From: "Stavros Macrakis" <stavros.macrakis at verizon>
> if the presentation of expression output in Maxima were built on
> [various systems], it would not be hideously difficult to present
> the output so that subexpressions would be selectable, representable
> in 1D form in the input area, and thereby editable.
Apparently none of the system you mention are available to us (open
source and compatible with our Lisp platforms)....
> a lot of liabilities come along with using any of these UI systems.
> I'm not even hinting that someone should duplicate these facilities
> somehow as open source in Maxima itself, since the liability of
> employing a particular odd UI modality still remains.
What are the liabilities?
For one, none of the system I mentioned are available (open source and
compatible with our Lisp platforms).... :-)
What is so odd about this UI modality?
Since it doesn't use the native UI substrate on the particular
platform, and doesn't interact with the user in the same way, there is
a larger learning curve and more resistance by potential users. Since
it doesn't use the native UI substrate on the particular platform,
there are whoe layers of UI code that is susceptible to changes in the
underlying platform and which would have to be maintained for all time
by the Maxima community.
How hard would it be to duplicate this functionality?
This is the wrong question, I think. The relevant questions concern
how portable any implementation of the functionality could be, and how
brittle it would be (across platforms, or in response to platform
evolution), how much further it would fragment the UIs used by the
Maxima community (probably already too fragmented), and how it would
be possible to maintain any implementation over a long lifetime.
At one point I had CLIM-like presentations running in Emacs. Of
course, the code no longer works, and not everyone using maxima uses
Emacs, and even the two main lines of Emacs development are completely
different in how they interface to mouse-sensitive text.
Think about these issues first, and only then think about how neat it
would be to rush in and bang out some code that would work, sort of,
for limited time on a few particular platforms.
Sorry to be so negative, but the kind of thing we are discussing is a
medium-sized project with a super-size downstream maintenance cost.
Probably there are more urgent things for the community to address
first.
Is anyone currently working on it?