TEAM-ADA Archives

Team Ada: Ada Programming Language Advocacy


Options: Use Classic View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Sender: "Team Ada: Ada Advocacy Issues (83 & 95)" <[log in to unmask]>
From: Michael Feldman <[log in to unmask]>
Date: Fri, 15 Nov 1996 14:15:42 -0500
In-Reply-To: <[log in to unmask]> from "Sy Wong" at Nov 15, 96 10:19:47 am
Reply-To: Michael Feldman <[log in to unmask]>
Parts/Attachments: text/plain (177 lines)
[Sy Wong wrote]
> >Sy's suggestion is more appropriate for SIGAda.
> This sounds like Washington bureaucracy or dog biting its tail.
> There is none in the entire Ada community (except one other person)
> supports the idea of a simpler version as a market invasion tool.  Not
> even in clarifying Annex-H wordings and validating compilers that is
> devoid of the restrictable constructs.  It was only after failure to gain
> SigAda action that I took up the suggestion to bypass SigAda and go to
> WG9 directly.

Perhaps, Sy, it is because for years, you have insisted narrowly that
only one solution is possible, namely, a syntactic subset of the language.
If you are having trouble  rounding up support for this idea, perhaps
it is simply that others don;t believe it's the way to go.
> >                            Some of SIGAda's working
> >groups have made valuable contributions to standards.
> There never was any SigAda working group that had worked to simplify the
> language, only add to it.  There were only papers agrandiosing how good
> is Ada versus C/C++ and calls for educating the (stupid) C/C++ users.  I
> have never seen a positive plan to meet or understand their needs.

Sy, your view is very narrow. _Syntactic_ simplifications are, in my
opinion and the opinion of most I speak to, unnecessary and undesirable.
> My objective is to introduce Ada to the following markets:
> (1) Hard real-time control systems area, often safety critical.
> (2) As HDL in lieu of VHDL (=1980 Ada plus unnecessary extensions).
> (3) Electronic Design Automation (EDA) tools.

A worthwhile objective.
> Riding on an existing ISO standard will be a one up on VHDL, which the
> members are arguing each other for a new version and with Verilog.  None
> has ISO status.

I agree.
> You, Dr. Mathis, concluded by saying the right principles.
> >To make Ada into a commercially accepted language, we have to start
> >thinking in business terms (investment, products, selling, profits) not
> >government funding and bureaucracies.
> I hope you will also carry out what you preach in practice by endorsing
> the following volunteer activities in some working group.  Why would
> anybody invest in implementing unnecessary and undesirable constructs in
> a compiler just to get validated and then be discarded by the user?  Who
> pays for the extra cost?

_You_ may deem them unnecessary and undesirable; you are evidently having
trouble finding many others who agree with you. There are alternative
approaches that you seem unwilling to consider.
> I want to point out that the LRM is ANSI/ISO/IEC-8652:1995 (Ada-95 for
> short) but the Rationale is not.  According to LRM:
> Ada-95 = essential Ada core + Annex-H restrictables
>          ------------------   ---------------------
>          |                           |
>          |----> part A               |----> part H
Right. If there is any disagreement, it is just in the definition of
> >Some of SIGAda's working
> >groups have made valuable contributions to standards.
> None seeks to clearly delineate Ada-95 into the above two parts.

The RM apparently does that already!
> The author of Annex H made excuse that he was not subsetting in the
> Rationale: "Ada compiler has to be validated for the entire language in
> any case."  This deeply entrenched idea of the community that users can
> discard any constructs they do not want is not based on economic
> realities.  This attitude certainly discourages the entry of new tools
> vendor that only caters to users for Part A, as I have listed.  Would you
> attempt to build a cheap runabout for in-town use that can be validated
> to carry 20 tons?

Maybe you are placing too much emphasis on _validation_. Why not get to
work defining what _you_ believe to be the right part A/part H definition,
try to produce a compiler, or at least a path to one, and worry about
validation later? C compilers are not (in general) validated, and C++
ones are certainly not, because there is no standard to validate against.

If there is clearly shown to be a market for your desired compiler, maybe
that will send a message that some validation process for it should exist.
> To be specific, the working group can (I may be repeating):
> (1) Edit the syntax summary, keywords and LRM pages to purge all that
>   is not in part A.

Who is defining Part A?

> (2) Separate ACVC into must-have and must-not-have for Part A.
> (3) Separating out GNAT parts that are not necessary for Part A and
>       replacing any C dependencies so that a self-standing product
>       can be implemented by tools vendors.

That is, in my opinion, foolish. Far better would be to put the effort in
on producing a back end for your desired platform(s), then cross-compiling
to it from a "normal" GNAT host. (Since GNAT is implemented already on
most plausible hosts, finding a host is no big deal. Even a DOS box
will suffice.)

> (4) Edit CAMP to remove boiler plates, especially the NO-Foreign label
>     which is no longer in force and markup Part H constructs used, if
>     any, in CAMP.  The statistics may be useful.

I agree. Sy, will you take that as your action item?
> Hal Hart had distributed a report called C41 by some panel.  I have only
> seen passive comments as though God hath spoken but no positive action
> plan from the Ada community.  First, I question the qualifications of the
> panel, many names are familiar to me as demi-gods to the computer science
> sector but unknown to the industrial sector.  By copy of this note to Mr.
> Paige, I request the background of actual Ada programming experiences
> (not just reading the LRM, if some did) of the panel.  If you want to
> comment on the Chinese language in competition with English, you should
> speak, read and write Chinese well.

I agree with you, but this is a completely separate issue. Since it will
be only a short time before the full report appears on the web, I suggest
that we not waste time commenting on summaries, and simply read the full
report. There will be PLENTY to talk about then.:-)

> Also to the Ada-team, I ask those
> having interest in the three applications areas I mentioned to contact
> me.  I specifically want to talk to persons that has designed integrated
> circuits, have used VHDL or had read IEEE-1076, and had programmed Ada
> for practical applications.  So far I have not found one as yet in the
> U.S.

I can't help you there, Sy. I agree with your basic desires, but not with
your suggested solution.
> By copy to Hal Hart:
> What are the qualifications used to select the Ada-team list and
> who compiled the list?

Team-Ada is nothing but an electronic mailing list; anyone can join
by sending a simple subscribe-me message to the listproc program
at ACM. The syntax escapes me; maybe someone else will post the subscribe

You are mistaken if you think someone has _chosen_ a Team-Ada. This is
NOTHING but a group of Ada fans who choose to correspond with each other
on advocacy issues.
> I welcome specific criticisms of my proposed working group efforts and
> reasoned alternatives to foster broader Ada usage in the three listed
> industrial areas.  Hopefully the present and future WG9 convener, both
> Americans, will help my lone effort to see Ada come to life in new areas.

Actually, I think you are better off to formulate your suggestions
_specifically_ and _constructively_ and send them to this list. We are
all here on this list because we want to be here. The list generates a
lot of mail, so I look forward to seeing your suggestions put into
specific terms, instead of complaints. You might find support here,
if we can understand exactly what you are proposing, and if you are
willing to be openminded in reading our replies.

> Please, no ATNA.
 Sorry, I don;t understand ATNA. Meaning?
> SY Wong, Tarzana CA.
Mike Feldman