim none of the key bindings are never used in any other program. That's antithetical to the manner in which I operate. Normally if a program doesn't support the odd keybinding or two I just use autohotkey or equivalent. But vim makes that impossible.
I can't comment on the 25% claim, but I've used both extensively and found them comparable.
I personally find the navigation of vim so convenient that I use vimperator for Firefox, as well as the built in vim mode of sublime text.
I totally get that vim is a completely unnatural fit for some people. I certainly don't hold that against them, and I hope they don't hold it against me or vim itself.
> Upon leaving vim none of the key bindings are never used in any other program.
That's actually not true at all. Lots of terminal programs have keybindings for stuff like hjkl. Even tig does. And if you're using git from the command line and haven't heard of tig, you need to look it up.
Even many roguelike games support vim keybindings for movement. Brogue is an example.
> Do you think it's possible some people just aren't wired like this?
Absolutely. Well, maybe I wouldn't call it "wired", but no one tool is right for everyone, and no one should feel bad or wrong for using something different.
I am a Vim lover. You can pry Vim from my cold, dead fingers. But of course it's not for everyone, for the same reason that we don't all use just one programming language or one web framework.
I was in your position for about a year until bit the bullet and uninstalled Sublime Text from my machine forced myself to use vim exclusively. It was an arduous and painful 2 weeks, but it got better and paid off in the long run.
Same here. I'm using Sublime now for larger tasks, as I like the plugins and code completion, but for smaller, more isolated things and remote work vim is still, and will be, my choice.
As an avid vim user, I certainly think it might just be a difference in inclinations. Use what works for you. A text editor is not the objective here, building you project is.
If you are burning all your cycles trying to map the natural way your brain wants to operate to the way your text editor operates, what do you have left for the real work?
If you want to learn vim, by all means go ahead. But feel no shame if it just doesn't fit you.
I've been in and out of vim, but never reached the 'extension' point. Probably due to lack of necessity, and Java work has pushed me further away :/ But mind if I ask why you select instead of e.g. "5yy"? Used to thinking visually, or does it do something different with buffers, or...?
My number one issue with VIM is the need to always count everything: number of lines, number of words, etc. That alone makes the short cuts take a lot longer because generally I'm thinking "I want to cut this function" or "I want to cut down to this line right here". I've never in my life instinctively thought, "I need to cut this line plus four more!".
I end up using visual mode a lot in VIM because it's often faster (or easier) than stopping to count how many times I need to repeat the operation.
I think everyone should find a workflow that works for them, and I've found that what works for me is working almost exclusively from a terminal.
Combining vim and tmux with the basix Unix command line and a few other tools like ack, tig, etc., gives me an environment that I can't really replicate with desktop applications, at least not on OS X, which lacks a true tiling window manager (i.e., one which, unlike slate, spectacle, etc., tiles the windows for you automatically). I've considered a switch to Linux for xmonad, but I still love OS X overall and am more than happy with my terminal workflow.
I work with two 27-inch monitors, and I like to dedicate one entire monitor to a single full-screen terminal with tmux as the window manager.
Admittedly, this has its downsides. For instance, I feel hopeless trying to do serious coding on Windows.
I believe I did try it once but couldn't get it to compile or something. I'm sure it was my fault, but I guess I just don't desire it badly enough over a terminal multiplexer to expend the effort.
Same here. My experience was with Emacs though. I learnt some of the key binding, but installing plugins for almost every need I had (syntax highlighting, debugging, color scheme) just wasn't cutting it. Sublime has very intuitive key bindings, and its blazing fast. Definitely the best editor out there. I still use emacs to edit files through the console, but the majority of my daily editing happens through ST.
I too LOVE sublime text (especially the mini document map, God I love that thing), I just wish there was an addon that turned a tab into a Rich Text Document editor (italic, bold, bullet points, etc...) I tend to write very long posts for my blog and I hate the Wordpress editor.
Oh God you're such a programmer. I'm a designer dude. We're visual people. We don't do terminals, we don't do "sudo apt get". We do clicking, UI, colors, things like that. It's just the way we are. I remember what code looks like (color and shape wise) and where on the page it is, it's not in my head like it is in yours.
I know some excellent developers who write all their code in Notepad. Hell, I wrote a fairly large Java application in Notepad when I was in uni. It wasn't that bad.
