[TxMt] Split Panes: who/what/when/where/why

[Allan Odgaard]

On Mar 14, 2005, at 7:48, Jonathan Ragan-Kelley wrote:

> Searching through the archives of this list, and other
> TextMate-related materials, I've seen that this has come up a handful
> of times, but not in a while, and rarely with much of any response

Unless you want an ETA (which I'm not too keen on providing) then the 
fact that I've added the feature request to my to-do is really all the 
response you need.

> Why does TextMate lack any sort of single-project/multi-window
> functionality

I know it's not what you asked for, but you can control-click the files 
in the project drawer and open these in new windows.

> Is there any intention of adding such functionality (in the near 
> future)?

Yes. I need to get the new syntax system completed, I hope beta 6 will 
be out this week, then a few betas will follow where the main thing is 
still syntax/scope/style sheet/selector/behavior related, and after 
that I intend to redo the project window/drawer, where split views is 
likely to be one of the enhancements.

> [...] I have a hard time imagining how the developers work on this 
> very 30+ KLC app, itself, without the ability to deftly maneuver many 
> files with keyboard alone and to see more than one thing at a time.

I only use the keyboard for navigation. I open new files mostly with 
cmd-T, go to header/source, or with the project wide search if I need 
to find the declaration of a class/function (I have function prototypes 
written so that I can search for these w/o getting a function call as a 
false positive).

When I need to work/see things in different sources, I arrange the tabs 
so I can go back/forth between them with option-cmd-left/right, and if 
it's in the same source, I use bookmarks.

> [...] foiled by a single, fatal (and bizarrely unique, among 
> programmer's editors) flaw
> in just one small aspect of its implementation which completely 
> prevents me from doing anything more than diddling around and getting 
> frustrated, time and again.

As Chris correctly stated, each user has his own thing that he thinks 
is a crucial show-stopper. To loosely quote from some of the private 
letters I've received recently:

  “your editor is the only one in the whole world without native ftp 
support”
  “some of the fundamental features missing: [ list of items ]”
  “Asian Language is the basic requirement of a text editor for 200 
billion CJK people”
  “a combination of the following prevent me from using this application 
[ list of items ]”
  “with significant improvements on the syntax system I might consider 
purchasing your application”

And this is just a fraction of what I get, the list of actual 
letters/requests and the list of what I would like to do with TextMate 
is starting to get absurdly long, and unfortunately I do not have 
access to a master switchboard where I can just enable and disable 
features as I please -- they all require work, planning and probably 
most importantly; motivation!

> [...] But I can't imagine that at least rudimentary split panes (or 
> some functionally similar interface feature) could possibly be nearly 
> as implementation-heavy as many of the features that seem to be coming 
> every few weeks.

I understand that development priorities can seem unintuitive when you 
only see the results coming out and see that whatever you'd hoped would 
appear didn't, and you think that this can't possibly be that big a 
deal to implement -- here are some things to remember:

  • that something is simple doesn't make it more likely to be 
implemented, I have so many “simple” items on my to-do that if I only 
did these, I would never get any of the complex things done,
  • that something appears simple is not always the same is it actually 
is simple, this depends a lot on how I designed the underlying parts 
that this feature builds on,
  • some “simple” items get attached to more complex rewrites and thus 
is indirectly not simple,
  • even though I sell it for money, TextMate is my hobby and not my 
job, so at least 50% of the work I do with it should be done because I 
find it interesting or find the result desirable.

The last point is probably what counts. If TextMate was my job and my 
work assignments were dictated solely by user requests, I'd quit my job 
-- it's that simple ;) This doesn't mean that I don't du requests, it 
just means I need to do them in my own time, so to speak. E.g. with 
regard to split views, I'm not adding this to the current 
implementation of the project window because I'd much rather redo the 
entire thing and take care of all the shortcomings in one shot, but 
this will have to wait till after I'm done with the current project, 
which is improving the syntax system.

> Is there anything I or anyone else could do to help it along (like 
> [not] sending more long, love-lorn emails)?

While I do read all feedback (and appreciate most ;) ), repeating a 
request that I've already acknowledged doesn't boost its priority.

I certainly don't lack things to do with regard to TextMate, and I 
think user requests should be treated equally, even when one user is 
more vocal than another ;)