Conference turris::languages

362.0. "Ada9X at OOPSLA" by BEGIN::BRETT () Thu Oct 07 1993 20:04

From:	TLE::US2RMC::"[email protected]"  7-OCT-1993 18:07:50.03
To:	[email protected]
CC:	
Subj:	Re: Ada9X at OOPSLA

I can not give a report on the tutorials, as I did not attend
any of them.  However, the reception of Ada 9X from the exhibition
floor was very encouraging.  There were a little more than 150
requests for the LRM and Rationale.  John Barnes sold several
copies of the fourth edition of his book, which contains a chapter
on Ada 9X as well as indicators of changes in Ada 83.  There were
several other Ada exhibitors too...EVB and Rational and ??.  The exhibition
included many training courses in object oriented methodologies such
as the Coad approach.  There were Eiffel booths, SmallTalk booths,
book vendors, and C++ booths.

From some of the presentations at OOPSLA it seemed that Ada's
real threat is not C++ but rather SmallTalk.  One of the comments
regarding C++ was that C remained encapsulated within the language
so although C++ offers a stronger type model, it is still fairly
easy to access the loose type model of C.  This was viewed as
a risk to the development of system software.  Tucker's presentation
was well received...and Tucker surprised the crowd with a description
of Ada 9X's support for concurrent OOP.  He showed that by using an
access to T'class as a discriminant to a task or a protected type
tagged objects could be passed to tasks and protected objects as 
descriminants.  Hence, within the task and protected body the 
class-wide operations could be accessed and dispatching could 
take place when necessary.

The Washington DC SIGAda held their montly meeting in conjunction
with the OOPLSA conference and welcomed conference attendees to
join the meeting.  About 10 non-Ada folks dropped by to listen
to the panel session on early experiences with Ada 9X.  The major
issue raised by one of the visitors was the lack of support for
highly parallel architectures in Ada 9X.  Erhard did an excellent
job in pointing out that while Ada 9X does not offer any direct
support for highly parallel architectures, the language has been
designed with concurrency and parallelism.  Unlike other languages
which are strictly sequential.

One additional observation from this conference was the attitude of
the attendees.  Unlike Ada conferences where attendees stand in the
hallways complaining that their favorite feature did not get into to
Ada 9X or that they are so irritated with this or that feature in Ada,
the attendees at OOPSLA took a noticeably different approach.  Instead
of complaining that C++ was missing this or that favorite feature or
did not do exactly what they thought it should do, they were effectively
bragging about the neat work around they came up with to get the 
desired affect.  Generally speaking, the crowd was more 'up-beat' than
what I have seen at various Ada conferences.  


% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us2rmc.bb.dec.com; id AA04343; Thu, 7 Oct 93 18:03:23 -0400
% Received: by inet-gw-1.pa.dec.com; id AA26553; Thu, 7 Oct 93 15:05:36 -0700
% Received: from TLSUND0.TARTAN.COM by ajpo.sei.cmu.edu (5.65/2.5) id AA01466; Thu, 7 Oct 93 18:01:17 -040
% Received: from tlsund11.tartan.com ([192.5.56.71]) by tartan.com (4.1/SMI-4.0) id AA24606; Thu, 7 Oct 93 17:59:16 ED
% Received: by tlsund11.tartan.com (4.1/SMI-4.1) id AA26343; Thu, 7 Oct 93 17:59:13 ED
% Date: Thu, 7 Oct 93 17:59:13 EDT
% From: [email protected]
% Message-Id: <[email protected]>
% To: [email protected]
% Subject: Re: Ada9X at OOPSLA

From:	TLE::US2RMC::"[email protected]" "Tucker Taft"  7-OCT-1993 18:40:00.10
To:	[email protected], [email protected]
CC:	
Subj:	Re: Ada9X at OOPSLA

I don't see SmallTalk as a threat, but rather as an opportunity.
We could position Ada as a follow-on for a system initially
built in SmallTalk, rather than as a replacement for SmallTalk.
That is, SmallTalk would be the prototyping language, and Ada
would be the production language.  Of course this means transition/
translation tools would be important, but I believe that is a
solvable problem (SmallTalk only has single inheritance ;-).
-Tuck

% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us2rmc.bb.dec.com; id AA05367; Thu, 7 Oct 93 18:35:28 -0400
% Received: by inet-gw-1.pa.dec.com; id AA29717; Thu, 7 Oct 93 15:37:41 -0700
% Received: from dsd.camb.inmet.com by ajpo.sei.cmu.edu (5.65/2.5) id AA01904; Thu, 7 Oct 93 18:33:49 -040
% Received: from spock.camb.inmet.com.camb.inmet.com by dsd.camb.inmet.com (4.1/SMI-4.1) id AA20865; Thu, 7 Oct 93 18:31:13 ED
% Date: Thu, 7 Oct 93 18:31:13 EDT
% From: [email protected] (Tucker Taft)
% Message-Id: <[email protected]>
% To: [email protected], [email protected]
% Subject: Re: Ada9X at OOPSLA

From:	TLE::US2RMC::"[email protected]"  7-OCT-1993 18:51:06.04
To:	[email protected], [email protected]
CC:	
Subj:	Re: Ada9X at OOPSLA


> I don't see SmallTalk as a threat, but rather as an opportunity.
> We could position Ada as a follow-on for a system initially
> built in SmallTalk, rather than as a replacement for SmallTalk.
> That is, SmallTalk would be the prototyping language, and Ada
> would be the production language.  Of course this means transition/
> translation tools would be important, but I believe that is a
> solvable problem (SmallTalk only has single inheritance ;-).

This opportunity may extend even farther than SmallTalk.  There
was some interest in the multilanguage environments of the future.
The opportunity for 9X is to provide the glue between the various
OO-worlds as well as the transition tools.

> -Tuck

JT

% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us2rmc.bb.dec.com; id AA05659; Thu, 7 Oct 93 18:46:39 -0400
% Received: by inet-gw-1.pa.dec.com; id AA00817; Thu, 7 Oct 93 15:48:52 -0700
% Received: from TLSUND0.TARTAN.COM by ajpo.sei.cmu.edu (5.65/2.5) id AA01976; Thu, 7 Oct 93 18:45:09 -040
% Received: from tlsund11.tartan.com ([192.5.56.71]) by tartan.com (4.1/SMI-4.0) id AA28010; Thu, 7 Oct 93 18:42:52 ED
% Received: by tlsund11.tartan.com (4.1/SMI-4.1) id AA26393; Thu, 7 Oct 93 18:42:50 ED
% Date: Thu, 7 Oct 93 18:42:50 EDT
% From: [email protected]
% Message-Id: <[email protected]>
% To: [email protected], [email protected]
% Subject: Re: Ada9X at OOPSLA
    

From:	TLE::US2RMC::"[email protected]" "Erhard
Ploedereder" 12-OCT-1993 01:42:29.50
To:	[email protected]
CC:	
Subj:	Ada9X at OOPSLA

A couple of additional observations....

on C++ at the OOPSLA Educator Symposium:  in a nutshell, very few attendees
liked C++, most actually expressed considerable distaste, but most still taught
it "because industry wanted it". Smalltalk was a clear favorite.

on an OOPSLA panel on multiple inheritance: the uniform opinion of the 
panelists and apparently of the audience, judging from the absence of
objections:  "we don't need it". A split opinion on "it's still nice to
have" vs. "it's actively dangerous" (minority).

General impression of trends: considerable negative vibes on C++; Eiffel
nice, but not practical; Smalltalk is what we like. Interest in Ada 9X is
not exactly rampant, but then neither is a dislike of Ada. Still, Ada is seen
as a niche player and most people can barely spell it.

Personal opinion: this is a market that would be ripe for a serious dose of
pushing Ada 9X as a viable alternative to a disliked C++, a weakly-typed,
interpretative Smalltalk, and an insufficiently supported Eiffel. It is a
market in search of a good language, for sure. Who's going to do it and
what's the line ?

Erhard
--------------------------------------------------------------------------------
From: [email protected] (Debora Weber-Wulff)
Newsgroups: comp.lang.ada
Subject: Ada at Systems'93 in Munich
Date: 11 Oct 93 15:51:56 GMT
Sender: [email protected] (Math Department)
Organization: Free University of Berlin

Here I was, looking through the Systems listings so I could be just
like Greg and complain that there was no Ada at the shows, and what
did I see?

Alsys is showing Ada Native Software Development environment and crosscompiler;

GSE is showing Tartan Ada, SunPro Ada and Verdix Ada;

And a little house, Dr. Rudolf Keil Gmbh, is showing a validated ADA-compiler
for the OS-9 operating system (whatever that is!)

6 Ada things. Five companies paid to advertise their C or C++ things. So either
you don't have to advertise for C anymore, or else Ada is actually starting to
get some respect!

--
Debora Weber-Wulff, Professorin fuer Softwaretechnik und
snail: Technische Fachhochschule Berlin, FB Informatik,
       Luxemburgerstr. 10, 13353 Berlin, Germany
email: [email protected]

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