Re: No top schools use Ada

Re: No top schools use Ada

[Ian S Nelson]

Excerpts from mail: 19-Apr-95 Re: No top schools use Ada by Robert
Dewar@GNAT.CS.NYU 
> if computer scientists and schools teaching computer science are happy to
> make it clear to their students, and the prospective employers of their
> students that computer science is quite different from software engineering
> and that they should look elsewhere if their goal is to become or hire
> competent programmers, that would be helpful!

Most certainly... In fact I wished they told me that before I got here..
 Actually, CMU is a pretty good SE school too, you just have to put up
with the boring stuff (prog languages 15-312, compilers, OSes, etc..)
before you can take the SE classes. I don't know hardly anyone who
wishes to be a "computer scientist" they stress proofs and functional
programming too much, anything useful is usually far too large to prove
and functional programming just isn't fun. ("hmmm, let's use a lot of
recursion to make this hard to read, but we came *prove it*") I don't
know how many projects I've seen come out of the labs in LISP and the
first thing done to them by the programmers is to convert it into C++,
it's almost like it is part of the scientific design process.
 
Anyways, why would MIT, CMU and Stanford say that they don't produce
software engineers and lose all that money they make?  
--
Ian S. Nelson  <bonovox@cmu.edu>                  finger for PGP key
Carnegie Mellon Computer Science/Math    
Home Page:http://www.contrib.andrew.cmu.edu/usr/in22/ian.html
My opinions are not the school's, although they should be!

Re: No top schools use Ada

[Jay M Martin]

> Actually, CMU is a pretty good SE school too,.....

Actually I thought that CMU was THE preeminent SE school!  Lets see,
its at the very top of the ratings and they actually care about SE
(one of the few!).  Looking at MIT,Cornell and Stanford, no software
engineering professors, gee I wonder what they think of SE.  My school
is in the top 12 and it has no software engineering professors.  In
fact, in the past a very famous SE professor was given the boot on the
basis that "SE is not worthy of study".  To many CS professors for
example feel that "modules" are not important CS concepts, this is
because modules are for making modular code which is important for
large projects.  But large projects are in the domain of Software
Engineering, which of course is not worthy of study by Computer
Science.  No wonder Ada and other SE languages don't have a chance 
in academia.

Jay

Re: No top schools use Ada

[Michael Sperber [Mr. Preprocessor]]

>>>>> "Robert" == Robert P Goldman <goldman@src.honeywell.com> writes:

Robert> [ ... ]

Robert> But I'm curious --- and let me just offer the futile wish that I can
Robert> do this without starting a religious war --- given the situation in
Robert> the languages world now, are there reasons aside from textbook,
Robert> etc. (that are peripheral to the language itself) for using Scheme
Robert> over ML as an intro programming language?

I regularly hack in both Scheme and SML.  SML has - as both a teaching
and, in some instances, as a programming language - some serious
shortcomings, IMHO:

- parametric polymorphism is great, but heavy on newbies

- type inference often separates the location of the cause of a typing
  bug, and the location of an error message which makes often it hard
  for newbies (and for me ...) to see what caused an error

- SML semantics has some quirks which are almost impossible to grok if
  you don't have a formal background already (such as the
  much-discussed problem with val ... and ...)

Having said that, Scheme is a much more lightweight language which
does not sick as much baggage about syntax, a typing system, currying
etc. on the student.  Admittedly, it's got serious shortcomings as a
development language.  However, everyone who knows Scheme, has, in my
experience, very little trouble dealing with SML.


-- 
Cheers =8-} Chipsy

Re: No top schools use Ada

[Ian S Nelson]

Excerpts from netnews.comp.lang.ada: 16-Apr-95 Re: No top schools use
Ada by Robert Dewar@cs.nyu.edu 
> Following up on Kevin's comment, yes indeed, it would be interesting to
> see an argument that teaching scheme is a valid way to introduce people
> to what software engineering is all about.