I've used Vim for many years and still found Sublime superior after trying it. While I still prefer Vim for quick edits and ssh, Sublime ended being much more productive for coding.
Poor documentation, closed source and an unresponsive developer? Sorry, while the editor has several really nice features, I'm not willing to risk commitment to such a floosy editor.
While on the first glance your comment might seem to be an unfair, snarky dismissal, it's actually a pretty accurate summary of what's making me stay with beta versions instead of buying it.
I can edit the theme file, but it's hit-and-miss guessing game that hasn't yielded any results. I also couldn't find any comprehensive docs for the theme file format. I can't look at the editor source code to see whether any solution is possible, because the source is closed. Likewise, there's no way someone could supply their own patch to the author, for the same reason. So anyone who needs this fixed is at the mercy of the author, who has not responded in any way yet.
Don't get me wrong: Sublime Text 2 has been a delight to use, for the most part. Sublime Text 3 seems to be equally great. I would probably shell out $70 if that would mean I would have better chances to have my pet peeves fixed in some foreseeable future, but I can't see any clear indication of that, either.
That is the bug that sent me back to Emacs too! On a high dpi display the tab fonts are too small and attempting to change the font size just results in this clipping behavior.
Going public with the Beta doesn't bode well for the numerous small bugs and inconsistencies that have plagues sublime for longer than the closed Beta. Some also have been introduced with new features, but still remain unaddressed. It also probably means no new plugin Api functions and features, although many have been requested for a long time.
I guess since ST is such a well-rounded product overall that these shortcomings really grind my gears.
Sublime once had genuinely amazing support. The solution to scaling that with the product's popularity has apparently been to drop it on the floor. Expect bug reports to go unanswered for months.
It seems to have been a victim of its own success. In the run up to the final ST2 release, we saw a string of new releases showing often subtle but useful improvements (and breaking a couple of things irritatingly as well). After the ST2 release, that stream of improvements just seemed to dry up, and now it turns out a paid upgrade model is being used. I don't mind that for new features, but for obvious bugs, in a premium product (which any text editor that costs real money is, given the competition), that got-your-money-now attitude does grate. This remains true even if the amount of money itself is small and it was an easy decision to buy it in the first place (and would be for the upgrade, if I had any confidence it wouldn't go the same way, which I no longer do).
Yes, sometimes I think jps intentionally ignores bug reports that are not in the Technical support forum. ;) that said, ST lacks a coherent stratgey to collect and evaluate bug reports and feature requests. There's a technical support forum, but that can mean anything from "how do I do X" to "Here's a bug" to "I want the API to do this". It needs a bug report flow with clear rules, like you need to add the steps to reproduce, and a status indicator such as acknowleged, wontfix, the option to reference duplicates etc. User Echo kind of works for feature requests, but it seems to be abandonded as well. There's now a community bug tracking project, but jps has been silent on that too. I know it's not easy to handle support, but there's a lot of room for improvement here.
Agreed. I had searched for a couple minutes to make sure I was putting the bug in the right place when I reported it. And even then there was no great way to search that it wasn't a duplicate considering how easy it is for forum posts to collide on a search.
I see new API stuff added all the time in the ST3 alpha update notes.
Plus: what are those "small bugs and inconsistencies"? I've been using ST2 for a year, and ST3 since it was publicly available an an alpha, and can't remember anything other than totally insignificant issues.
I have been using version 3 for a couple months now and I must say, this thing is fast. Starts in less than half a second and the 'Goto Symbol in Project' is an amazing new feature which I now use frequently. Window layout management has also been greatly improved. For me, at the moment this is the best code editor out there.
When you press Ctrl+Shift+R (Windows), it opens a panel similar to Goto Anything. You can then type the function, class or whatever and it will bring you right to it.
Also, when your cursor is on a word, and you press F12, it will bring you to the definition of the class or function. This also works with CSS classes and ID's from your HTML.
It indexes classes and methods (symbols)in your project folders based on the syntax definition of the language in the file. You can then look up and search all symbols in your project, or Go to a symbol where your cursor is. It isn't "smart", so it doesn't know where you use which symbol, or which symbol is assigned to a variable. I haven't used it for mixed projects yet, so I can't say how it works there.
Sadly It is really difficult to make all Linuxes as First Class Citizens, You have to resort to tarballs, Only Ubuntu has gotten to be a true first class citizen.
