DEC Ada: why is it so good? 
Author Message
 DEC Ada: why is it so good?

Everything in my experience and what I read in this group
and what I have heard from other people confirm how solid
the DEC Ada compiler is.

What happened that that compiler is so good, while others
are well, not so good. (The recent post about aggregates
comes to mind).

I can think of a reason of large user base, maturing the
product faster than other, lesser used products. Is that
true?

What other reasons are there?

Alex Cellarius
--
Opinions mine



Wed, 22 Jan 1997 04:21:47 GMT  
 DEC Ada: why is it so good?

|>
|>Everything in my experience and what I read in this group
|>and what I have heard from other people confirm how solid
|>the DEC Ada compiler is.
|>
|>What happened that that compiler is so good, while others
|>are well, not so good. (The recent post about aggregates
|>comes to mind).

DEC was fairly late in getting their compiler to market.  We were all
surprised that the vast majority of Ada compilers available in the first
few years were ALL on the Vax (or had a version for the Vax), but DEC
had NOTHING.

Then the DEC Ada compiler comes out and it is fantastic!

The rumor I heard was they threw out their first attempt and made a really
solid product by doing it again.  I don't know if this is true, but I have
always wondered if that would be a good strategy for designing any software.
Make an almost complete prototype first and then redo it.  Maybe then we
wouldn't have fiascos like the C++ code that drives the Denver baggage
handling system.

|>I can think of a reason of large user base, maturing the
|>product faster than other, lesser used products. Is that
|>true?

Probably has a lot to do with it.

Wayne.



Wed, 22 Jan 1997 07:45:38 GMT  
 DEC Ada: why is it so good?

Quote:

>DEC was fairly late in getting their compiler to market.  We were all
>surprised that the vast majority of Ada compilers available in the first
>few years were ALL on the Vax (or had a version for the Vax), but DEC
>had NOTHING.

>Then the DEC Ada compiler comes out and it is fantastic!

What is early and what is late?

Back in 1985 I was working at GTE Gov. Systems in Mountain View, CA.
We were a beta test site for DEC Ada version 0.8 and 0.9. It was okay. Had
some bugs, but then it was beta. I believe 1.0 came
out in late '85 and was validated. Still had some bugs. By the time v1.5 (1986?)
came out, the compiler was solid, but slow. We had some sequencing problems
with ASTs being delivered to task entries, but other than that we were pleased.
When v2.0 came out, task rondezvous were being sped up as were exception
propagation
times. DEC began to talk about shared generic object modules. This was in the
1987(?)
time frame.

If I recall correctly, the FIRST validated compilers got their validations
some time in 1985. Thus, it seems DEC wasn't that far behind.

I always figured DEC bought in on Ada going big and focused thier considerable
engineering resources (at least in those days) into producing
a top notch product. Also, the back end of their compiler was the optimizer
already inuse in their C, Pascal, and other offerings. The back end
was well tested and optimized considerably. Thus, DEC didn't have to
start from scratch.

Just thought I'd clear out some cob webs and revisit some fond old memories...

        -- ART STADEN
           New member of Team Ada



Wed, 22 Jan 1997 14:44:49 GMT  
 DEC Ada: why is it so good?

Quote:

>If I recall correctly, the FIRST validated compilers got their validations
>some time in 1985. Thus, it seems DEC wasn't that far behind.

The first few validations came in 1983. NYU-Ada/Ed was first, with the
original SETL implementation. Then came ROLM, Data General, and DEC,
(I am not sure if this order is correct). But 1983 is not "late" by any means
- the standard was adopted in Jan. 1983.

I think we need an Ada history course somewhere. :-)

Mike Feldman
------------------------------------------------------------------------
Michael B. Feldman -  chair, SIGAda Education Working Group
Professor, Dept. of Electrical Engineering and Computer Science
The George Washington University -  Washington, DC 20052 USA

"Pork is all that stuff the government gives the other guys."
------------------------------------------------------------------------



Thu, 23 Jan 1997 23:43:24 GMT  
 DEC Ada: why is it so good?