I think it would also be interesting to see an argument that teaching
ada is a valid was to introduce people to _computer science_
CS and SE are different you know.  I better topic would be to see what
the top SE schools teach in their first SE classes.
--
Ian S. Nelson  <bonovox@cmu.edu>                  finger for PGP key
Carnegie Mellon Computer Science/Math    
Home Page:http://www.contrib.andrew.cmu.edu/usr/in22/ian.html
My opinions are not the school's, although they should be!

Re: No top schools use Ada

[Robert Dewar]

Regarding using Ada to teach computer science

 this actually works very well, I am teaching the equivalent of CS2 (i.e.
 data structures and algorithms), and there are lots of advantages, most
 notably the completely checked environment of Ada, over say C or C++.

 I can't see *any* advantage in using C++ over Ada 95 in teaching computer
 science or SE -- can you really make a case for that? Usually the choice
 is based simply on commercial popularity rather than any pedagogical
 argument (that is in those few cases where the people arguing know both
 languages well).

 Scheme is a different issue: to me it is limiting to live only in the
 functional world. For example, today I was presenting heapsort. The
 essence of this algorithm is the in place tree mapping, that's what
 makes the algorithm interesting.

 Same thing with quicksort. I have actually seen people teaching in a
 functional style and presenting a two way division sort as quicksort,
 missing the most interesting part of the algorithm, which is the
 in place partitioning.

 Yes, I know you can model this functionally, but it seems confusing to
 me to do this.

P.S. I plead guilty to subversive activities in my course, like trying
to teach students about abstraction, reuse and modularity. Hopefully
this will not *too* seriously warp their capability to do good work in
computer science.

Re: No top schools use Ada

[Brian Hanson]

In article 798344756@gnat, dewar@cs.nyu.edu (Robert Dewar) writes:
> Regarding using Ada to teach computer science
>  Scheme is a different issue: to me it is limiting to live only in the
>  functional world. For example, today I was presenting heapsort. The
>  essence of this algorithm is the in place tree mapping, that's what
>  makes the algorithm interesting.
Huh?  Over in the comp.lang.scheme there have been long drawn out arguments
about how scheme is not really a functional language as it is not a pure
functional language.

I have written a number of significant programs in scheme and never
really written in functional style.

Scheme is a lean and elegant lisp.

;; The following sorts are short but horribly inefficient

(define (insertion-sort a)
  (do ((i 1 (+ i 1)))
      ((>= i (vector-length a)) a)
    (do	((v (vector-ref a i) v)
	 (j i (- j 1)))
      ((or (< j 1) (< (vector-ref a (- j 1)) v))
       (vector-set! a j v))
      (vector-set! a j (vector-ref a (- j 1))))))

(define (swap! a i j)
  (let ((v (vector-ref a i)))
    (vector-set! a i (vector-ref a j))
    (vector-set! a j v)))

(define (bubble-sort a)
  (do ((i (- (vector-length a) 1) (- i 1)))
      ((< i 0) a)
    (do ((j 1 (+ j 1)))
	((> j i))
      (if (> (vector-ref a (- j 1)) (vector-ref a j))
	  (swap! a j (- j 1))))))

(define l (list 1 5 7 8 2 4 1 5 6 2))

(define (run sort a)
  (sort a)
  (display a)
  (newline))

(run insertion-sort (list->vector l))
(run bubble-sort (list->vector l))

$ vscm
> (load "../src/scm/sort.scm")
#(1 1 2 2 4 5 5 6 7 8)
#(1 1 2 2 4 5 5 6 7 8)
;value (after 170ms... comp: 0 (0 GC), exec: 170 (0 GC)):
#t
> 

-- 
-- Brian Hanson
-- brh@cray.com

Re: No top schools use Ada

[Robert Dewar]

