Re: Ada 9X paper

Re: Ada 9X paper

[bu.edu!inmet!inmet!spock!stt]

You can get information on the Ada 9X project using anonymous
FTP from ajpo.sei.cmu.edu.  There is also an Ada 9X bulletin board, 
whose phone number is 800-ada-9x25 (800-232-9925).

S. Tucker Taft    
Ada 9X Mapping/Revision Team (ada9x-mrt@inmet.inmet.com)
Intermetrics, Inc.
Cambridge, MA  02138

Re: Ada 9X paper

[dog.ee.lbl.gov!overload.lbl.gov!agate!spool.mu.edu!sdd.hp.com!zaphod.mps.]

In article <1992Jan07.142237.9959@news.mentorg.com> samg@roguecom (Sam Griffith
 x7588) writes:
>I was wondering where I can get an Ada 9X draft paper?  Also, I took Ada
>in school about 4 years ago, and I was wondering what the job market is like
>for Ada programmers.  I have been doing OOP work for 5 years now, and most of
>it has been in C++, Obj-C and Smalltalk.
>
We're working on a small Ada project (about 50,000 lines of code).  My SW
contractor has had difficulty finding people who _really_ know Ada to work for
him. One of the questions they ask an interviewee is "What is the difference
between a derived-type and a sub-type?"  Surprisingly enough, not everyone
they interview (who supposedly "knows" Ada) can answer this question.

Apparently, most DOD Ada applications have had complexities on the order of
simply replacing Fortran with Ada.  Our application involves a distributed
arcitecture, multi-tasking, and a number of binding requirements (X-Windows
and POSIX).

We've had a number of difficulties finding an Ada compiler which actually
performs well enough to meet our needs.  Most of the Ada compilers simply
have too many bugs to be useful.  Even more "mature" compilers have demonstrate
d
enough bugs to make them worthless.  In fact, we are betting a lot that the
"latest" version of a certain compiler will have much needed features and
"fixed" bugs.  

This situation is very frustrating because my software contractor has spent
the past few month running around trying to get an Ada compiler that actually
works (or working bug fixes) rather than designing code.  I can't blame them
too much because a working Ada compiler is sort of on the critical path to
success in our development effort.

Some of the more "unique" (are they really all that unique?) things we're
now doing are:

1.  Writing TCP/IP Ada code which provide the functionality of pipes and
remote procedure calls for this distributed architecture system.  We need
to be able to connect (for example) an personal computer to a workstation.

2.  Building a subset of the Ada POSIX interface as defined in the IEEE drft
spec 1003.5 (which recently failed in balloting).  We still need these 
services and must now build them ourselves.

3.  Trying to find X-Windows bindings which work well with our design (we
have a fairly good handle on this now).

Every time we find an Ada compiler bug, we spend about two days trying to
isolate the problem and finding a robust work-around solution.

I've spent enough time complaining about Ada compiler bugs, does anybody
have any information pertaining to Ada code which provides TCP/IP interprocess
communication functions and also POSIX interface code which is available
to the DoD?

---Dan---
-- 
+ Captain Daniel F. Van Der Werken, Jr., USAF |  I do not speak for the Air
+ Rome Laboratory/OCDS                        |  Force, otherwise I'll be
+ Griffiss AFB, NY 13441                      |  in Kansas making big rocks
+ (315) 330-4441/DSN 587-4441                 |  into little rocks!

Re: Ada 9X paper

[Pat Rogers]

In article <1992Jan15.142527.5277@lonex.rl.af.mil>, vanderwerkend@lonexb.rl.af.
mil (Dan Vanderwerken) writes:

> [ ... most Ada compilers]
> have too many bugs to be useful.  Even more "mature" compilers have demonstra
ted
> enough bugs to make them worthless.  In fact, we are betting a lot that the
> "latest" version of a certain compiler will have much needed features and
> "fixed" bugs.  
> 

What target ISA are you using? I know of several compilers I would categorize
as quite mature and of high quality, but perhaps your target is one of the
more recent ones to be avialable with Ada...

We're doing a distributed Ada implementation (2/3 done) on bare 68030's
and R3000's. The compiler we use for the 68K's is one of the best I've seen 
in 12 years (OK, its early '92, so 11 years) of Ada use.  It is 
certainly as good as the compilers for the languages I used before Ada.
That is not meant to be a soapbox defense of all Ada compilers, just a
suggestion that Ada compilers are like ALL compilers -- some are better than
others, some are very good, and some just plain stink.

RE: The TCP/IP stuff, we did one for our project, it works, and I never 
want to see it again.  I strongly recommend using a vendor's implementation 
whenever possible, such as Ready's ARTX.

Regards,

Pat Rogers
SBS Engineering
Houston
progers@ajpo.sei.cmu.edu