On Wednesday 13 October 2004 19:35, C Y wrote:
> The "interaction" interface I think is probably very clear in all cases
> except perhaps the command line itself. Xmaxima communicates via
> sockets, and IIRC the other GUI attempts have done the same thing.
Actually, Kayali uses Expect, which simulates a terminal (pty). This can
prevent some odd problems when apps don't behave properly in respect of
pipes.
> I
> see no reason why a Lisp GUI should mean a non-clean API - the intent
> is to have a maxima-kernel and separate GUI, regardless of the language
> used for the GUI. A Lisp GUI would have an easier time of it as far as
> the translation between GUI and math and internal forms are concerned,
> but the structure of that interaction I would argue should still be
> well defined. It's up to other GUIs to talk Maxima's language, or
> teach maxima to talk their language.
>
I'm sure that whoever is working with Maxima would appreciate an optional
machine-orientated interface.
> > Since Qt requires a license for the Windows environment, the project
> > suddenly becomes much less interesting.
>
> I don't think his intent is to do "the" GUI for Maxima, and I rather
> suspect KDE users will appreciate a GUI written to interact well with
> their environment. I still think McCLIM is our best long term bet, but
> the emphasis is on long term since McCLIM is still maturing. Which
> works, since an effort by the Maxima team to make "the" GUI is likely
> still years down the road.
>
That's right, I'm coming from KDE/Qt first and Maxima second. To be honest, I
hadn't heard of Maxima until a few weeks ago - my aim is first to provide a
Math GUI to those who are interested, and secondly to help the Maxima
community as much as I can. This is a learning experience and whoever does
develop 'The' Maxima GUI might be able to take a few things from Kayali and
save himself some time - even if it means knowing how NOT to do it!
> As far as I can see, the third party GUIs server two purposes - they
> will take some of the strain off of us to Produce a GUI Right Now, and
> they might help ease Maxima onto more computers. Hopefully when we get
> to it we can redefine the standard for a good mathematical GUI, but
> math bugs, docs, case, and a variety of other issues loom large.
>
It is true that the Linux community is far smaller than the Windows one, but
nevertheless I hope that Kayali, once released to the great unwashed, can
help spur the take-up of Maxima.
Abdulhaq