Hacker News Reader

Comment by batista | Hacker News Reader

If we are completely honest, Textmate was always a sub-par editor.

No Vim or Emacs style brilliantness, no BBEdit style tons of features and mature engine, no IntelliJ like, er, intelligence, no ST2 comprehensiveness, etc etc.

Plus, the Textmate 1.x text engine was probably a mess too -- I remember the very first versions being laggy (and that's coming from someone who doesn't find even Eclipse laggy). That he couldn't easily fix the one-character-undo is another pointer to that (and, for all I've seen, the 2.x engine is not that better).

It's main saving grace was the many extensions it had, and looking half-decent and native on OS X. Basically, it caught on because it appeared on the right time, and appealed to OS X users like web programmers etc, that wasn't old-time unix buffs, and wanted something native looking without forking for BBEdit (which itself was/is Carbon based and with a custom text display widget).

I don't think Textmate deserved all that success --it should have happened to a better editor.

View on HN

I disagree with your assertion. TM 1.x was an amazing piece of software and way better than numerous text editors out there at the time.

Usability wise it was impressive, keyboard shortcuts made sense, creating your own snippets and extensions was a breeze, block mode was simpler than on any other text editor out there and there was support for most programming languages out of the box.

Consider for a moment that Gedit and Notepad++ are used professionally even today by many devs.

While I'm a big fan of Vim today and wouldn't go back I think TM 1 was a step up for me as it was for numerous other developers at the time. I went from using Gedit to Textmate and I was blown away by the feature set.

But gedit is pretty impressive, still! It's extendable too, maybe just not to the extent of Textmate.

> gedit is pretty impressive, still!

Yes, and it feels completely alien on OS X.

I think you're either forgetting, or never directly experienced, how much nicer TextMate was (and is) for dealing with files that mix different languages. PHP/HTML, Javascript/HTML, CSS/HTML, etc. TextMate automatically detects the language shifts and modifies syntax highlighting and snippet shortcuts. This was huge. No more jumping from PHP mode to HTML mode to JavaScript mode: it just worked.

Yes, the mixed mode editing, tabs and the great snippets were killer features for me to switch from 10 years emacs to TM1.

Yah, same here. I still have the muscle memory for emacs and I've gone back a few times and tried it, but honestly, the configuration was such a pain in the ass. I have .emacs files from long ago, but I tried a few months back bringing it up to date along the lines of TM1 and after about 15 mins, I just said forget it.

TM1 out of the box had all my web dev needs already there. Plus, tabs worked great. The only things I've replaced on it over the past few years were search in project (via grep in project) and the file drawer on the left (forget what that plugin was called). Speaking of plugins / bundles - those were super easy to install. For what I needed, they pretty much just always worked.

For the OP who said that Textmate didn't deserve its success I say, whatever. Everyone that I worked with over the past 5+ years doing web dev used TM. And NO ONE complained that it got in their way of getting their work done. So, cool, that's your opinion, but I think TM deserved all the success it got - it's helped me make a good amount of money over the past few years and I still use it today.

The fact that TM2 is now on Github, that's great. Will anything come of it? Dunno, but I guess I can't blame Alan for putting it on there and maybe moving on. TM1 is almost 8 years old and that's a long time to be beholden to something. Oh, and I never felt ripped off; it was money well spent and has lasted a long time..

As an aside, I believe there's a textmate package for emacs.

http://code.google.com/p/emacs-textmate/

No idea how well it works, but I suppose the only major difference in some respects is having to use Ctrl/Alt/Esc instead of Command.

More saving graces for TextMate 1 even today: It is right at home on OS X and also acknowledges the usefulness of UNIX. / It has no learning curve. / It requires zero configuration to be perfectly usable. / The shortcuts are as intuitive as it gets. / The only shortcut that I found weird was "cmd+T", and that one has been widely copied.

