Ada95 for Windows 95 Reviewers Wanted

Ada95 for Windows 95 Reviewers Wanted

[Ian Goldberg]

Product Reviewers Needed

R.R. Software is in the process of building an Ada 95 package for =
Windows 95 that will contain a "thick", object-oriented Windows binding =
and a GUI application builder.

We have an immediete need for product testers as part of the new product =
release. The review team will be set up in two groups, Tiers I & II.   =20

Tier I testers will be provided with hard copies of both the Binding and =
our Janus/Ada 95 Professional Development System for Windows 95/NT.  =
They will also recieve, at least,  two versions of the bindings, and at =
least two versions of the tools.  Tier I testers will also be provided =
an opportunity to comment on early versions of the design documents.

Tier II shall be comprised of interested programmers that already own an =
Ada 95 compiler package for Windows 95, which can be from any =
manufacturer.  These testers will have the opportunity to download off =
either our Web site or our BBS a soft copy of the bindings.  The design =
documents and the GUI application builder will not be available to Tier =
II testers.

Accepted Tier I testers will be shipped a copy of the Janus/Ada 95 PDS =
for Microsoft Windows NT/95.  This will provide them the opportunity to =
familiarize themselves with the latest Ada 95 features and enhanced =
capabilities.  The use of our compiler will not be required for the =
final product, but will be essential in the early development stages.

Tier I testers will be expected to provide timely feedback on the =
binding and tools.  This should include comments and suggestions, as =
well as bug reports.  They also will be expected to provide example =
application(s) using the bindings, as well as code demonstrating bugs.  =
These testers will be asked to sign a minimal non-disclosure agreement =
for mutual protection.

To apply to be a  tester, please send E-Mail to: randy@rrsoftware.com.  =
Please include the following information: your name, shipping, mailing, =
and e-mail addresses, daytime (work) phone number, Ada 95 compilers =
used, a statement relating your experience using Ada 95, object-oriented =
programming, and Windows programming, and a brief statement on your =
ideas on Windows bindings. =20

We will contact the selected testers in April.  All applicants will be =
automatically placed on the Tier II mailing list, it is not necessary to =
send a separate message to do that.

To be placed on the Tier II tester mailing list, please send E-Mail to =
randy@rrsoftware.com with your name and E-Mail address.  Be sure to =
identify the message as a request for Tier II tester status.  We will =
then let you know when to begin with your review.

Thank you for your interest,

Ian Goldberg
General Manager

Re: Ada95 for Windows 95 Reviewers Wanted

[Mitch Gart]

Ian Goldberg (Ian@RRSOFTWARE.COM) wrote:
: Product Reviewers Needed

: R.R. Software is in the process of building an Ada 95 package for =
: Windows 95 that will contain a "thick", object-oriented Windows binding =
: and a GUI application builder.

(snip)

As you probably know Intermetrics and Labtek produced Win32Ada,
a thin Ada 95 binding to the Win32 API.  Due to legal complications 
with Microsoft it took a while to get this released but now is 
finally released and available on the Web, see a separate 
announcement I am about to post.

Win32Ada is thin but full, covering all the Win32 API, over 70K 
lines of Ada just in the package specs.  I think a thin binding is 
useful partly in its own right and partly as a basis for future 
thick bindings.  It sounds like R.R. is now producing a thick binding.

I think the best thing would be to have an underlying thin layer
that gives direct access to everything in the system, probably
not with the best Ada style, and then one or more higher level 
bindings to selected areas, for example OLE, which are more 
limited in scope but more carefully crafted to give the best
possible interface to Ada 95 programmers.

I don't know how far along you are in your implementation but
I hope you will consider using Win32Ada as an underlying layer 
and making your higher level thick bindings compatible with this.

- Mitch Gart

Re: Ada95 for Windows 95 Reviewers Wanted

[Stephen Crawley]

In article <DoMsF5.5py.0.-s@inmet.camb.inmet.com>,
Mitch Gart <mg@asp.camb.inmet.com> wrote:
>Ian Goldberg (Ian@RRSOFTWARE.COM) wrote:
>: Product Reviewers Needed
>
>: R.R. Software is in the process of building an Ada 95 package for =
>: Windows 95 that will contain a "thick", object-oriented Windows binding =
>: and a GUI application builder.
>
>I think the best thing would be to have an underlying thin layer
>that gives direct access to everything in the system, probably
>not with the best Ada style, and then one or more higher level 
>bindings to selected areas, for example OLE, which are more 
>limited in scope but more carefully crafted to give the best
>possible interface to Ada 95 programmers.
>
>I don't know how far along you are in your implementation but
>I hope you will consider using Win32Ada as an underlying layer 
>and making your higher level thick bindings compatible with this.

