Conference bulova::decw_jan-89_to_nov-90

Someone else asked this in the DW_EXAMPLES conference, but nobody seems to
have supplied an example.  It would be rather straightforward to adapt the
Ada version to Pascal.  I would not recommend starting with the C version.

				Steve

RE :.1
>
>Someone else asked this in the DW_EXAMPLES conference, but nobody seems to
>have supplied an example.  It would be rather straightforward to adapt the
>Ada version to Pascal.  I would not recommend starting with the C version.
>

Thanks Steve for the quick reply. Is the any Pascal Guro that is willing to do
the adapt?
				*-Doron-*
P.S.
I have never got into pascal, so I believe that there is at least one guy that 
can do this better than me.

    I wrote a Pascal version (based on Steve Lionel's Fortran
    and Ada versions) that is distributed with the Pascal V4.0
    kit.  Look in SYS$COMMON:[SYSHLP.EXAMPLES.PASCAL].
    (I think the installation asks if you want examples; if
    you said no, re-install and say yes.)
    
    /steve

>
>    I wrote a Pascal version (based on Steve Lionel's Fortran
>    and Ada versions) that is distributed with the Pascal V4.0
>    kit.  Look in SYS$COMMON:[SYSHLP.EXAMPLES.PASCAL].
>    (I think the installation asks if you want examples; if
>    you said no, re-install and say yes.)
>    
Steve
 Great work. I was wondering if U R planning to write the DECburger's pascal 
version? Is there any one plan doing it? The Helloworld examp[le is essential 
for beginners but I find the DECburger very important as well.

					*-Doron-*

    The VAX Pascal development group does not have plans to
    do DECburger.  I wrote another DECwindows application in
    Pascal; if you want a look, install CMSGRAPH V3 from toolshed
    on METOO::.
    
    I think the major part of the problem is learning 
    DECwindows (a full-time job), not doing DECwindows in Pascal...
    
    /steve

    Actually, I glanced at the Ada version as Steve Lionel suggested.
    I think it should be an "easy" job to do a blind Ada to Pascal
    conversion.  Perhaps if I find time (not for the next 2 months),
    it might happen.
    
    				-John

It should be easy - the only thing to worry about would be that the
"valued procedures" in Ada would be functions in Pascal.  If John doesn't
get to it first, I might give it a go, since I seem to already be maintaining
three versions of it - what's one more?

			Steve

>    
>    I think the major part of the problem is learning 
>    DECwindows (a full-time job), not doing DECwindows in Pascal...
>    
Steve,

 U R right. Learning the DECwindows it's self is the major issue. But since
there R some people that getting some difficulties implementing it (in Pascal
for example), we HAVE to help them or they'll try getting this help from 
some one else. We have to keep our customers happy, and the Pascal customers 
fill a little bit 'neglected' when DIGITAL did not provide themp some DW 
exmaples. I know every one can say that there not enough examples in Fortran,
but Fortran is VMS 'oriented' so actually it is very easy to implement almost
every thing here.

				Thanks for the replies,
					*-Doron-*

    There is not a significant difference between FORTRAN and Pascal (or
    Ada or many other languages for that matter) when it comes to doing
    DECwindows programming.  The reason there exist FORTRAN and Ada
    versions of DECburger is that I was on the Ada project and working
    on the Ada DECwindows bindings, so I decided to do DECBURGER in Ada,
    translating from the C.  That was a learning experience, where I
    "learned" how poor the DECwindows documentation was and how many bugs
    were in the toolkit.  (I also added more reasons to think C is the
    assembler of the 80s.)
    
    Then I moved to the FORTRAN project.  Taking the working Ada example
    and translating it to FORTRAN was a matter of two days.
    
    You only somewhat "learn DECwindows" from translating examples.  I
    suppose it would really take trying to develop a new application and
    beating one's head against the wall to figure out how to do things
    in DECwindows.  But if what you want to do is similar to one of the
    examples, it's a start.
    
    					Steve

Anyone who knows of customers who feel "neglected" because their 
language is not C should let DW product management know about it.  
Steve was good enough to do the Ada bindings and the examples, but 
who's going to do the Motif conversions for other languages???

DW product management needs to know that just putting out Vax 
bindings does not a product make.  Resources must be dedicated to 
creating and maintaining language bindings, examples, and reasonable 
conversions to Motif.  We'd be in deep water if Steve Lionel hadn't 
done so much work for the product, and he's not even on the team!

Bruce

When I can, I will probably do Motif versions of DECburger for Ada and
FORTRAN (and maybe C) - not because it's my job (which it isn't), but
because I think it's important.  I am worried that the Motif toolkit
interfaces will be as poorly designed as those we have for DECwindows
today, making them difficult to use from languages other than C.  But
it won't be impossible, and the examples will help.

Since access to Motif is still restricted, I can't comment on how much
of a job it will be.  I also haven't seen or heard of any SDL files for
the Motif routines.  (If there are some, I'd be happy to look at them.)

			Steve

  Take a look at note 138 in the MOTIF conference.  This has become a touchy
subject with me.

  I feel that as far as Ada is concerned the interface spec should be done
to the C routines.  What this gives you is an Ada package that can be used
on both VMS and Ultrix/Unix.  With the latter we (Digital) could provide/sell
this package interface to customers who use other hardware platforms.

  From the Motif conference note it doesn't appear that there are going to
be any 'other' language interface for quite a while.

  A friend and I are currently working on a plan to create an Ada interface
spec for OSF/Motif since we have customers who NEED to have this last month
and we feel that if something is not done VERY VERY soon that Digital is
going to lose out to other vendors who are currently working on providing
such a capability.

  Note however I'm speaking from an Ada point of view...the other languages
are on their own :^).

						-=> Jim <=-

    1) I've started doing the Ada -> Pascal DECBURGER conversion.
       I'm about 1/2 finished, but I'm out of town next week at a
       standards meeting.
    
    2) At DECUS, a customer with over 500,000 lines of Pascal/DECwindows
       code asked (actually demanded) about Motif Pascal bindings.  I
       of course gave them the DECUS answer of "Excellent suggestion.
       I'll take that back and I'm sure it will be considered."  
    
    				-John

Given the experience so far, I would say that the "VAX bindings" were
ultimately a mistake, in that they were unique to VAX implementations.  That
they were far more usable from non-C languages than C bindings would have
been is moot.

For Motif, there should be one standard set of bindings for any given
language, that are portable across OSF/Motif platforms.  Given the C blinders
that are inevitable, this means a bit more pain for high-level language
users.  But ultimately the customers will demand this, so their code is
not DEC implementation dependent.

Also, for languages such as Ada and Pascal, the bindings should take full
advantage of the strong typing and typing flexibility offered by those
languages.  The current Ada bindings still have very much of the untyped
C nature, especially with the data structures.  You lose many of the benefits
of Ada and Pascal by using all of the generic declarations (in C, everything
is an integer or a pointer).

Make no mistake about it, this is a LOT of work.  It will probably mean
separately constructed, maintained and documented packages for each language.
We need to make that commitment now so that our customers won't be 
disappointed in us again.

In the real world, C is NOT the language of choice for writing windowing
applications.  Just ask the CSC people about what languages customers most
frequently are using when they receive calls about DECwindows programming.
FORTRAN and Ada are very important, along with Pascal.  And we cannot afford
to leave behind BASIC and other languages.

Can we do it right the second time?

				Steve

	Should not someone be beating up DECwindows Product Management
	about this rather than here?

						mike

Re: .15

Working on it.  Prognosis is not good...

		Steve