If you say that another editor should have been as successful, then what do you mean? That TextMate users have never heard of vim, emacs or ST2? That developers have been lazy and should have out-done TextMate (which ST2 arguably did)? I think this as clear of a deserved success as it gets - it's a programmer's editor and people decided it was the best one to program in.

I agree with you insofar that I tried it and didn't quite get the appeal personally. But I had alot of experience with VIM coming in.

The degree of success is something I allude to as being congruent with the success of Apple products in general - well designed, cohesive, but definitely not the core demographic being power users. If anything you could probably say it was many's first powerful editor... and as people people grew in their abilities lament formed from its limitations, much of which was to be addressed in TM2, but the extended wait opened the door for people to move onto other, more powerful editors, and esp. set the stage for Sublime - which I'm neither here nor there with and may go back to VIM. But hey, its plugins are in Python and it has a VIM-like mode with an elegant UI so it's definitely a solid step in the right direction.

I think its success was well deserved given its core demographic and esp. tie-in with the Rails community which seemed to grow with it. I'm glad it existed, I consider it a great 'gateway' editor :) And now Sublime is its spiritual successor and I consider that a pretty good thing. Some people, esp. web designers, are just not going to give up UI polish no matter what so it is what it is. And it may very well have introduced others to VIM or EMACS who wouldn't have given them a look initially, finding the old school look and complexity daunting at first blush.

What drew me to TM1 were the several innovative features for the time (mid 2000s) that were promptly copied into other editors (and sometimes improved there):

  * snippets
  * fuzzy file finder
  * project file list
  * multi-column editing

This is in addition to the non-text related good things TM did: grouping languages, snippets and scripts into a single folder, the bundle community, etc.

However, those 4 features were what moved me over to TM - I was more than willing to plunk down my $50 for these features that no other editor (at the time) had

The fact that the file list correlated to the contents of a directory was a huge deal. XCode and other editors require "projects" to be assembled. TextMate could also ignore certain paths if configured correctly, cleaning up projects considerably.

I used SubEthaEdit prior to TextMate and was quite happy, but it lacked multi-file editing and tabs.

"brilliantness"? Are you kidding? Have you worked with vimscript? Oh well, at the risk of feeding a troll...

Textmate introduced a huge number of features that now show up in other editors. Snippets, command-t, a tightly integrated community-driven extension system.

Textmate was a lightweight editor for lightweight dynamically typed languages. It was the perfect contrast to Eclipse and Java, and was instrumental in changing the status quo for web development (for the better, imho).

>"brilliantness"? Are you kidding? Have you worked with vimscript? Oh well, at the risk of feeding a troll...

No, I haven't worked with vimscript. What does it have to do with anything?

Vim/vi is a brilliant piece of editor software, with a solid concept (an editing language, the dual modes etc). And it is in continuous use for 30+ years now. Does anybody disagree with that?

Now,vimscript being crap, or even it being totally absent, would not change those facts. In fact, most of the decades of vi/m using by hundreds of thousands of programmers have been without vimscript existing at all.

Btw, please don't use the word "troll" out of context. I made an argument. Reply or ignore it, but don't use conversation killing words like "troll".

vimscript/viml, which is the basis of extending vim, regularly shows up as one of the worst languages to work with. http://geeksta.net/geeklog/exploring-expressions-emotions-gi...

vim is great, but most of its great features are at the end of a very difficult (but necessary) learning curve. The learning curve for developing plugins is even steeper (and unnecessarily so).

In contrast, most people can pick up and use textmate instantly. They can alter the extensions easily, since everything is done in a sane script environment (python/ruby). They can also share/alter/manage their extensions more easily without causing conflicts. Pathogen/vundle are recent vim plugins that finally make this easier for vimmers.

In a sense, textmate changed the nature of what an editor was about. I think extensions are a large part of an editor's value now. ST2, vim, and others are now following textmate's approach, and I admit, they're doing a better job of it than textmate. However, you still have to give credit to the originator.

> In a sense, textmate changed the nature of what an editor was about. I think extensions are a large part of an editor's value now.

You've never used Emacs, have you?