Further to this, I would like to make a plea for R.R. Software to
consider placing the package specs for their bindings into the public
domain >>as soon as possible<<, and for other suppliers to consider
adopting the specs as an interim (defacto) standard for Windows 95
thick bindings.

I'm in no position to judge whether or not RR's bindings are of good
quality, but I think that my point is a valid one anyway.  The Ada
user community would not be well served by having lots of mutually
incompatible W95 binding products.  I would hope that the Ada
community is now mature enough to strongly resist such a trend!

-- Steve

Re: Ada95 for Windows 95 Reviewers Wanted

[Robert Dewar]

"I'm in no position to judge whether or not RR's bindings are of good
quality, but I think that my point is a valid one anyway.  The Ada
user community would not be well served by having lots of mutually
incompatible W95 binding products.  I would hope that the Ada
community is now mature enough to strongly resist such a trend!"

I disagre with the sentiment here. Ther is no point in standardizing
thick bindings of low quality. I am not saying anything about the
quality of the RR bindings, about which I too an not in a position
to judge. 

It is the "valid one anyway" in the above comment that I question. There is
certainly value in haveing generally available bindings, and in fact the
ARA recently formed a group specifically oriented to looking at bindings
to see which of them might qualify as being in this category.

I actually think it is much mre important to have a reasonably universal
thin binding since that is what tool vendors need (they are modifying
C generation stuff, they want something as close as possible to the
C binding anyway).

A well done, well documented, thick binding is a useful addition, but
only if it meets these criteria. That being done, I agre that it would
be very helpful if anyone producing bindings would disclaim both
copyrights and interface copyrights on the specs of what they do,
otherwise people will not want to bother trying to get involved with
making them comon anyway.

Re: Ada95 for Windows 95 Reviewers Wanted

[Jere W. Retzer]

Stephen Crawley wrote (in part):
> 
> I'm in no position to judge whether or not RR's bindings are of good
> quality, but I think that my point is a valid one anyway.  The Ada
> user community would not be well served by having lots of mutually
> incompatible W95 binding products.  I would hope that the Ada
> community is now mature enough to strongly resist such a trend!
> 

Two of the advantages of Ada being its readability and portability, I would 
be interested in what people think of the readability of windows Ada code.  
My first reaction to the win32 bindings code I have seen is that it is 
difficult to read, but I am admittedly new.  Also, what happens to 
portability in general when the program is targeted to an environment that 
requires bindings?  For that matter, will we have portability problems if we 
use bindings from different vendors as implied above?  Are these two 
attributes affected by the 'thickness' of the binding as it is being 
discussed here?

Re: Ada95 for Windows 95 Reviewers Wanted

[Jere W. Retzer]

Stephen Crawley wrote (in part):
> 
> I'm in no position to judge whether or not RR's bindings are of good
> quality, but I think that my point is a valid one anyway.  The Ada
> user community would not be well served by having lots of mutually
> incompatible W95 binding products.  I would hope that the Ada
> community is now mature enough to strongly resist such a trend!
> 

Two of the advantages of Ada being its readability and portability, I would 
be interested in what people think of the readability of windows Ada code.  
My first reaction to the win32 bindings code I have seen is that it is 
difficult to read, but I am admittedly new.  Also, what happens to 
portability in general when the program is targeted to an environment that 
requires bindings?  For that matter, will we have portability problems if we 
use bindings from different vendors as implied above?  Are these two 
attributes affected by the 'thickness' of the binding as it is being 
discussed here?

Re: Ada95 for Windows 95 Reviewers Wanted

[Simon Johnston]

Please consider me for the following, I am also a beta tester for the
Thomson ObjectAda7 compiler and would be interested in your bindings
to the Win32 API and GUI builder.

Thanks.