I've been using ST3 for Ruby/Rails development. I was particularly interested in the "go to definition" functionality. I am a regular user of RubyMine and while code inspection is it's strong suit, the editor never felt responsive. Hence the ST3 trial.
The method jumping functionality is very good. Not on par with RubyMine; e.g you can't jump to an ActiveRecord method or any other 3rd party gem. But for navigating your code base, it reliably finds method definitions across the whole project and find them quickly.
It's not on par with RubyMine, but it's still a solid addition.
You can search through any files that are in your project workspace. When you start a project, just "Add Folder to Project" and include the ruby gems directory and the ActiveRecord source directory.
Will check this out on my chromebook pixel. I recently stopped using ST2 even though I paid for it because there doesnt seem to be a reliable way to increase the font size in the sidebar, tabs and menus without introducing rendering glitches. Back to emacs with terminus as the font for me.
the main reason i am not buying or using it is because he doesn't work off the Userecho page he set up, which in my opinion should be his priority. He adds stuff that he feels is good, but ignores real user's needs. http://sublimetext.userecho.com/
all the feature requests are simply mothballed for years, including some that are showstoppers for me.
As someone who uses sublime everyday let's go through that list...
Syntax / Code folding
- feature for dealing with bad code. Solution: write good code. We don't need additional cruft in the editor slowing it down.
Terminal panel (specifically for windows)
- Would actually be nice for windows users. I am not one. Probably why the console in sublime is so limited, supporting windows in this regard is terribly difficult.
Matching HTML/XML tags
- A package called 'All Autocomplete' already does this.
- surprised this doesn't exist, surprised I never noticed.
Armv7 or armv6 version of sublime text 2 for linux
- I don't even know what to say about this one.
Autocomplete hints
- Sublime is a text editor, not an IDE.
Have an API for Tooltips
- Reasonible request.
Documentation!!!
- Reasonible request.
Icons in the sidebar
- Checkout 'Theme - Flatland' to see how this is done.
Background transparency at the color scheme level
- Couldn't care less.
Move tabs to the title bar like in Google Chrome
- Not sure what this means.
Rename files/folders via Project Sidebar
- Feature added.
Better, or customizable match_brackets style
- Can't say I've looked into this, probably a package exists.
Add intellisense-like support
- Sublime is a text editor, not an IDE.
Double clicking in an empty tab area should create a new tab
- Feature added.
Side Bar API
- Reasonible request.
SCM integration
- Packages exist.
feature is useful
Quickly going through that list, the biggest outstanding issues are documentation and printing. It's not really legitimate to say the developer is ignoring user feedback. It would appear that bad feedback is being ignored and good feedback is being added.
i don't use code-folding either, but blaming the users for writing poor code is a pretty shitty excuse not to implement the most requested feature.
- Matching HTML/XML tags
- Better, or customizable match_brackets style
just because something can be done via an extension, doesn't mean that it should be done that way. The most popular extensions, i think, belong in the core.
Whether you think any of these points are valid is your personal opinion, against hundreds/thousands of others. If he didnt want to take user feedback seriously, then why bother with a prominent feature tracking system?
EDIT: looks like Matching HTML/XML tags is implemented in the new beta (with some quirks). Unfortunately, the fact that the highlight style cannot be customized makes it just as useless as the existing, barely-noticeable match_brackets feature. Is it really that difficult to let users customize the editor's marking style?
I was hesitant to add that comment, but it needs to be said. Pushing people into the pit of success is a good thing. Sucks to read 10,000 line files? Stop writing 10,000 line files. I also really like how fast sublime is. Coming from an IDE for about 6 years, this is one of the features I like the most. Code folding is bloat.
>Whether you think any of these points are valid is your personal opinion
I didn't post this under any other assumption. These are all my opinion, as a full time sublime user.
>If he didnt want to take user feedback seriously, then why bother with a prominent feature tracking system?
I would argue that he is taking user feedback seriously. One of the most difficult things to do with any product is to decide what feedback is valuable and what feedback is not. I think sublime is doing a very good job of this.
Pushing people into the pit of success is a good thing.
I'm not sure history supports that position, at least in the context of software development.
For example, in C++ the syntax for the new-style type cast operators was made deliberately awkward. The argument was that casting is generally not to be encouraged, so doing it should be ugly. What this actually did was encourage almost everyone to ignore the new-style casts and just write the old C-style ones instead, because they worked and didn't make your lines 157 characters long.
