Re: GtkAda vs Windex, Gwindows

["Warren W. Gay VE3WWG" <ve3wwg@cogeco.ca>]

Stephen Leake wrote:
> Pascal Obry <obry@act-europe.fr> writes:
>>Stephen Leake <stephen_leake@acm.org> writes:
>>>I'm considering giving up on GtkAda for Windows development, and
>>
>>Well that's not the moment to give up :) Gtk+ on Windows is maturing quickly
>>and the 2.4.7 version is quite better than previous ones. Note also that the
>>Gtk+Wimp (Native Windows look&feel) is now part of the Gtk+ project.
> 
> That's all good.
> 
>>Will this makes you stay on the GtkAda side ?
...
> For example, I can't figure out why an EventWindow doesn't respond to
> a Focus event. (If you read the GtkAda mailing list, you've seen this
> issue). The Gtk manuals are not clear on this topic, and neither the
> GtkAda nor Gtk mailing lists have been much help. It's extremely
> frustrating. When I write Ada code at home, it's for relaxation and
> fun; I have a very low tolerance for frustration.

I have to agree that the documentation could be better. I have had
similar "feelings" in the past about GtkAda, but looking for a
portable solution, I was forced recently to re-examine and get better
acquainted with it. I now keep notes ;-)

What it really needs is a better guide and manual. While some of the
GtkAda package organization makes logical sense to me, I have trouble
understanding other ones. I find myself constantly grepping specs
to find what I think I need. Sometimes I am forced to look at bodies
as well.

While some of this is unavoidable as the project matures (like some
of the support moving to this Pango package), I found it rather
confusing. What is really needed is a comprehensive user guide,
with full blown examples (not the few that come with the sources).

One area that I found weak in the extreme was how to use GtkAda to
draw on some sort of a graphics area. Many questions came to mind
about that (and I have done X Window graphics), but I had to do a
lot of experimentation and spec grepping to get through that.

If the GtkAda team is reading this, I would be willing to volunteer
some time on the documentation, though I cannot always be available
(I can work on it in spurts).

I have personally expended a lot of energy on my APQ documentation
because I know how important it is for people. An example can save
the programmer from reading much verbiage and abstraction, and can
cut to the chase. This is the reason the APQ documentation has an
example for every API function.

I hope the GtkAda folks don't take offense to these remarks. I know
that they have travelled a long and perhaps even a thankless road.
I appreciate GtkAda for being there, since application portability
is very important to me. But I think it is fair to say that better
documentation is needed to get more people to use it.

> I suppose I should try reading the C code, maybe it's not all that
> bad. I'd feel better if I felt that others in the GtkAda community
> shared my concerns.

I haven't had to look at the C code much yet, but give me time.
Certainly in fairness to GtkAda, Gtk could be better documented
as well!

> At the same time, although I keep planning to get a Linux system to
> "play with", it never seems worth it; Windows works well enough for
> me, and I really don't have time to spend on two operating systems and
> two GUI systems. I hate sounding like an ad for Microsoft, but it's
> just the facts of life.

I guess it depends upon your end goals. I like to produce software
not only for myself, but software that can be shared with others.
If I only did software for myself, I'd probably just do Linux/*NIX
projects and use X Window bindings or something. Something else
for those few Windows apps that I run on the laptop. Most of the
time, I want to write the app once, and be usable wherever it is
important.

-- 
Warren W. Gay VE3WWG
http://home.cogeco.ca/~ve3wwg