> Product Reviewers Needed
>
> R.R. Software is in the process of building an Ada 95 package for
> Windows 95 that will contain a "thick", object-oriented Windows
> binding and a GUI application builder.
>
> We have an immediete need for product testers as part of the new
> product release. The review team will be set up in two groups, Tiers
> I & II.
>
> Tier I testers will be provided with hard copies of both the Binding
> and our Janus/Ada 95 Professional Development System for Windows
> 95/NT.  They will also recieve, at least,  two versions of the
> bindings, and at least two versions of the tools.  Tier I testers
> will also be provided an opportunity to comment on early versions of
> the design documents.
>
> Tier II shall be comprised of interested programmers that already
> own an Ada 95 compiler package for Windows 95, which can be from any
> manufacturer.  These testers will have the opportunity to download
> off either our Web site or our BBS a soft copy of the bindings.  The
> design documents and the GUI application builder will not be
> available to Tier II testers.
>
> Accepted Tier I testers will be shipped a copy of the Janus/Ada 95
> PDS for Microsoft Windows NT/95.  This will provide them the
> opportunity to familiarize themselves with the latest Ada 95
> features and enhanced capabilities.  The use of our compiler will
> not be required for the final product, but will be essential in the
> early development stages.
>
> Tier I testers will be expected to provide timely feedback on the
> binding and tools.  This should include comments and suggestions, as
> well as bug reports.  They also will be expected to provide example
> application(s) using the bindings, as well as code demonstrating
> bugs.  These testers will be asked to sign a minimal non-disclosure
> agreement for mutual protection.
>
> To apply to be a  tester, please send E-Mail to:
> randy@rrsoftware.com.  Please include the following information:
> your name, shipping, mailing, and e-mail addresses, daytime (work)
> phone number, Ada 95 compilers used, a statement relating your
> experience using Ada 95, object-oriented programming, and Windows
> programming, and a brief statement on your ideas on Windows
> bindings.
>
> We will contact the selected testers in April.  All applicants will
> be automatically placed on the Tier II mailing list, it is not
> necessary to send a separate message to do that.
>
> To be placed on the Tier II tester mailing list, please send E-Mail
> to randy@rrsoftware.com with your name and E-Mail address.  Be sure
> to identify the message as a request for Tier II tester status.  We
> will then let you know when to begin with your review.
>
> Thank you for your interest,
>
> Ian Goldberg
> General Manager
>

with StandardDisclaimer; use StandardDisclaimer;
package Sig is
--,-------------------------------------------------------------------------.
--|Simon K. Johnston - Development Engineer (C++/Ada95) |ICL Retail Systems |
--|-----------------------------------------------------|3/4 Willoughby Road|
--|Unix Mail: skj@acm.org                               |Bracknell          |
--|Telephone: +44 (0)1344 476320 Fax: +44 (0)1344 476302|Berkshire          |
--|Internal : 7261 6320   OP Mail: S.K.Johnston@BRA0801 |RG12 8TJ           |
--|WWW URL  : http://www.acm.org/~skj/                  |United Kingdom     |
--`-------------------------------------------------------------------------'
end Sig;

Re: Ada95 for Windows 95 Reviewers Wanted

[Stephen Crawley]

In an earlier article, I wrote:
>>"I'm in no position to judge whether or not RR's bindings are of good
>>quality, but I think that my point is a valid one anyway.  The Ada
>>user community would not be well served by having lots of mutually
>>incompatible W95 binding products.  I would hope that the Ada
>>community is now mature enough to strongly resist such a trend!"

In article <dewar.827590178@schonberg>, Robert Dewar <dewar@cs.nyu.edu> wrote:
>I disagre with the sentiment here. Ther is no point in standardizing
>thick bindings of low quality.

I agree with that.

I guess my original message was not clear.  The point I was trying to
make was that other manufacturers should >>consider<< adopting RR's
bindings as an interim standard.

If RR's bindings are "no good" [ ... I have no data ... ] then it
would not serve anyone's purposes to standardise on them, and the
consideration process should be a simple one [:-)]

OTOH, if RR's bindings are "good" or even "half-way good", they could
and possibly should serve as an interim defacto standard.

Either way, making the package spec's public domain is a good first
step.

-- Steve

Re: Ada95 for Windows 95 Reviewers Wanted

[Robert Dewar]

"OTOH, if RR's bindings are "good" or even "half-way good", they could
and possibly should serve as an interim defacto standard."

I actually think that by FAR the mor important issue is to have a relatively
widely implemented thin binding. This is what tool vendors need, Relatively
few programs these days are written by making direct calls to the X or
windows interface, instead GUI's are built with high level tools. It is
these tools that need porting to Ada, and for that purpose a well designed,
universally implemented thin binding is what is needed. The intermetrics
bindings are intended as a candidate for this purpose.

Re: Ada95 for Windows 95 Reviewers Wanted

[Ted Dennison]

Jere W. Retzer wrote:
> 
> Two of the advantages of Ada being its readability and portability, I would

Any language's portability ENDS when you venture out of the defined
language and into the OS. Ada is no different here.