If you make software development tools, fighting against what software developers want to do is almost never a winning strategy IME. You promote better practices by providing better tools that make those practices easier, not by providing worse tools that make some things harder and hoping that will stop people doing those things instead of ignoring your inferior tools.
Re folding: what about inherited codebases, via github or your company or whatever? That seems to me to be the vast majority of source code being read / edited in the world, and odds are you can't simply change it on a whim.
I think that pain is a great motivator to refactor the code base. I understand that pain, I have been there. But coming from an IDE that supported code folding, I can say two things about it. First, it only helps marginally. Second, it makes occurrences of large files even more prevalent.
> Move tabs to the title bar like in Google Chrome
What it means is like chrome there should be no title bar. Precious vertical pixels wasted for something that is never used. Like chrome or most modern browsers have moved the tabs in the title bar sublime should do this too.
I've tried to give sublime a shot many times -- and I really would like to use it. It does have great plugins and features, however the deal breaker has always been that I needed to use the mouse at some point in my workflow. I have to admit that the "vintage" mode doesn't completely emulate vim, which is really a shame.
I've said the same thing every time I've tried to use Sublime or any other modern editor or IDE. My specific beef is the inability to manipulate splits using a mouse. It's pretty much a deal breaker for me as I find splits far too useful to give up now.
Block or multi selection, though I don't use it as often is another thing that is often poorly done or requires a mouse.
Try ST2 with a ton of plugins on an underpowered netbook one day ;) ...it wasn't "slow" like a big fat IDE can get, when you end up typing faster then text can show up on the screen, but it can still get "slow for an editor", like ">1s startup time", ">200ms to open a new tab" (yes, it's noticeable damn it, anything above 10ms if fking noticeable!) if some plugin tries to lint or spell check in the bg, etc.
I'm really glad to hear this since as far as I could diagnose it, sublimelint was probably the biggest contributor to my ST's slowness :) ...but it's a great plugin to have, so kudos and keep up the good work!
I don't run many plugins with ST2, so I haven't noticed much issue.
That said, I feel odd about this being a primary upgrade point. Speed improvements to me indicate that there were speed issues (confirmed by others here). I read that as "bug," not "feature request." It almost feels like you're going to be paying $15/$30 for a patch in this instance.
Well, for example opening a "small" 20MB SQL file could lock my ST2 for minutes on my slow 8 core i7 with 8GB RAM. For comparizon, Vim and Emacs could open 500MB SQL file in the same amount of time. And I'm not talking about saving, just opening.
I was editing JSHint - most of which is in a single 5000 line file - yesterday, with the same file open in split view, and was disappointed to see ST2 chugging a bit, with visible input lag.
If you have a somewhat recent computer (Core 2 Duo or better?) it was probably caused by one of your plugins. It's pretty easy for a plugin to block the UI in ST2.
If you listen for on_modified as a plugin and do any work, the next keystroke won't appear until you return from on_modified. Multiply that across many plugins, some of which don't care about speed, and your editor becomes very slow.
It's simple, Sublime 2 starts after 2 or 3 seconds, sublime 3 starts in less than a half second, 3 seconds don't seem like a lot, but could you imagine every program you have starting in less than a second? It would be a really different experience.
I started using Linux on a machine with a 386 and 4 megs (yes, megs) of ram. Emacs was, I must admit, a bit slow on it, so I used jed until I got a better computer.
jEdit was basically my first introduction to powerful text editors while learning programming and Lilypond. It had some nice plugins/macros for text processing. But, yeah, it was so slow that eventually I moved to Notepad++ for basic text viewing/editing. Recently I decided to give ST2 a try and am very pleased with its speed and features. Right now I'm trying unlearn a few Notepad++ shortcuts that I occasionally misapply to ST2.
Having jEdit load on startup helps a bit with the slowness.
First thing I tried out was - does it close <div/>. Yes, it does: "<div/></div>". I'm really happy with lots of sublime text features. But this simple untested completion error stunned me.
Sublime is a great product with shitty community and plugins, unfortunately. It could be so great but there is no feedback, no word from Jon regarding the communities requests.
I bought and use ST2 daily, it's really great and I've been an emacs user for eons. Only other editor that I use now apart from ST2 is pycharm (and not even that much). Only thing I haven't figured out yet how to have multiple syntax highlighting in ST2, for example PHP within HTML or vice versa when I sift through old code, or JS in HTML. I know there must be a solution, but I haven't really looked into it.
