Re: RE: Common Lisp window standards

[Scott McKay <SWM@STONY-BROOK.SCRC.Symbolics.COM>]

    Date: Tue, 7 Jul 1992 12:16 EDT
    From: moon@cambridge.apple.com (David A. Moon)

    [info-mcl removed]

    > Date: 7 Jul 92 10:22:45 U
    > From: "pierce" <pierce@at-mail-server.vitro.com>
    > Subject: RE: Common Lisp window standards
    > To: Info-MCL@cambridge.apple.com, SWM@STONY-BROOK.SCRC.Symbolics.COM
    > 
    > Scott,
    > 
    > David Moon writes:
    > 
    > >I haven't looked at the technical details, but my somewhat ill-informed
    > opinion >is that although Bedrock and CLIM have some overlap, each also does a
    > lot that >the other doesn't do. 
    > >You can also bet the farm that Bedrock will receive a lot more development and
    > >support resources than CLIM, more than just the extra resources required to
    > >compensate for the disadvantages of doing it in C++.  I'd like to see the Lisp
    > >community get some benefit from all that expenditure.
    > 
    > >This suggests to me that a good strategy would be for CLIM 3.0 to be based on
    > >Bedrock; that is, for the overlapping part of CLIM (probably a little less
    > than >half) to be removed and replaced with Bedrock, and CLIM to concentrate on
    > its >unique advantages.  In the long term I think this would be better for
    > >everybody.
    > 
    > As one of the main CLIM designers and implementors,
    > 
    > Could you please comment on this strategy?

    Since I assume Scott has never even heard of Bedrock, that might be difficult for 
    him!  I don't think Apple and Symantec have published any specifications on 
    Bedrock, only press releases.  Scott, for purposes of answering this question, 
    should you choose to do so, you can assume Bedrock is like MacApp (I've seen a 
    very early document which justifies that assumption at this very high level of 
    detail) except that it will be implemented on Mac, MS-Windows, OS/2, and 
    Unix/X-Windows (I don't know what dates for what platforms, and I hope I copied 
    the list of platforms correctly from my memory of the press release; don't count 
    on what I just said to be the same as what Apple says).  So in addition to 
    windows, menus, dialogs, and controls (you'd call some of those widgets, I'm not 
    sure which ones), it also has concepts of commands, undoing, and drawing 
    canvases.

If my reply looks a bit mushy and vague, that's because it is...

On the surface, doing a Bedrock port of CLIM is by no means a bad idea,
and may be a good one.  Certainly the "windows, menus, dialogs,and
controls" and "drawing canvases" sorts of things are already modelled by
CLIM 2.0, so porting those is simply a matter of doing the work, and
extended CLIM's internal model where necessary.  CLIM natively supports
commands, but in order to support Bedrock "commands", CLIM would need to
have a more abstract model that has per-port implementations (which is
how CLIM does gadgets right now).  Again, this is probably relatively
straightforward.  A toolkit I am working on that is built on CLIM
supports undo/redo, but I don't know how it would fit with Bedrock.

If someone can get Bedrock design documents to me, I would be happy to
research the matter further.  Drop me a message if you have documents
for me, and I'll give you my US Mail address.