[TxMt] Quick find and open last closed file

Rob Shearer textmate at v.cx
Wed Feb 7 23:35:57 UTC 2007


> I have been working with an editor with built interpreter for more  
> than ten years.  I understand what you say about Emacs, Stallman  
> decision was really argued over.  I don't know about anyone that  
> uses elisp, for any other thing than make emacs to work.  I think  
> you, like me, like better and modern languages.

I don't think the problem with Emacs is (just) the choice of  
language. *Everything* in emacs is programming, not configuring. How  
many people really understand how C++-mode works in Emacs? The issue  
isn't that the language is obscure, it's that the parsing logic for  
each mode is genuinely complex, the Emacs API is large (look at the  
length of the documentation!) and every bit of the system does things  
its own way. If you want to write an extension the "right" way you  
should expect to invest a full day (at least) cramming on the basics  
of the programming environment.

> Its not true that having an interpreter embedded avoid the using of  
> AppleScript, or any other scripting languages.  That hasn't to do  
> with that fact.  You always can do "system(args)" with any  
> scripting language calling any other with args.

And immediately lose whatever "integration" with the editor's  
internal data structures you were going for. Inevitably having a  
"preferred" language would just become an excuse to leave all the  
other languages/interfaces impoverished.

> Things like Menu control (building them, etc), open/close hooks  
> (procs that must be called when anything open or closes),

You want to screw with the menu structure, and do special things on  
open and close? Stay away from my editor. The each-mode-with-its-own- 
UI stuff is what drove me away from Emacs. And crazy *nix UIs in  
general. There's a reason I'm using the Mac.

> Access to the document:  position, selecting parts of it, changing  
> the insertion point position!!, are all of them impossible with the  
> TM model except that in future versions TM hard-coded them and  
> gives us direct access by new tricky interface gadgeds.

I'd like access to these things. And some of them commands already  
have access to. I expect we'll get more, whether through more env  
variables, or through a more robust interrogation interface (which  
can be called from any scripting language).

Opinions differ. I like TextMate's philosphy better than that of  
Emacs. For me, simple + good citizen + easy to extend > infinitely  
flexible + quirky interface + complex API.

(Have I mentioned that chunked undo > character undo?)

-rob




More information about the textmate mailing list