Fairly good. I've been using ST3 as my primary editor for the past month or so, and I've mostly had good results. Many plugins have a python3 branch on their github, or just work outright.
It's worth checking if any essential ones will work for you before switching.
Personally, my biggest issue is that wbond's Sublime FTP doesn't work, but the overall speed increase is so great that I'm trying to make my workflow work without it.
Glad for the speed improvements but the UI seems to be getting worse. For example, the minimap now only shows your viewport view on hover making it mostly useless as an orientation tool. I emailed the developer about this issue many weeks ago and never received a reply, so I have no idea if he also views this as a problem (not encouraging).
How come on OS X the theme colors are considerably darker in sublime than textmate? Something to do with gamma settings or color profiles? For me solarized dark is almost unusable in st, in textmate it is not as dim (solarized light in st has inverted selection, which annoys the hell out of me).
Check whether the theme definitions actually contain the same colour values. There are a lot of unversioned XML files floating around the TM/ST universe.
I'm not a Sublime user yet, but I've always been interested in it's interface. I'm trying to figure out if it's a replacement for a text editor like VIM or a heavier IDE like IntelliJ?
It's a replacement for VIM so good that in most cases the IDE becomes unnecessary, because the notion of 'project' and various GoTo and search commands solve all navigation needs. It even has a rudimentary build system support.
I talk about c++ and Python though, where no 'real cool' IDE exists. Not so for statically-typed reflexive languages like Java and C#, where your well-informed IDE can make code transformations of arbitrary complexity.
Stino does a really nice job of integrating Arduino into Sublime. The default IDE for Arduino is horrifying; if you do much more than copy and paste a few sketchbook examples, you'll probably be looking for a way to get Arduino development working in your favorite text editor right away.
In Sublime's case, any time I'm working on a ".ino" file, I get an "Arduino" menu added to the menubar, which gives me all of the options needed for compiling & uploading code to the Arduino &etc.
I absolutely love ST2, Just downloaded the ST3 beta and it is significantly faster then version 2. I'll keep 2 though while I see how stable 3 is in day to day use.
I bought it even though emacs is still my primary editor. $70 is a drop in the bucket. I keep all of my notes open in a couple different screens (grouped by project). It's really quick to navigate to buffers (Cmd-t) and search all buffers. I'm finally learning a little Python because that's how plugins are written. Sublime has pretty smart completion too.
Anyone have an org mode? I'd at least like to make urls clickable.
The biggest bug for me is that it's $70 - Although I'm a professional developer, I'm still fairly underpaid (I don't live in London, San Fran, NY, etc.) and thus it's out of my price range when I can barely afford travel, rent and food.
I use vim and reasonably happy with it, but honestly 70$ is little more than a new console game, or a dinner for two in a decent restaurant... we are not really talking about a license for Photoshop or professional CAD software.
If you are a professional developer living in a reasonably modern country and cannot afford 70$ for what is supposed to be to be your primary tool, then you are not underpaid, you are being scammed.
That the features it offers are not worth 70$ compared to free editors is another matter entirely, and a purely personal matter.
I bought a license for TM years ago, and currently use TM2. This would be a "nice to have", rather than a must have. I should probably save up and take my girlfriend out soon. Thanks for reminding me.
Still, I don't think it's that much money. I bought TextMate, then moved to phpStorm (this one was almost 100€, they had a sale not too long ago) and I also got Coda 2.
I buy a lot of software, if a tool is useful it's going to pay for itself in no time.
Not sure what your point is, apart from not being able to afford it.
Then again, not sure why, if you're really employed as a developer, your company isn't paying for your development tools. DId you have to buy your own laptop?
Being underpaid isn't a function of being in a particular region (assuming your in America). Middle America Jr. Devs with little/no programming experience are getting started around $70K.
Software developers in the US tend to be paid much more than anywhere else in the world, and often don't realise this.
In the UK, a salaried senior software developer would typically make £45-60K before tax (currently around $70-90K US). Working for someone else as an employee, you'd expect to reach that level after maybe 5-8 years, but then stick there for the rest of your career. Obviously the rates can be higher with special circumstances and they are somewhat higher if you're working in London, but anything close to what is normal in the US would be a rare exception in the UK (and most other places outside the US).
There has been a 25% fall in the GBPUSD exchange rate over the past few years which needs to be factored in when comparing salaries. The Bay Area can make London seem an affordable place to live.