Yes, I know you can program like this in Scheme, but what's the point?
To me if you are going to use a lisp-like language, it is horrible to
use this kind of procedural style. I would prefer to teach Haskel, as an
honest functional language, and Ada as an honest procedural language. I
find Scheme OK reasonable only if you confine your view to its functional
subset. To understand that your code really is in an insort place requires
a view of the semantics of the Scheme implementation that seems quite
inappropriate to me.

oops I mean inplace sort, not insort place, but I rather like the latter
thought, so I will keep it :-)

Re: No top schools use Ada

[Robert Dewar]

In an earlier note, I commented in response to the observation that CS and
SE are different, that I hoped that these splendid computer science schools
are sure to let their students and their students' prosepective employers
become aware of this important difference. We would not want students to
go to such schools thinking they would be competent programmers at the 
end, or for employers to make the mistake of thinking that their graduates
might be able to build software systems :-)

Re: No top schools use Ada

[Jay M Martin]

> Kevin V. Sobilo
> Sorry, but can you substantiate the fact that these are "TOP" schools???
> MIT, obviously has a reputation for RESEARCH, but I don't know of
> any strength in Software Engineering which is the center of this
> issue.

 1 Scheme    Massachusetts Institute of Technology
 1 C         Stanford University 
 1 Scheme    University of California, Berkeley
 1 C++       Carnegie Mellon University  (Corrected)
 5 Pascal    Cornell
 6 Scheme    University of Illinois,Urb
 7 C         University of Washington
 8 Pascal    University of Texas, Austin
 8 Pascal    University of Wisconsin, Madison (2/3-term Pascal, 1/3-term C)

Top 9 CS departments ranked in last months US News and World Report.

>Colin James:
>I don't see academics from Harvard, Yale, Stanford, Princeton, MIT, 
>CalTech, Rice, Monash University (AU), Dartmouth College, or any school
>I omit from ignorance which is ACTIVELY engaged here in any fruitful 
>discussions whatsoever.
>Could it be that this news group is either trivial or that those academics 
>are too busy doing things that matter to even care ?  ...
>This newgroup is not worthy of any outside respect, and, by in large, 
>it shows it.  By contrast, read other comp.lang newsgroups where the
>discussion actually proves something. 

What comp.* newsgroups has all the MIT/Stanford/CMU/Berkeley
professors contributing, I want to go read it!!!  Sure comp.lang.C++
has Stroustrup, but he is a "wild and crazy guy" because he is only
one of the few eminent Computer Scientists who lowers himself down to
actually press the electronic flesh with the masses (and we benefit
and are grateful for it!).   I just don't see many prominent
academics wasting their time with internet news.  Heh, posting
news and 50 cents will get you a cup of coffee! 

> too busy doing things that matter to care?
Well, more like doing theoretical masturbation, but thats close enough.
They got lots of other stuff to do (in order Letterman style ) besides
posting in comp.lang.ada:

1000 : Thinking of how Computer Science could actually benefit 
       industry/mankind.
  ...
   6 : Doing Research
   5 : Teaching undergrad students
   4 : Hand-holding Grad Students
   3 : Getting that paper out before the deadline.
   2 : Milking those Ol funding sources and managing that money.

 And the #1 time consumer for professors is!:

   1 : Machiavellian departmental politics!


Personally, I think this newsgroup (comp.lang.ada) is pretty good.   

Jay

Re: No top schools use Ada

[Norman H. Cohen]

In article <NEWTNews.19974.798127420.cjames@cec-services-01>,
cjames@envisionet.net writes: 

|> This newgroup is not worthy of any outside respect, and, by in large,
|> it shows it.

Then clearly it is not worthy of your valuable time.

We understand why we will no longer be able to enjoy your carefully
reasoned and dispassionate analyses, but we will miss you.

Good luck in your future endeavors.