> You've never used Emacs, have you?

I'm talking about light weight editors that support light weight scripting in a number of different languages. I consider emacs to be more IDE-like. That's admittedly a weak distinction though.

emacsclient is about as lightweight as it gets, and Cocoa Emacs itself (24.1.50.1, built from the Git repository on Savannah this afternoon) starts up in less than 5 seconds on my Mac. As for "scripting in a number of different languages", that's more of an anti-feature for me. Specifically, running Emacs in a Windows, FreeBSD, Linux, Solaris, ... VM in VMware Fusion, I can symlink the init files and site-lisp directories in my guest OS home directory to the analogous items in my OS X home directory, and, given reasonably comparable fonts and Emacs versions, my Emacs configuration and most extensions work identically in the VMs.

If, on the other hand, these extensions were written in a number of scripting languages rather than Emacs Lisp, I'd need to maintain a like number of scripting language runtimes in each guest OS. Given the amount of code I write in Emacs pales in comparison to the amount of contributed third-party code I use, the fact that I have to write it in Emacs Lisp is a vanishingly small price to pay. Besides that, for lightweight scripting at least, I actually like Emacs Lisp.

I agree that emacs is losing its stigma of being/feeling bloated. Also, lisp is finding more practical applications through clojure. It seems like emacs should be on the upswing. So, charts like this (taken with a small grain of salt) are puzzling: http://www.google.com/insights/search/#q=emacs&cmpt=q

Vim had "extensions" before TextMate even existed.

I don't speak Ruby so I've never been able to alter much of anything in TextMate beside creating/customizing snippets in my 4 years of heavy usage (user ID in the lower 4 digits).

However, picking up vimscript for my small needs was relatively easy.

It's main saving grace was the many extensions it had, and looking half-decent and native on OS X.

One of the key features it had in the early days, that most other OS X editors did not, was a file tree drawer that worked well. From the terminal, you could type "mate ." and you'd have an editor with a file drawer right there. No "project" BS, no modal dialogs to navigate around files.. all very handy for working on file heavy Rails projects.

Exactly. Window is folder and tabs are files, that arrangement feels completely natural, and Vim never got there with all its own buffers and windows and tabs and so on. Nothing else ever got there (on OSX) until Sublime.

Yes, the extensions are its saving graces. I'd use it for typing up a web page in HTML, but as soon as I try to type in Japanese characters, it all goes to hell.

No, seriously. Typing Japanese characters in TextMate is a worse experience than, say, typing them into a terminal window, or typing them into an aged Carbon application that was obviously not designed for Japanese input.

I've been a TextMate user since... well as long as I can remember (I was so glad to be done with all those BBEdit trials– seemed liked every time a new MacWorld magazine was out it would come with the latest and greatest... BBEdit). But, over the last year or two, this is the one problem that has caused me to revert multiple times to other editors. I was hoping TM2 would improve on this a bit but so far nothing's happened.

It won so many adopters because it was simple and pretty. The advanced features are great, but TM won because it didn't have Emacs/Vim learning curve, and it mostly just worked.

VIM learning curve.. It takes 2 hours (at least it did for me 8 years ago) to go through vimtutor, which means you can do all the basic stuff like with any editor. That's it. From that point you can stay or go how ever far you want. From using the editor "the notepad level", go nuts by building a database of key shortcuts and commands in your head.

Is 2 hours a considerably "steep learning curve" for the thing you are going to look at and interact with 8 hours/day, 40 days/week, for the rest of your engineering career?

Compared to zero hours for a more traditional editor like TextMate or Sublime Text 2 that has command-S as save? Sure is.

Also, at the point that you first pick up an editor like Vim, you have no idea whether it's going to work for you, so the 8 hours/day is a furphy.

MacVim supports stuff like Cmd-S for Save just fine.

2 hours? Once thing is reading a tutorial in that time and another is use vim without having to think to enter command mode and moving around comfortably.

