Pseudo-Adas for PCs

Pseudo-Adas for PCs

[Stavros Macrakis]

It's interesting to hear of the availability of Ada for PC's.

Unfortunately, it appears that none yet exists.  Janus/Ada was cited as
lacking "tasking, generics, exception handling, multi-dimensional arrays
(ouch!), Ada standard strings, operator overloading."  I threw away the
brochure they sent me because the product was so incomplete, but if I
recall correctly, it also lacks discriminants, enumeral overloading,
aggregates, named parameters, and several other features.

This, gentlemen, is not Ada.

> A company called ALSYS is also supposed to be comming out with
> an Ada compiler which compiles something near full Ada.
	Alsys, by the way, is Jean Ichbiah's company.  (Ichbiah was
	Ada's chief designer.)

>     Commentary
>     Ada is a huge language and I  do  not  believe  that  it
>     would  be  viable  without  all of those DoD giga-bucks.

I think it's commendable that the DoD has realized that large
investments in software technology are necessary.  Actually, the
investments so far are tiny compared to the amount they spend on
programming.

>     Although Ada is not without its good features, there  is
>     no  excuse  for  its size.  In my opinion Modula-2 would
>     win hands down over Ada but for two things:
>       1. There are only a few  Modula  compilers  available....

What do you think that DoD money is being spent for!?  Perhaps they
should have spent it on Modula instead (I don't think so), but support
for a language ends up being as important as its technical qualities--
look at the continued success of Fortran, Basic, and C.  It is rational
to compromise technical quality for support.  (And of course I continue
to believe that Ada has the technical quality as well.)

>       2. Modula is not a well standardized language  as  Ada....

Why do you think it took so long to complete the Ada language design?

>     I  think  that  Ada has taken block structured languages
>     about as far as they can go.  I think that the  meaning-
>     ful   language   research  will  concentrate  on  object
>     oriented languages (e.g., offshoots  of  SmallTalk)  and
>     data flow languages.
>                      Ian Kaplan, Loral Data Flow Group

Yes, well, read up a bit on Ada and see how such features as packages
and tasking and overloading really give you object orientation.

	-s

Ada cheap shots

[Dick Dunn]

> ...  Janus/Ada was cited as
> lacking "tasking, generics, exception handling, multi-dimensional arrays
> (ouch!), Ada standard strings, operator overloading."  I threw away the
> brochure they sent me because the product was so incomplete, but if I
> recall correctly, it also lacks discriminants, enumeral overloading,
> aggregates, named parameters, and several other features.
> 
> This, gentlemen, is not Ada.

Agree.  I've always been irritated by the fact that DoD has refused to hear
of the idea of subsetting Ada (particularly in view of its size), but when
I read of this nonsense of Ada-without-Ada, I have to admit, however
grudgingly, that there's a point for the no-subsets rule.

> >       2. Modula is not a well standardized language  as  Ada....
> 
> Why do you think it took so long to complete the Ada language design?

Well, it might be because they started by trying to standardize it, then
produce full implementations, and when they're done they'll go about
getting some user experience...some tasks take much longer when you do them
completely backwards!

> >     I  think  that  Ada has taken block structured languages
> >     about as far as they can go.  I think that the  meaning-
> >     ful   language   research  will  concentrate  on  object
> >     oriented languages (e.g., offshoots  of  SmallTalk)...
> ...
> Yes, well, read up a bit on Ada and see how such features as packages
> and tasking and overloading really give you object orientation.

I agree with > > that Ada is a point near the end of the path of block-
structured languages.  (I think that ALGOL 68 took them much further
than Ada has, in some ways.)  However, trying to regard SmallTalk as either
truly object-oriented (semantic quibble possible here) or a radically
different future direction invites the response > that Ada offers almost as
much.  Frankly, most of what SmallTalk offers is cosmetic, and its little
game of active data / passive code (data sends messages) is an annoying
renaming of traditional concepts that smacks of the "Emperor's New
Clothes".  In other words, (1) Ada is at the end of the overdeveloped
algorithmic languages, but (2) SmallTalk isn't the new direction we need.
Backus' Turing lecture of several years back provides much more insight.
-- 
Dick Dunn	{hao,ucbvax,allegra}!nbires!rcd		(303)444-5710 x3086
   ...A friend of the devil is a friend of mine.