--
Norman H. Cohen    ncohen@watson.ibm.com

Re: No top schools use Ada

[cjames]

In article <3n5qss$rgl@watnews1.watson.ibm.com>, <ncohen@watson.ibm.com> writes:
> Path: 
nsgate.envisionet.net!news.sprintlink.net!howland.reston.ans.net!gatech!udel!news.
mathworks.com!solaris.cc.vt.edu!swiss.ans.net!newsgate.watson.ibm.com!watnews.wats
on.ibm.com!ncohen
> From: ncohen@watson.ibm.com (Norman H. Cohen)
> Newsgroups: comp.lang.ada
> Subject: Re: No top schools use Ada
> Date: 20 Apr 1995 14:25:32 GMT
> Organization: IBM T.J. Watson Research Center
> Lines: 15
> Distribution: world
> Message-ID: <3n5qss$rgl@watnews1.watson.ibm.com>
> References: <D759Az.GHM@cs.fredonia.edu> 
<NEWTNews.19974.798127420.cjames@cec-services-01>
> Reply-To: ncohen@watson.ibm.com
> NNTP-Posting-Host: rios8.watson.ibm.com
> 
> In article <NEWTNews.19974.798127420.cjames@cec-services-01>,
> cjames@envisionet.net writes: 
> 
> |> This newgroup is not worthy of any outside respect, and, by in large,
> |> it shows it.
> 
> Then clearly it is not worthy of your valuable time.
> 
> We understand why we will no longer be able to enjoy your carefully
> reasoned and dispassionate analyses, but we will miss you.
> 
> Good luck in your future endeavors.
> 
> --
> Norman H. Cohen    ncohen@watson.ibm.com
> 

Thanks Norman, and go to hell.

Re: No top schools use Ada

[Vladimir Vukicevic]

In article <NEWTNews.19974.798127420.cjames@cec-services-01> cjames@envisionet.net writes:

> That fact is a sad commentary on the QUALITY of what goes on here:  
> 
> about 30% of the posts are from Robert Dewar who is a notorious ...
> about 10% of the posts are from David Weller who has a very ...
> about 5% of the posts are from Michael Feldman who is a ...
> about 1% of the posts are from Tucker Taft who sometimes ...
>
> and the rest of the posts are either honest questions from honest people
> or ridiculous flames against Greg Aharonian and others.

(Hmm. I need to start posting more.)

So, according to this, you and Greg never post?  Neither of you has
honest questions, and I don't think either of you have ever flamed
each other...

Well, let's see. Robert got the highest marks, 30% of all the posts.
However, since Robert has contributed to many more discussions and
topics than either you (or Greg) have... many of those posts are just
telling people to direct gnat questions to gnat-report. Many, many
more are answering questions about the language and discussing
language issues (Something that I've -never- seen either you or Greg
do... do you two actually know Ada? Or do you feel your time is better
spent making empty, repetetive comments and long-winded posts that all
look like reincarnations of last week's long-winded post?). Only a
small fraction are spent actually discussing Ada being abandoned by
the DoD, NASA, Air Force, Navy, every tool vendor, and heck, let's
say it's being abandoned by the White House kitchen staff as well.
Likewise with David Weller, Michael Feldman, and Tucker Taft... all
have contributed to many discussions that I doubt you have any idea
of what is being said.

So, what % amount do you and Greg contribute to this newsgroup?
Certainly at least 75% of the drivel. Definitively 0% of anything
useful whatsoever (other than to get a good laugh).

> This newgroup is not worthy of any outside respect, and, by in
> large, it shows it.  By contrast, read other comp.lang newsgroups
> where the discussion actually proves something.  Is this a necessary
> evil of Internet.  I think not, because other comp.lang groups are
> valuable.

Wait, does this mean that you'll be leaving? And Greg too? Did Christmas
come early this year or something?

> Raise the quality of discussion here through self-discipline and by not 
> being selfish, then comp.lang.ada won't appear as a FidoNet talk feed.

Colin, a few sentances ago you called Robert Dewar a "notorious academic
snob and patronizing Internet bully of newbies," stated that David Weller
has a "very limited intelligence" and "writes meaningless invective," and
that Tucker Taft is "overpaid for narrow views and an inability to write
clearly."