I use vim daily and it's my editor of choice (so far, at least), but it needs commitment and time to get use to the modal system and to be used without feeling that is absolutely broken. After 8 years I'm sure it feels very natural (as it feels now to me), but I sure remember that it took me time to get use to it. It's not just reading a tutorial...

I must disagree with you. I have used TextMate 1 for many years. Lighter weight than Emacs for quick edits and rapid file browsing. My usual use case is working in a shell and using the mate command line alias to quickly view/edit files. BTW, I use IDEs for serious development: I don't open TextMate and stay in it for an hour, rather using it for fast little tasks. I have received huge value for a low price.

TextMate lacked polish, I think that is the fairest indictment you can make of it. It really did bring some awesome features and customizability to the table with probably the lowest difficulty-of-customization to flexibility ratio in the history of editors. It's easy to dismiss how impressive it was in 2004 because so much of it has trickled down to other editors. It really goes hand-in-hand with Rails which people are so quick to criticize today because of how much improvement it spurred in the competition.

This is pretty harsh, but you're right: it was always sub-par, with some compelling features. As a long-time BBEdit (and thus Mac) user, my impression was Textmate appealed to "switchers" who may not have been aware of better options.

I tried Textmate many times, but single-character undo was always unforgivable.

> I tried Textmate many times, but single-character undo was always unforgivable.

I've heard this complaint before. On the other hand, I can't forgive multi-character undo. I think it's just user preference.

Furthermore, I still use TM1 all day every day. When people call it sub-par or compare its features to other editors it often reminds me of the checklists comparing the first iPhone to other (crappy) phones at the time. You can check all the boxes you want, but if the product doesn't get the details right it doesn't matter. TM isn't for everyone, but it's certainly for me, and I shopped around. (I don't mean to imply all other editors are crap.)

edit: I'm a life long Mac user, 15 year emacs user. I think it has a lot of appeal to non-switchers. That said, it definitely rode the Rails wave.

Are you kidding about the undo thing? I can understand that some people might like TextMate despite its flaws, but I cannot imagine that anybody would really prefer one-character-at-a-time undo.

I've been developing for the Mac for since 2000, and TextMate is the only app I can recall seeing with that "feature".

Nope, not kidding. I prefer single character undo for programming.

Everyone has a right to their opinion but "No Vim or Emacs style brilliantness" followed by "I don't think Textmate deserved all that success --it should have happened to a better editor." doesn't really make sense does it?

What TextMate did do was act as a nice gateway drug to Vim and Emacs for many developers and also paved the way for other editors like SublimeText to prosper. Who would have thought that a one-man text editor startup on a single platform could actually be financially viable?

TextMate came up with the brilliant idea of "textmate bundles"- the fact that both Sublime Text as well as E-TextEditor on Windows will import/use these bundles is a pretty good sign of how good that idea was.

Regardless of whether TM deserved its success, its time has past. I took this as a reminder to finally delete the binary from my Mac. Sublime all the way, and without irritating licensing restrictions.

The way its macros worked was brilliant and much imitated (so much so that it's no longer a differentiator).

>I don't think Textmate deserved all that success --it should have happened to a better editor.

This doesn't make much sense. You say it appeared at the right time and was pretty and native on OS X. What better editors deserved the success it had that were available when TextMate was popular? Which editors were available on OS X that weren't giant IDEs or VIM and Emacs style editors that have a ridiculously steep learning curve?

I've never understood when people say something or someone doesn't deserve something they've earned. TextMate is successful because lots of people bought it. Who are you to judge what deserves to be successful and what doesn't, especially when history seems to disagree with you?

The early and enthusiastic support of DHH / the Rails community had a lot to do with the initial success of TextMate. Whether this was 'deserved' is too subjective to be worth debating. It happened, and TextMate took off as a result.

Community effects can never bring a poor technology to the top. They can keep a great technology from overtaking a good technology, though. But TextMate's success was not only from the community support, but also from its capabilities. Snippets were one of its cool features, for example, that helped it gain success.

