Would a standard 'universal' GUI specification be useful?

["Nick Roberts" <Nick.Roberts@dial.pipex.com>]

One of the big wins for projects programmed in Ada is the relative ease with
which Ada code can be ported, especially when this is just between different
compilers.

However, there are many barriers to porting Ada programs between different
environments (operating systems). Not least among these barriers is the wide
disparity between the various different GUI (Graphical User Interface) -
what used to be called WIMP (Windows, Icons, Mouse, & Pop-up Menus) -
environments currently extant. In large part, they simply provide different
ways of doing the same thing.

A standard 'universal' GUI facility (in the form, I fancy, of a hierarchy of
library package specifications, together with defining documentation) could
provide a single vehicle for Ada programmers to use the facilities of
different GUI environments, and the portability of their code would be
tremendously enhanced. I believe such a facility could also provide a
slightly easier way for Ada programmers to get to grips with a notoriously
slippery branch of modern programming (by hiding much of the messy details).

However, such a standard would need to be large, in order to cover the wide
area of functionality involved (especially considering the necessity for
efficiency of implementation), and it would have to be truly impartial and
non-proprietary. It would also have to be designed to be flexible enough to
allow for the special facilities of individual GUI environments to be made
available to programmers, and also to allow extra special facilities to be
added when necessary.

I would be willing to 'start the ball rolling', by proposing a skeletal GUI
structure, which could then be fleshed out and improved upon by those most
knowledgeable in the various disciplines involved. I would also be willing
to co-ordinate and oversee the project, but hopefully there would be, of the
many others more suited to this task than myself, someone who would be
willing to take it on. There would undoubtedly have to be a degree of
co-ordination with related projects (e.g. OpenGL). I suspect particular care
would be required to ensure that the project did not become bogged down by
arguments of form (rather than substance).

This project would be a big enterprise, so it would surely have to be a
highly collaborative effort; it would only work with the participation of a
fairly large (if not very large) number of people. If this commitment is not
going to be forthcoming, the project is unlikely to succeed.

So: (a) do you think this idea worthwhile (would it be used, in practice);
(b) would you be willing to make a contribution (even if it's only a small
one)? And (c) any volunteers for co-ordinator?

To save bandwidth, please e-mail replies to me (at the e-mail address given
below), unless you have a comment which will be of wider interest. I will
post a summary of the responses, to this newsgroup and wherever else is
relevant.

I would be particularly grateful if you would disseminate this message to
your colleagues, and to anyone else who might be interested.

--
Nick Roberts
ThoughtWing Software, Croydon, UK
ThoughtWing@dial.pipex.com