Re: gtkada/glade info needed

[Marin David Condic <mcondic.nospam@acm.org>]

Ted Dennison wrote:

> I used it at work to build a little app for showing our application's
> varaible values using vxWorks' WTX (debugger) protocol. It was very slow
> going at first. It struck me that Glade would be very tough to figure out
> for anyone who isn't fairly familar with how Gtk or Motif GUI building
> works. The main trip up would probably be in figuring out what widgets you
> have to use as your base (window). Only certian ones are allowed. You can't
> just create a GUI with one pushbutton.
>

Well after a bunch of bashing around, I kind of figured out how to get a few
simple things into a window so I could try to make some code. It still looked
like fewmets. I couldn't quite figure out how to make buttons that were the same
size or get good control of where things appeared. I understand that this allows
for automatic resizing, but there must be some tricks for how to make a window
that doesn't simply have everything scattered over God's Green Acres. This is
where some version of a "How To" manual would come in really handy.

>
> It also has some rather annoying crash bugs on Windows (NT at least). But
> since Windows isn't its primary port, this is probably to be expected.
>

Didn't succeed in killing Glade under Win2k. Of course I didn't try everything
yet. :-) What I *did* kill was the code generation.

> I had problems with that too. For one thing, I think there was a file that
> has to be deleted before you save (after the first time). I kind of liked
> this, as it prevented me from accidentally wiping out my old GUI before I
> was ready. But I did have to write a script to backup my GUI package bodies
> and delete that file. Then another script I wrote restored my GUI package
> bodies after gate was run. I'm working from memory here, so I can't be much
> more specific for certain. If no one else responds before I go back in on
> Monday, I'll take a better look at what I did.
>

When looking at the CLI window, I was able to see errors that had to do with
trying to create directories that didn't exist. I could probably post the
*specific* error messages, but I just didn't save them the last time around.

> Actually, I believe ObjectAda does more or less what you are asking for, it
> just isn't free. :-)
>

I had ObjectAda a few years ago. It was O.K., but I thought there was a bit of a
derth of documentation. Too much was assumed about what you knew of Windows
programming. Sort of like saying: "Well, first go out and buy MSVC++ and learn
how to use that. Then you'll understand how to use this."

Of course it has probably improved in the last couple of years and perhaps its
time to take another look at it. I don't mind paying as long as I think I can
get what I need and the $$$ stay expressible in three digits. :-) Naturally,
this does limit you to Windows programming (as does VC++) and I rather thought
I'd be able to sell the concept of doing the next project in Ada if I could show
the capacity to port easily to more than just Windows.

>
> As for doing something free, C++ doesn't really have such a system (on
> Windows) either. What would such a set of tools include? Should the GUI be
> Windows with ActiveX support and all that jazz, or should it be TK, or
> should it be Java, or should it be Gtk+? How about 3D libraries? There are 3
> popular ones on Windows right now, OpenGL, Direct3D, and GLIDE (do we have
> enough GL*DE acronyms yet?). Even Sockets support is in a similar situation.
> Any choice you make is not going to satisfy everyone.
>

Please don't mistake me for a C++ advocate. Its just that after working with
MSVC++ I can see that the *toolkit* is done rather well and I had not seen such
a complete and well integrated toolkit for Ada. Its the same sort of thing you
get when looking at kits for doing embedded work on small computers. Lots and
lots of well integrated development tools that use a C compiler and not much to
say for Ada.

As for trying to please everyone, I understand you can't do that. But its almost
as if just picking some set of tools and pulling them together nicely, might be
enough to create a de facto standard. Pick a set of GUI tools you can expand on
and can make work on a variety of platforms. Pick a debugger. Pick whatever you
like. Put it in a bundle that is easy to install, works together smoothly,
doesn't crash and is well documented and I think you'd have something there.

> That's because they give you no choices! Your GUI will be Windows (with
> ActiveX support and DirectX). Your 3D will be Direct3D. Your sockets will be
> WinSock.
>

Freedom of choice is what I've got. Freedom *from* choice is what I want. :-)
You can push choice too far sometimes. For example, assembler on a bare machine
lets you do *anything* that is possible to do with that machine. Your choices
are limitless. The instant you throw an OS on the machine, you have, by
definition, limited your choices. With Windows, ActiveX, DirectX and whatever, I
can pretty much build an app that will work "Good Enough". And as we know "Good
Enough" is not nearly so wonderful a thing as "Perfect". But its "Good Enough!"
:-)

I'd still like to get portability out of the deal since that is one of Ada's
strong suits. With the right tools in place, it would be a *very* good selling
point to show developers how they could build it on Windows and run it on Unix
(or OS/2 or VMS or Lord knows what else!) Its a big project potentially, but I
think its one that would make Ada a force to be contended with.

>
> I'd agree here, particularly with the first sentence. I don't even think
> "like" is strong enough. Perhaps GtkAda does need a bit more polish.
> Certianly GLADE on Windows needs a bit more polish. But what they have done
> so far in such a short time is very impressive.

Oh, I may complain, but I do have to respect the effort that has gone into it.
From what I have seen, it has a lot of potential and could easily expand into
something extremely powerful. I guess what frustrates me is that I am hoping to
push for the next big project to go with Ada, GtkAda, etc., on the grounds of
portability. But for me to do that, I've got to have enough confidence in the
toolset to believe we can get the job done and that getting the rest of the crew
up to speed using it isn't going to be an impossible burden. With a limited
amount of documentation and uncertainties about quality, its hard to make the
case. I was building a toy application here with the hope that I could take it
around and say "See? It runs on Windows and looks pretty good, eh? What do you
say we glom onto a Sun and see if it works just as well there?" I'm just not
there yet. :-(

MDC
--
======================================================================
Marin David Condic - Quadrus Corporation - http://www.quadruscorp.com/
Send Replies To: m c o n d i c @ q u a d r u s c o r p . c o m
Visit my web site at:  http://www.mcondic.com/

    "Giving money and power to Government is like giving whiskey
    and car keys to teenage boys."

        --   P. J. O'Rourke
======================================================================