Re: Why no TextMate for Linux?

On Saturday 06 March 2010 10:05:04 am Paul Harrington wrote:
> Using
> proprietary technology actually does not stop the progress of open
> source.

That much is true, most of the time.

There are notable exceptions -- for instance, the fact that people are 
starting to incorporate textmate URLs into HTML stacktraces and such, so that 
you can just click on something and have it pop up in TextMate, but those 
aren't portable at all. Far better would be to come up with some sort of 
standard "Go to this line of this file in a text editor" URL scheme that the 
open source community could adopt also.

But that's a ridiculously trivial, almost contrived example. Most of the time, 
using proprietary technology doesn't cause problems for open source -- so long 
as the open source project remains independent of said proprietary technology.

I'm much more interested in it on a personal level. Switching text editors at 
this point might be, for most of us, far trickier than switching OSes, and 
could be almost as bad as switching keyboard layouts. (Dvorak, anyone?) By 
picking a proprietary technology, you're doing several things that I can't 
really see being worth the risk:



 * You're locked-in to a single provider -- in this case, a single _person_.
   - If Allan Odgaard doesn't want to implement a feature you want, you're 
SOL.
   - If Allan Odgaard can't fix a bug that's annoying you, you're also SOL.
   - if Allan Odgaard wants money for a new car, you might find the next 
version of TextMate costs significantly more.

 * You're tied to an OS which is notorious for breaking backwards-
compatibility.
   - The next version of OS X is as likely as not to break the current version 
of TextMate. It's entirely up to Allan whether there will be a fix, and how 
much such a fix would cost -- maybe it's only available in a new version of 
TextMate.
   - Once you do upgrade, the new version of TextMate is as likely as not to 
refuse to work on old versions of OS X, so you'd better upgrade all your boxes 
at once.
   - If you don't like the new OS X, for whatever reason, some new version of 
TextMate might force you to upgrade anyway.
   - Switching OSes -- to Linux, to Windows, to Plan9, to whatever -- is out 
of the question for you.

 * You're a programmer, yet you can't program your own programming tools.
   - I don't care how extensible it is, you don't have the source.
   - Look at the tricks tools like Diakonos can do. Can TextMate do that?
   - Basically, TextMate may be at the top of the heap now (though there's 
certainly room for debate), but if it continues to innovate, you can't be a 
part of that. I would hope tools like Diakonos would win out in the long run, 
because people who use them would inevitably contribute needed changes -- if 
there was ever a "scratch your own itch" app, a text editor is it.

 * Again, a SINGLE PERSON is responsible for the destiny of this editor.
   - If Allan should be hit by a bus (not that I am wishing this on anyone), 
what happens to TextMate development?



Now, I don't personally hack on my own text editor. It's also MUCH less of a 
commitment than, say, a framework -- if you were on ASP.NET, for instance, 
you'd be tethered to Microsoft for the lifetime of your app or until you port, 
and it's much harder to port an entire app than to learn a new text editor.

Still, think about it. Would you choose a proprietary programming language? 
Library? Framework? If not, why not, and why would you use a proprietary text 
editor, or debugger, or _any_ proprietary programming tool?

I generally don't bother people about developing on OS X. It's annoying, but 
most of the Ruby stuff is going to be general Unix stuff anyway, not Mac-
specific. But then, switching OSes is easy when your tools are portable.

But this is just my own curiosity. If you were using IE6, I'd hate you 
forever, and this rant would be much longer, and it would never stop until you 
stopped using IE6. If you're using TextMate, I think you're making a mistake, 
but it really doesn't affect me that much.