Re: GWindows and David Botton

[David Botton <david@botton.com>]

On 2004-10-05 18:40:36 -0400, Stephen Leake <stephen_leake@acm.org> said:
>> * GWindows for the most part (but not all) is already a _superset_ of
>> CLAW both in design and functionality
> 
> Ok, so it won't be hard to merge GWindows into CLAW.
> Although I agree with Randy; "superset" is not a proper description here.
> 
>> * Where GMGPL CLAW has more functionality, the code can easily be
>> integrated in to GWindows
> 
> Or vice versa.

Actually my point is that it is easier to integrate the subset, i.e. it 
is easier to port CLAW "bindings" in to the GWindows "framework"

> 
>> * GWindows ease of use for teaching GUI programming do to model and
>> good specs
> 
> I don't get this impression from reading the code. I have not had much
> experience in teaching any Ada GUI system to users. But I understand
> the CLAW code more easily than the GWindows code, because the CLAW
> code has better comments.

CLAW is good code, I've never claimed else. I've also never heard 
complaints from lack of documentation in the specs, but certainly open 
to add more where needed.

>> * GWindows shows good Ada design while demonstrating how to be
>> practical in design choices instead of Academic, ie. it is "real
>> world"
> 
> I think CLAW has a better design. But I guess that will always be a
> matter of opinion.

Perhaps. We would have to compare on specific metrics to really make a 
fair statement.

> 
> 
>> * GWindows integration of COM/DCOM/ActiveX/.NET using all Ada, no
>> C/C++ or MFC code needed (as in ObjectAda or older GWindows
>> releases)
> 
> That's _not_ GWindows, that's GNATCOM. As we have already established,
> CLAW can easily take advantage of that as well. Please be consistent
> in this discussion!

No, it is GWindows, what you are missing is that there needs to be 
support all the way around to do it right. That is why GWindows was 
created to start with (to be part of GNATCOM).

> 
>> * GWindows integration of Database support
> 
> That is a point. It looks easy to port to CLAW.

It is.

> 
>> * GWindows design for future integration in to a Delphi like
>> environment
> 
> It would be interesting to hear how CLAW lacks in this department.

Its reliance on only an inheritance model, but everything can always be 
worked around, the question is how easily.


> 
>> * GWindows offers UNICODE API support out of the box making it a
>> better choice for 64 Bit Windows, NT, 2000, XP and Longhorn - ie. the
>> ANSI API (and thus CLAW, the thin Win32Ada binding, GWindows ANSI
>> build, etc) should only be used where Win9X/ME is a must (and then one
>> should offer both binaries).
> 
> Hmm. I guess I haven't looked at the Unicode issue in CLAW. Clearly it
> could be added.

There is a lot of work involved and there are other similar issues.

> 
>> * GWindows can be used for windows CE programming as well if a cross
>> compiler was available
> 
> What prevents CLAW from doing that? Although I don't rank this high on
> a list of requirements.

UNICODE is the primary reason.

> 
>> * GWindows has always been MGPL'd
> 
> As CLAW is now (or will be soon, if we agree to start a community
> project with it).

CLAW is not currently, although a decent subset is. It is up to them, 
but I think they should make sure they have a solid financial model 
before going forward.


> 
>> and as such has always been very much alive even if I was not :-)
> 
> Well, that's not true. That's exactly why this whole discussion got
> started; there was no active support for GWindows, and we were
> discussing the best way to start a support group.

The list was dead for a few months granted, but development didn't stop either.

> 
>> * I have and attachment to the FSF's GNU/GPL philosophy and practice
>> and want to support such projects. (Despite being a Mr. Windows in the
>> Ada world, All of my own work and the majority of programming I do is
>> on Linux, Darwin and Mac OS X. GWindows was created to advocate Ada
>> and GPL/MGPL _not_ Windows, somehow that point always gets lost in the
>> sauce....)
> 
> Me too, but I'm not sure how it applies here. Once CLAW becomes
> GMGPL, it meets the same philosophical goals.

Time will tell. As I said GWindows is already doing it, GMGPL, new code 
being developed by different groups, etc. Given that there is not much 
that CLAW is going to add to what GWindows already is and that it is 
easier to port CLAW code to GWindows, seems smarter to push forward in 
that direction.

When dealing with volunteer efforts the bottom line is what your heart 
desires. From my perspective community code under GMGPL is good 
regardless of what base you start with. In the end if it looks good and 
some one wants it as part of their project they will port it in. So I 
would support any effort to have more code written.


>> * More than one set of GMGPL'd professional extensions has been made
> 
> I'm not clear how this is relevant.

ie. It is already happening even if there is not a lot of discussion 
about it publicly.

> 
>> * The GMGPL Claw version has been around for some time now yet no one
>> pushed to extend it, I expect that anything less than a full GMGPL
>> version of CLAW will mean the same lethargic response
> 
> I personally didn't work with CLAW because it was not GMGPL. Making it
> GMGPL changes that.

If it was GMGPL to start with I likely would have never written 
GWindows, but now that I have and the result is a solid open source 
framework that

> 
>> that despite the availability of many of my tools and some examples
>> that could have been used with out compromising lic.,etc. (as
>> pointed out by others) to give CLAW additional support where
>> GWindows was stronger.
> 
> There could be many reasons why your code was not added to CLAW; lack
> of customer demand, for one. Lack of clear title, for another;

Clear title exists at this point to the code, but at some point that 
may change as GWindows garners more community involvement.

> 
> CLAW has a goal of compiler independence; that's actually a point in its
> favor that I've forgotten to mention.

Not sure that matters so much these days, but I've seen ports of 
GNATCOM/GWindows to other compilers (at least one vendor even ships it 
AFAIK).

David Botton