But that being said, there is still a significant difference in wages for salaried positions. Contracting rates seems much more similar. And the tax treatment for contractors in the UK is so favourable as to make it more worthwhile than all but the very highest salaried developer roles in the City.
And the tax treatment for contractors in the UK is so favourable as to make it more worthwhile than all but the very highest salaried developer roles in the City.
IME as someone who does freelance work in the UK, the tax treatment usually isn't a big win. There are some tax advantages, but in practice they are usually balanced or significantly outweighed by the extra overheads of running a business (accounting, legal, insurance, equipment purchases, etc.), the loss of typical employee benefits (paid time off, pension contributions, etc.), and the risk of downtime when you don't have work. If these kinds of downsides don't apply to you, then you're likely to be treated as a disguised employee and the IR35 rules will negate the tax advantages as well.
I find the big difference in moving to contract/freelance work in the UK is that you change the frame of reference from employer-employee negotiations and their associated expectations to business-business negotiations and their very different associated expectations. If you're good, and therefore generate a lot of obvious value to clients, the latter removes the glass ceiling. I suspect not coincidentally, as you mentioned, rates for this kind of work seem to be much more consistent among different countries.
As another person mentioned, being in the UK you face different challenges. My reply was partially to help you, but also to let other American programers know that there are better options out there.
To give you the tools to do your job? I think perhaps you should start looking for a position elsewhere. Bristol does have other employment opportunities.
It's a developer tool. I bill more than that per hour. Lots of people bill more than that per hour. If it saves you time or stress, it's easily worth $70.
The potential problem is not so much the price, which is as you say very reasonable, but getting locked into a proprietary product that you are then attached to. What if they stop developing it? What if they take it in a direction you don't want? What if they move to a higher and higher price? All things to consider.
As for me, you can pry my Emacs from my cold, dead fingers.
The answer to all of the above is: 'Don't upgrade'. As it stands ST is an excellent editor with a rich plugin system around it, even if the dev decides to call it a day it's unlikely it's going to stop working in the next 2-5 years, and additional features can be introduced via the plugin system.
I do understand the concern though, but for a lot of people (myself included) ST is a better solution than a non-proprietary solution.
A waste to who? It's the editor of choice for a large number of people. It provides value to developers. In turn, the people who spend all day building it (I assume from a comment above) get paid to do so. Do you want to work for free?
I may be in the minority, but I don't think all code should be open. The people behind Sublime have made a great product and continue to improve it. Do you think they'd receive even half of their income if they made the code available?
Step 1. Remove nag message.
Step 2. Release to the world.
Day 3. Tonnes of people jump to non-nag version and never pay.
The plugin system and API is available for those that wish to complement or extend the editor. Making the code available directly harms their ability to make money and their ability to continue to invest time and money building the product.
If you are a developer billing that much per hour, then it is undoubtedly worth it. For hobbyist programmers and people just getting started, however, it's a bit more uncertain. It's the age old question of how to price a product that is of great value to professionals, but also useful for novices (think Photoshop and Excel). I think that ST2's approach of being hassleware that doesn't go overboard is a good system.
It's the same editor whether you pay to use it or not. The tiny hassle at the beginning of clicking "OK" is just to remind you to finally pay for it when you move from hobbyist to professional. :)
And if you work in a big company, who knows! They might already have a group license you can you.
My problem is that I am invested in vim to the extent that I will need to relearn quite a bit if I give up vim. I can do that but what does ST provide me over vim - that's my main point of debate along with the fact that the $70 price for one version with paid upgrades means things are not in the impulse purchase zone over the long run. (Not upgrading doesn't really work - ST2 might stop working if I upgrade to say Ubuntu 14.04 and I am not sure if ST2 will continued to be supported and how long.)
I would have bought it if it was $29 and proprietary OR $70 opensource donation ware or $49 proprietary with 3 free version upgrades. That's my comfort feel around editors - YFMV of course.
You are right, but the problem is exactly that it's a developer tool, and most developers tools are $free these days, so it kinda sets the expectation of price for a developers tool. If Emacs, Vim, Eclipse, Visual Studio (Express) and many, many others IDEs and editors a free, it's pretty hard to justify paying for a developers tools. It's not that it's not worth it, it's just that the common expectation is that developers tools are free.
