Re: ADA COMMON ENVIRONMENT (comments)

Re: ADA COMMON ENVIRONMENT (comments)

[Tapani Rundgren]

Mark Taube {90518} wrote:
> 
[snip]
> 
> >1.  The X Window System (X11)
> >2.  Win32 (MicroSoft Win32)
> >3.  ODBC (Open Database Connectivity)
> >4.  MFC (Microsoft Foundation Classes)
> 
>   Wouldn't it be better to define a portable platform independent
> "PIGUI" Ada class specification that could "liberate" Ada somewhat.

Objective Interface Systems (OIS) is working on a Fresco implementation
in Ada which is platform independent (OO interface to X11).
(More info about Fresco -> http://www.faslab.com/fresco/HomePage)
(OIS -> http://www.ois.com)

/Tapani Rundgren

Common GUI Environments (was: ADA COMMON ENVIRONMENT (comments))

[Larry Kilgallen]

In article <Pine.ULT.3.91.961021072343.3261X-100000@swlrap.msd.ray.com>, Mark Taube {90518} <mmt@swlrap.msd.ray.com> writes:

>>1.  The X Window System (X11)
>>2.  Win32 (MicroSoft Win32)
>>3.  ODBC (Open Database Connectivity)
>>4.  MFC (Microsoft Foundation Classes)
> 
> Wouldn't it be better to define a portable platform independent
> "PIGUI" Ada class specification that could "liberate" Ada somewhat.
> Even Microsoft has made MFC a somewhat portable class framework for
> the C++ world. Given the considerable language differences between
> Ada and C++, it seems a clean start reworking ideas already out there
> would benefit the Ada community. MS Windows is not the be-all, end-all
> environment for everybody.

I am convinced that a sufficiently portable GUI programming
environment will not satisfy the tastes of platform zealots
in the customer base (those who do not write Ada programs,
but might purchase them).

In reading the manual for the MacApp GUI framework for
MacOS, I found additional ammunition in the strong support
provided for AppleScript and the Macintosh Open Scripting
Architecture.  These are checklist requirements for many
in purchasing Macintosh software these days.

Likewise as a user of the Common Desktop Environment for
Motif, I would want to buy an application which provided
support for reactivating itself in the same workspace (I
may have the terminology slightly wrong) the next time I
logged in.

Working with ObjectAda for Windows (and reading the
Windows 95 book which was recommended on c.l.a) I have
recently learned about "Multiple Document Interface"
windows, and they seem to have no counterpart on the
Macintosh or in Motif.

This experience leads me to presume that OS/2 will
also have at least one GUI programming mechanism
which is unique.

The commercial (non-Ada) efforts to provide a uniform
GUI programming environment even between just Windows
and Macintosh have all come up short of the requirements
of platform advocates of either persuasion, and those
which truly are designed in an agnostic fashion (not
Apple porting to Microsoft or vice versa) generally
are found wanting in _both_ environments.

There may be some inhouse efforts where once can force
a non-conforming (to the platform GUI) application down
the throats of end users.  Those efforts will probably
save money (even tax money in the case of government
projects).

But for Ada to be a success in the commercial world it
is necessary to be able to build commercial applications,
and at this point in time that means rigorous compliance
with platform GUI standards.  Thus, I feel the current
approach of separate bindings for Motif, Windows, etc.
is the appropriate one.

Larry Kilgallen
using Ada without a Mandate

ADA COMMON ENVIRONMENT (comments)

[Mark Taube {90518}]

 
>The Ada Common Environment Working Group of the Ada Resource Association
>is accepting candidates for adoption as "Ada Common Environment
>Bindings."  One binding from each area will be designated as "ARA Common"
>and will be recognized by member companies of the Ada Common Environment
>Working Group.

    Excellent idea, we have a fair amount of "broken" code around here
due to the Ada83 X windows bindings fiasco.

>The nominated binding must be an existing implementation. It must be
>freely distributable, though not necessarily in the public domain.
>Bindings which do not rely heavily on vendor-specific language features
>are a plus.

  Hey this is Ada, Why not define the the specs as standardized public domain
and allow vendors to make whatever body implementations they want. The Ada
community needs portability between Ada implementations which doesn't exist
in the Ada83 world.
  .
  .
  .

>1.  The X Window System (X11)
>2.  Win32 (MicroSoft Win32)
>3.  ODBC (Open Database Connectivity)
>4.  MFC (Microsoft Foundation Classes)

  Wouldn't it be better to define a portable platform independent
"PIGUI" Ada class specification that could "liberate" Ada somewhat.
Even Microsoft has made MFC a somewhat portable class framework for
the C++ world. Given the considerable language differences between
Ada and C++, it seems a clean start reworking ideas already out there
would benefit the Ada community. MS Windows is not the be-all, end-all
environment for everybody.

  PIGUI References:

   http://www.zeta.org.au/~rosko/pigui.htm  
   http://www.cs.unm.edu/~wampler           "V" for C++
      see also  C/C++ Users Journal August 1996, page 21
  
 I realize "time to market" is important but sometimes HASTE makes WASTE.

-----------------------------

       Mark Taube 
       Raytheon Electronic Systems
  

Re: Common GUI Environments (was: ADA COMMON ENVIRONMENT (comments))

[Mark Taube {90518}]

On Mon, 21 Oct 1996, Larry Kilgallen wrote:

> In article <Pine.ULT.3.91.961021072343.3261X-100000@swlrap.msd.ray.com>, Mark Taube {90518} <mmt@swlrap.msd.ray.com> writes:
> 
> >>4.  MFC (Microsoft Foundation Classes)
> > 
> > Wouldn't it be better to define a portable platform independent
> > "PIGUI" Ada class specification that could "liberate" Ada somewhat.
> > Even Microsoft has made MFC a somewhat portable class framework for
> > the C++ world. Given the considerable language differences between
> > Ada and C++, it seems a clean start reworking ideas already out there
> > would benefit the Ada community. MS Windows is not the be-all, end-all
> > environment for everybody.
> 
> I am convinced that a sufficiently portable GUI programming
> environment will not satisfy the tastes of platform zealots
> in the customer base (those who do not write Ada programs,
> but might purchase them).

 You are probably correct. Actually the only binding I disagree with
is the one to MFC. I agree that this is the defacto industry standard 
for the C++ Windows environment, but I'm pretty convinced that a binding 
to MFC (I just starting working with MFC in C++) is not the way to go. It
would probably be better to define an Ada based class on top of Win32 or
maybe Open32 (can't find any tech docs on this only press releases). I think
Thomson has done something in this area. But the base window type classes 
could be used for any GUI. The specifics for each platform could be added in 
at the appropriate subclass definition. 

> 
> There may be some inhouse efforts where once can force
> a non-conforming (to the platform GUI) application down
> the throats of end users.  Those efforts will probably
> save money (even tax money in the case of government
> projects).

 Actually there are a lot of commercial activites in this area for
C++, including MFC on top of X/Motif. How successful, I don't know.

> 
> But for Ada to be a success in the commercial world it
> is necessary to be able to build commercial applications,
> and at this point in time that means rigorous compliance
> with platform GUI standards.  Thus, I feel the current
> approach of separate bindings for Motif, Windows, etc.
> is the appropriate one.

 But maybe the Ada world could do a better job than the C++ world where
everything wins by market-share and not superior technical implementation.

 (I wish some very experienced OO people would jump into this thread)