Quote:

>I think we need an Ada history course somewhere. :-)

Good idea.  If somebody does do that, please don't forget to mention
BiiN (Billions invested in Nothing)

:-)

--
Proud (and vocal) member of Team Ada! (and Team OS/2)        ||This is not your
<Witty, but politically correct, comment under construction> ||  father's Ada
For all sorts of interesting Ada 9X tidbits, run the command:||________________



Fri, 24 Jan 1997 01:21:40 GMT  
 DEC Ada: why is it so good?

|>

|>>
|>>If I recall correctly, the FIRST validated compilers got their validations
|>>some time in 1985. Thus, it seems DEC wasn't that far behind.
|>>
|>The first few validations came in 1983. NYU-Ada/Ed was first, with the
|>original SETL implementation. Then came ROLM, Data General, and DEC,
|>(I am not sure if this order is correct). But 1983 is not "late" by any means
                                                ^^^^^^^^^^^^^^^^^^
|>- the standard was adopted in Jan. 1983.
|>
|>I think we need an Ada history course somewhere. :-)

Yes, perhaps we do.  First, I didn't write that (please use care on the
attributions).  Second, DEC did not validate in 1983 as you implied.  By
the time DEC validated (in late '85 or early '86 as I recall) there were
several Ada compilers available on the Vax from other companies.  It simply
appeared (in 1985) that DEC was a little behind.

By the way, Jan '83 was when the ANSI standard was adopted.  I believe the
Mil-Std was in December of 1980, so it took DEC five years after the first
standard.  There was definitely the perception that Ada compilers in general
were slow to appear.  I remember this well, and I believe it had a large
impact on why C, instead of Ada, dominates today.

Wayne.



Fri, 24 Jan 1997 05:10:38 GMT  
 DEC Ada: why is it so good?

|>

|>>
|>>I think we need an Ada history course somewhere. :-)
|>>
|>
|>Good idea.  If somebody does do that, please don't forget to mention
|>BiiN (Billions invested in Nothing)

Can you elaborate on that?  Surely, you are not referring to Ada.  Ada was,
IMHO, one of the best investments the government has made in recent years.

Wayne.



Fri, 24 Jan 1997 05:13:04 GMT  
 DEC Ada: why is it so good?

Quote:


>|>


>|>>
>|>>I think we need an Ada history course somewhere. :-)
>|>>
>|>
>|>Good idea.  If somebody does do that, please don't forget to mention
>|>BiiN (Billions invested in Nothing)

>Can you elaborate on that?  Surely, you are not referring to Ada.  Ada was,
>IMHO, one of the best investments the government has made in recent years.

Sure.  I wouldn't want to be associated with other "ranters" in this
newsgroup :-)  I can understand how Wayne might misinterpret my post.
 The "Billions invested in Nothing" was a play on the company name:
BiiN.  That company had an Ada-based hardware system, if I recall correctly,
that was almost, but not quite, fielded.  I don't have any of my old
Ada propaganda with me, but I remember BiiN making a big push
sometime around 1988, before it disappeared in obscurity.  

I don't think the guy who thought up "Team Ada" would consider the Ada
language as "billions invested in nothing"!   I certainly agree with
your last sentence!

--
Proud (and vocal) member of Team Ada! (and Team OS/2)        ||This is not your
<Witty, but politically correct, comment under construction> ||  father's Ada
For all sorts of interesting Ada 9X tidbits, run the command:||________________



Fri, 24 Jan 1997 06:19:58 GMT  
 DEC Ada: why is it so good?

Quote:



>>If I recall correctly, the FIRST validated compilers got their validations
>>some time in 1985. Thus, it seems DEC wasn't that far behind.

>The first few validations came in 1983. NYU-Ada/Ed was first, with the
>original SETL implementation. Then came ROLM, Data General, and DEC,
>(I am not sure if this order is correct). But 1983 is not "late" by any means
>- the standard was adopted in Jan. 1983.