> be interested in what people think of the readability of windows Ada code.
> My first reaction to the win32 bindings code I have seen is that it is
> difficult to read, but I am admittedly new.  Also, what happens to

Your first reaction would be correct in this case. Any Ada code that 
makes heavy use of thin OS bindings is ugly, and generally tough to
follow. GUI code is particularly heinous. (Note that I don't say "C" 
bindings. In Win32's case, they are technically Pascal bindings.)

That's why I tend to write thick bindings whenever I'm going to use a
significant amount of OS bindings.

> portability in general when the program is targeted to an environment that
> requires bindings?  For that matter, will we have portability problems if we
> use bindings from different vendors as implied above?  Are these two
> attributes affected by the 'thickness' of the binding as it is being
> discussed here?

1 - It goes out the window. 
2 - Yes.
3 - Somewhat.

3 deserves a more detailed answer. Yes, any "thick" bindings you have
will probably be completely different when you switch vendors (assuming 
your vendors even give you thick bindings). However, the difference in
vendor's thin bindings is likely to be something that will require minor
changes to EVERY binding call. And with thin binding there are a LOT of
calls.

Also, note that if you write your own thick bindings, you will have the
ability to make the nessecary changes in only the body of the bindings 
packages when you change vendors. This significantly reduces the impact.

-- 
T.E.D.          
                |  Work - mailto:dennison@escmail.orl.mmc.com  |
                |  Home - mailto:dennison@iag.net              |
                |  URL  - http://www.iag.net/~dennison         |

Re: Ada95 for Windows 95 Reviewers Wanted

[Stephen Crawley]

In a previous article I wrote:
>>OTOH, if RR's bindings are "good" or even "half-way good", they could
>>and possibly should serve as an interim defacto standard.

In article <dewar.827847069@schonberg>, Robert Dewar <dewar@cs.nyu.edu> wrote:
>I actually think that by FAR the mor important issue is to have a relatively
>widely implemented thin binding. This is what tool vendors need, Relatively
>few programs these days are written by making direct calls to the X or
>windows interface, instead GUI's are built with high level tools. It is
>these tools that need porting to Ada, and for that purpose a well designed,
>universally implemented thin binding is what is needed. The intermetrics
>bindings are intended as a candidate for this purpose.

I won't disagree that thin bindings may be more important than thick
ones (especially when GUI builder tools are involved), but that's not
what RR have produced!

Presumably RR think that there is a demand for thick bindings.  And
assuming that there is a demand, it makes sense that that demand is
filled by one (defacto) standard, not a number of non-standards.

Actually, there's another point that RR and other vendors should
consider.  If multiple thick X11 bindings emerge without any sign of a
standard, Ada users may be more likely to use the standardised thin
bindings and / or GUI builder tools.  Thus, lack of standardisation
may in fact tend to REDUCE the thick binding vendors' market share.

Anyhow, this is all speculation (on my part anyway).

-- Steve

Re: Ada95 for Windows 95 Reviewers Wanted

[Tom Griest]

With respect to RR's thick windows bindings two questions come to mind:

   1) are they built on some commonly available thin binding (or can
      they be moved to such a binding fairly easily)

   2) are the think bindings proprietary?

The "open-ness" issue is important for the users because they don't want to
be locked into any one vendor if they don't have to.  The ability to use
a common thin binding is also important because it is impossible to keep
the thick binding in perfect sync with all the new stuff MS is coming out
with.  The thin binding can cover all the APIs, whereas the thick binding
can make life easier for those writing code to the APIs directly.

BTW, it would be a big mistake to assume that the set of Microsoft
APIs are primarily for GUI development.  This is simply not the case.
Most of them have to do with other things like networking, mail,
object access, URLs, device access, file management, process management,
internationalization, loading dlls, manipulating the registry,  etc., etc. 
and a bazillion other things.  There really are no high level tools that 
handle the middleware necessary to combine all these services.  So,
people will be acessing the APIs directly for some time to come.  
Furthermore, new APIs crop up every quarter.  At a minimum there is a 
lag between when the APIs come out, and when 4gl tools will incorporate 
them.

I would encourage RR to make their thick binding public.  RR should
probably maintain control over it for the time being while
"suggestions" are considered.  If at some time they decide to stop
"managing" the effort, or there is widespread need for an alternate
thick binding, it would seem that the SIGAda Ada Bindings Working Group 
(Windows subgroup) should pick up the process.

-Tom Griest
LabTek Corp.