Re: PL/I Visibility.

Robert,

I would suggest that topics regarding the AS/400 be sent also to the
midrange-l@midrange.com listserver.  

We have a fairly decent cross section of users and IBM'ers on this list
and I'm also cc'ing this reply with your original to the m-l list...

Frankly, I'm one of the probably few folks left that thinks PL/I was and
should still be (if the damn standards folks hadn't abandoned it) one of
the best languages out there.  

Don in DC metro

On Sun, 2 Nov 1997, Robert Barnes wrote:

>       Perhaps, but I think that PL/I has become more visible in recent years.
> As one piece of evidence of this process, the number of PL/I positions
> available has been on the increase -- as is apparent from postings
> on the newsgroup.
> I'm sceptical that this is any more than Y2K panic starting to bite.  
>Situations
> where companies are doing real, new, development in PL/I are few and far
> between.  Show me a company (even a single one would do!) that is using
> PL/I for development for anything other than legacy-related reasons.
>         ...
> By promoting it at any opportunity.
> But promotion has to be realistic to be credible.  For example, you say:-
>       I can see no technical reason for using C over PL/I.  There's
> so much lacking in that language . . .
> An inability to coexist easily with the GUI toolkits is one VERY compelling
> technical reason. At one stage, for example, we were doing development with
> OLIT (Open Look Intrinsics Toolkit). The interfaces to OLIT were all described
> with C header files.  How would you interface this with PL/I?  Remember that
> OLIT, like every other bit of systems software, would be coming out with later
> releases, and unless we used the vendor-supplied interface descriptions, we
> wouldn't have any forward compatibility.
>          and then C++.  Besides, the language
> [I think you mean the implementation]
> While I'm well aware of the distinctions between a language and its implement-
> ation, PL/I (or any other language) is only as good as the implementation that
> you have to work with, so any distinction would be academic.  In any event, I
> meant both "Language" and "Implementation" in this context:  the PL/I that
> was available (and is still all that is available in UNIX, except for AIX) is 
>a
> subset G implementation, which is not really very exciting.
> 
> My comment:-
>           Whereas there were strong
>         technical arguments in favour of PL/I in the MVS-type environment, 
>and in
>         particular there were some types of tasks that were very difficult to
>         achieve with any other high-level language ....
> relates both to the language, and to the available implementations.  You have 
>to be
> pretty determined to use a "Non standard" language in any environment - ie a 
>language
> not solidly supported by IBM when you're working in an IBM environment, SUN 
>in a
> SUN environment, etc.
> 
> I think that what's offering is decidedly better than 10-15 years ago.
> As far as AS/400 is concerned, ten years ago PL/I was not available at all,
> on its predecessor, the S/34-S/36-S/38, so in a narrow sense you're right.
> "Better than no PL/I at all" however is not good enough.  I received this 
>comment
> in a response to my posting:-
>         >Finally, "PL/I" for the AS/400 is a joke, and a bad joke at that.
>         >It's not even subset G, and subset G is a pretty tiny slice of IBM
>         >(or VAX) PL/I.  For any application that actually exploits PL/I, the
>         >AS/400 PL/I environment is far too frail.  And the AS/400 PL/I
>         >compiler has assumed room temperature; you'll never see another
>         >enhancement.
> Hardly encouraging, wouldn't you say?
>          Should I promote PL/I and Visual Age?
> If you have used them, why not . . . ?
> I haven't used Visual Age at all, and personal PL/I experience doesn't come 
>into it as the
> question is whether I should bear the cost of educating staff in the 
>environment, or should
> [my team] use Visual Basic, Delphi, etc.
> 
> As with C and OLIT, the key question is tools support.  For example, I am 
>looking at a
> product that provides a two way CASE tool to analyse our present database, 
>showing its
> structure graphically, and allows us to both to modify it, and to export it 
>to other platforms.
> It includes features like the ability to generate Visual Basic programs to 
>manipulate our
> AS/400 database, but it does not include any PL/I. As long as tools such as 
>this continue
> to ignore PL/I, I have no choice but to do likewise.
> 
> If you'll pardon my bluntness, advocacy of PL/I is just not tenable in the 
>real world where
> interfaces, productivity, portability, support, available skills, etc have to 
>be considered.
> 
> So, what is there to promote?  I continue with the view that IBM's PL/I 
>effort is far too little,
> far too late. I wish I didn't believe this!
> 
> Rgrds, Robert.
> 

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to "MIDRANGE-L@midrange.com".
| To unsubscribe from this list send email to MAJORDOMO@midrange.com
|    and specify 'unsubscribe MIDRANGE-L' in the body of your message.
| Questions should be directed to the list owner/operator: david@midrange.com
+---