Sublime Text and Textmate likely wouldn't have existed without many people being willing to pay for them. If we refuse to accept paid for developer tools we are only ever going to get certain types of tools build by certain types of people/ companies.
It does in Australia. It's a work expense. Just because there are other options that are free it doesn't mean it's not deductible. That'd be like telling a carpenter they have to use a tree branch as a hammer. But I'm no tax expert in your region, so it might be worth investigating if you're not certain.
The thing is I need to do Itemized Deductions if I want to tax deduct that expense here in the US. The problem is that it many a times comes out less than the standard deduction.
I think a lot of it depends on what you do and what you use. I have it primarily for quick C and Go projects because I find the built-in compiler feedback to be very well done.
If were purely a Python, PHP or otherwise Web developer, I haven't found a whole lot in ST2 that you can't get in Notepad++ (even without additional plugins).
The great news here is you can test drive Sublime Text to definitively decide if the price is right. For me, it was, but I would also caution against jumping on it just because a lot of people do/have. There are developers who wouldn't necessarily benefit from this over the alternative(s).
I'm in the same boat. I've used vim for a while now and I'm faster at a couple of small things. So much so that Sublime feels slower to me. (Having to take my hand off the mouse to do a two handed key combo. Or having to move my hands too far from the home keys to do a key combo.)
I spent some time last year using Sublime Text 2. [1] But I didn't get very far before I switched back to vim. My guess is that I'll have to unlearn my vim-ism or go all-in on vim and spend time practicing the more esoteric commands until they become second nature.
Good. Vim is a great, extremely powerful tool. One of the things that I really couldn't stand about Sublime was that so many users were trying to make it more like Vim. I don't want to use Vim for my daily work. I want to use Sublime. Vim is free, if I wanted to use it for anything more that modifying conf files, I'd use it. But please, don't turn Sublime into Vim... we already have a GREAT Vim.
I agree. I do have a Sublime Text 2 license, but I miss modeful editing, ex, etc. I had used vintage mode for a while, but as usual with such emulations, you quickly run into annoyances.
The only thing that I found really attractive in Sublime Text 2 is Goto Anything. ctrlp.vim brings some of its advantages to vim.
You don't really need multiple cursors in vim. Visual mode is much more powerful than dropping a bunch of cursors at the right places in lines or blocks of text, and if that's not enough, :s commands handle pretty much every usecase. If you're _still_ unsatisfied and want some corner case that's best satisfied by multiple cursors, then you can use the plugin. I installed it thinking it might be useful, but I still haven't found a use case where I prefer multiple cursors to visual mode or ex commands.
Yes, considering it has one of the least annoying trial versions ever (I believe it pops up a message every 50? or so saves, and otherwise is exactly like the paid version)
"When 3.0 is released, upgrades will be available for $30, or $15 for users who have purchased recently. The cut off date for the reduced upgrade price will be based on the actual date of the 3.0 release."
Thanks, it's worth full price - but I'm finding more and more upgrades that are offering very little concession to users that have supported them from the start.
I'm a bit ashamed.. I still use sublime2 with the dialog popping here and then asking me to pay for it :-/ The only reason I don't pay to upgrade is simply because my current setup is working perfectly.. i.e. Why would you pay 100$ for more trouble? (Trouble here is re-installing everything, re-configuring the binding, etc etc.)
6 years ago, I bought TextMate. Like Sublime Text, I enjoyed using it. Sadly, TextMate became abandonware. The sole developer suffered burnout. Bugs were never fixed. Promised features never showed up. Users wanted to fix these issues themselves. They begged the developer to release the source code, but it never happened. TM 2.0 was open-sourced, but TM 1.5 remains proprietary.
Developers marry their editors. They spend years learning every aspect of them. They customize their configuration. They install and develop plugins and extensions. If an editor becomes abandonware, the users of that editor have wasted an immense amount of time. All of their knowledge and work surrounding their editor becomes useless. They're forced to start over with something else.
All that said, I would recommend Sublime Text to others if the developer made an anti-abandonware pledge. The pledge would be to release the source code if no new version had been released in more than a year. I realize he has to make a living. Heck, I don't mind if he gets filthy rich from selling Sublime Text. I'm simply worried that despite his good intentions, he'll end up hurting his users in the long run. With an anti-abandonware pledge, users would be able to avoid the fate of TextMate users.
1. https://github.com/Floobits/floobits-sublime
2. http://en.wikipedia.org/wiki/Bus_factor