Vim has snippets too. I'm sure there's an Eclipse/Visual Studio plugin that provides the same functionality. There's really nothing revolutionary in here.

Thanks for pointing this out, and I don't disagree. It's debatable whether TextMate was revolutionary or a huge leap ahead of the competition at the time. But batista claimed that "Textmate was always a sub-par editor" and implied that its success was due to hype, not its capabilities, which I find hard to accept as true.

I'm not claiming that TextMate was the best choice, just that it's unfair to call it a bad choice or sub-par. It was a very solid, decent text editor.

They came built-in with TextMate - big difference to be a first-class member of the editor.

Vim does have snippets, the plugin is called snip-mate. Why do you suppose that is? ;)

Vim has dozen of snippet scripts: http://www.vim.org/scripts/script_search_results.php?keyword...

They do, but 2 of the top 3 mention TextMate in the title!!

Vim had abbreviations way before TextMate has been created. Think about that.

Doesn't matter, you'd still have to learn Vim first.

Textmate is about bringing unix and scripting power to people who don't have time to learn an esoteric system

> people who don't have time to learn an esoteric system if they had time to learn a programming language why they don't have time to learn how to speak with Vim? also I don't think that vim commands are more "esoteric" than (eg) objective-c and actually, if you are short of time then you should learn to code faster with a faster text editor like Vim (or Emacs of you prefer..)

imo the real problem is laziness

textmate is appealing because it always had a great default configuration. For many people getting old-school editors to play well required too much config ninjaing.

>This doesn't make much sense. You say it appeared at the right time and was pretty and native on OS X. What better editors deserved the success it had that were available when TextMate was popular?

The fact that there wasn't another editor that fulfilled all those "needs" at the time (mainly: simple, native looking) does not mean that it deserved it's success. What I'm saying is, it got its success not because it was better than the competition but because there was no competition.

So, I just wish that something better was available at the time with the wanted features (mainly: simple to use and native looking). It would have been better for all in the long run.

>I've never understood when people say something or someone doesn't deserve something they've earned.

You can earn something for lots of reasons, not always the best of reasons. Nixon got the vote, doesn't mean he also deserved it.

>TextMate is successful because lots of people bought it. Who are you to judge what deserves to be successful and what doesn't, especially when history seems to disagree with you?

Actually history seems to agree with me. Did you miss the part when the promised 2 was worked on for 6 years, got nowhere, delivered a disappointing alpha and got dump half-heartedly (as in with GPLv3, with the intention to continue commercial development) on GitHub?

> The fact that there wasn't another editor that fulfilled all those "needs" at the time (mainly: simple, native looking) does not mean that it deserved it's success. What I'm saying is, it got its success not because it was better than the competition but because there was no competition.

What kind of doublespeak is that? If there's a race, and you're the only one who shows up, you deserve to win, period.

Uhh, the same doublespeak as pointing out that if you're the only one to show up for a race, you're also the one that came in last?

>What kind of doublespeak is that? If there's a race, and you're the only one who shows up, you deserve to win, period.

It's not doublespeak, it's common sense. If "you're the only one who shows up" it's not much of a race (or a victory).

> Actually history seems to agree with me. Did you miss the part when the promised 2 was worked on for 6 years, got nowhere, delivered a disappointing alpha and got dump half-heartedly (as in with GPLv3, with the intention to continue commercial development) on GitHub?

I dunno. I think the part where he could afford to spend 6 years working on some pie in the sky product which never released, subsisting entirely on income from TextMate 1 which was minimally enhanced during that time, suggests that history agrees with the "TextMate was very successful" view.

Two points worth mentioning. Emacs 2.4 comes with many features by default that shrink the learning curve.

And most importantly, check out Technomancy's emacs starter kit. It even comes with a package that lets Textmate users keep using their familiar keyboard shortcuts if they want.

Not 2.4. 24.

You're 22 and a half version numbers short. Emacs is more than 30 years old, remember.

typo :)

Reply on news.ycombinator.com