>I think we need an Ada history course somewhere. :-)

>Mike Feldman
>------------------------------------------------------------------------

The guy in the next office to mine has been in the validation business since
the beginning (well, almost the beginning).  His recollection is that the
first Telesoft validation was at the same time as the Ada/Ed validation --
he was in California while his supervisor was doing Ada/Ed remotely from
the Boston area.  This was in mid '84.  I think it possible that this was
the second cycle; that's back when the ACVC had a new release every 6 months.
If we really want to know the history, John Goodenough was intimately
involved in validation at the time, and I'll be he can tell us.

Phil Brashear



Fri, 24 Jan 1997 16:15:47 GMT  
 DEC Ada: why is it so good?

Quote:

>The first few validations came in 1983. NYU-Ada/Ed was first, with the
>original SETL implementation. Then came ROLM, Data General, and DEC,
>(I am not sure if this order is correct). But 1983 is not "late" by any means
>- the standard was adopted in Jan. 1983.

>I think we need an Ada history course somewhere. :-)

>Mike Feldman

I recall that the validations were:
  1) NYU
  2) Rolm/Data General
  3) Western Digital (group was spun-off and renamed... name escapes me,
     but I think it eventually became Incremental Systems)
  4) TeleSoft/LabTek  (This was May 21, 1994)

Anyone else have others and dates?

BTW, another posting indicated that the ANSI standard was approved in
January of 1993.  I seem to remember a version coming out in January,
(perhaps a release by the ACM), but one of my copies states:
  "Approved February 17, 1983"

-Tom



Fri, 24 Jan 1997 21:40:53 GMT  
 DEC Ada: why is it so good?

Quote:

>Good idea.  If somebody does do that, please don't forget to mention
>BiiN (Billions invested in Nothing)

And don't forget the Intel-432 either!

However, surely the prize for the greatest waste of money and
resources in pursuit of that ignis fatuus, the high level
language machine, must go to Tron.



Fri, 24 Jan 1997 17:29:43 GMT  
 DEC Ada: why is it so good?

|>
|>I recall that the validations were:
|>  1) NYU
|>  2) Rolm/Data General
|>  3) Western Digital (group was spun-off and renamed... name escapes me,
|>     but I think it eventually became Incremental Systems)
|>  4) TeleSoft/LabTek  (This was May 21, 1994 [should be 1984?])
|>
|>Anyone else have others and dates?

That sounds about right to me.  Also, Alsys and DDC weren't far behind those.
I and another engineer wrote a code generator for the DDC compiler in 84/85
for some obscure 36-bit mainframe that no one has ever heard of.  We validated
in 1985 and I don't believe DEC validated before we did.  Things started to
really take off about then.

Wayne.



Fri, 24 Jan 1997 22:25:05 GMT  
 
 [ 155 post ]  Go to page: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

 Relevant Pages 

1. DEC Ada 3.3/DEC Unix Memory leak

2. INFO-ADA Digest - 13 Dec 1997 to 14 Dec 1997

3. INFO-ADA Digest - 3 Dec 1996 to 4 Dec 1996 - Special issue

4. DEC support for Ada 95 (was Re: DEC Fortran 90 for OpenVMS

5. INFO-ADA Digest - 26 Dec 1993 to 27 Dec 1993

6. DEC Ada/ DEC tools needed

7. Ada History (was Re: DEC Ada: why is it so good?)

8. Why Ada is failing - comp.lang.ada #4245 - comp.lang.ada #4273

9. Summary of port of DEC Ada Motif binding on Sun with Alsys Ada

10. Oracle/Pro*Ada/LSE/DEC Ada utility

11. DEC Ada run-time error messages (Was: More Ashamedness of Ada - this time OBJECT WORLD)

12. Why Python is like BASIC (and why this is a good thing)

 

 
Powered by phpBB® Forum Software