Things to do before GUI-developers can start their work

--- Raymond Toy <toy@rtp.ericsson.se> wrote:
> 
> Neither do I.  There is no standardized socket interface so you'll
> either have to get some socket library that runs on the supported
> systems or make such a socket library for maxima.

Ick.
 
> And CMUCL/SBCL only has threads on x86.  No threads anywhere else.

How hard is it to impliment threads?  Is it something we can hope to
see mature more within the next few years on the various lisps?  I
didn't realize it but they could be a critical piece of the puzzle.
 
> Another alternative GUI might be McCLIM, which is entirely in Lisp (I
> think).  My understanding is that you could then have a nice display
> and also be able to use the mouse to select various parts of the
> output and paste it back in.
>
> But, of course, no such system exists today.  But McCLIM supports
> such things.

The alternative I favor actually is Garnet, which is also native lisp
and very well done.  It has really excellent documentation as well,
which is always good.

> P.S.  McCLIM makes heavy use of CLOS and gray-streams, so gcl might
> may not be able to compile it at all.  And clisp may be unusably slow
> with McCLIM.  This might not be a great alternative.

Garnet has its own object system, and already currently runs with good
speed on CMUCL and Clisp.  I have poked around with it a little, and I
think it can eventually be made to do pretty much whatever we need it
to.  The problem will be threads, because otherwise the interface will
have to wait for the execution of each math command before it can
accept any input.  I don't know why that didn't occur to me before.  I
know clisp doesn't have them but has some sort of partial
implimentation, and I don't think gcl has even thought about that yet. 
I have no idea about openmcl.

CY

__________________________________
Do you Yahoo!?
Yahoo! Calendar - Free online calendar with sync to Outlook(TM).
http://calendar.yahoo.com