And -you-'re telling -others- to raise the quality of discussion?

	- Vladimir

Re: No top schools use Ada

[Chris Morgan]

Hello everyone,

  Ditto everything Vladimir just said!

  No need to repeat the obvious, but I agree with Vladimir that these
things need rebutting.

Chris

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

-- Chris Morgan
-- SMCS Project, BAeSema Ltd, Scientific House, 40-44 Coombe Rd,
-- New Malden Surrey, England, UK
-- e-mail : chris.morgan@baesema.co.uk
-- Currently grappling with Ada95 via GNAT2.03 on Sun-SPARC-SOLARIS2.3
-- Member, Team Ada       when does team GNAT start?
------------------------------------------------------------------------

Re: No top schools use Ada

[Kenneth Almquist]

jmartin@kaiwan009.kaiwan.com (Jay M Martin) writes:
> Lets look at which languages the top CS department use for their first
> class:

According to the full list (summarized by Richard Pattis in comp.edu),
the most widely used language for CS1 is Pascal (36%).  Ada is in second
place with 17.5%, C is fifth with 8.6%, and C++ is sixth with 6.5%.  The
fact that none of the nine schools listed by Jay Martin use Ada for their
first class is not statisticly significant.

The use of C, however, is statisticly significant (P <= 0.05, barely).
If one wants to try to read anything into this preference for C, it is
important to keep in mind that the CS departments listed by Jay Martin are
the ones with top ranked *graduate* programs.  Graduate program rankings
mainly reflect the quality of the research done at the department.
				Kenneth Almquist

Re: No top schools use Ada

[cjames]

In article <D79HGD.Bo0@nntpa.cb.att.com>, <ka@socrates.hr.att.com> writes:
> Newsgroups: comp.lang.ada
> Path: 
nsgate.envisionet.net!news.sprintlink.net!howland.reston.ans.net!news.moneng.mei.c
om!uwm.edu!fnnews.fnal.gov!gw1.att.com!gw2.att.com!nntpa!not-for-mail
> From: ka@socrates.hr.att.com (Kenneth Almquist)
> Subject: Re: No top schools use Ada
> Message-ID: <D79HGD.Bo0@nntpa.cb.att.com>
> Sender: news@nntpa.cb.att.com (Netnews Administration)
> Nntp-Posting-Host: socrates.hr.att.com
> Organization: AT&T
> References: <3mq0jd$r10@kaiwan009.kaiwan.com>
> Date: Wed, 19 Apr 1995 02:55:25 GMT
> Lines: 16
> 
> jmartin@kaiwan009.kaiwan.com (Jay M Martin) writes:
> > Lets look at which languages the top CS department use for their first
> > class:
> 
> According to the full list (summarized by Richard Pattis in comp.edu),
> the most widely used language for CS1 is Pascal (36%).  Ada is in second
> place with 17.5%, C is fifth with 8.6%, and C++ is sixth with 6.5%.  The
> fact that none of the nine schools listed by Jay Martin use Ada for their
> first class is not statisticly significant.
> 
> The use of C, however, is statisticly significant (P <= 0.05, barely).
> If one wants to try to read anything into this preference for C, it is
> important to keep in mind that the CS departments listed by Jay Martin are
> the ones with top ranked *graduate* programs.  Graduate program rankings
> mainly reflect the quality of the research done at the department.
> 				Kenneth Almquist
> 

Any discussion of statistical significance here is meaningless without stating the 
number of subjects of interest, ie, the Sample Size.

Re: No top schools use Ada

[Kenneth Almquist]

I wrote:
>> The use of C, however, is statisticly significant (P <= 0.05, barely).

It appears that CMU uses C++ rather than C.  Statistical significance goes
out the window.

> Any discussion of statistical significance here is meaningless without
> stating the number of subjects of interest, ie, the Sample Size.

That's why I was careful to write:
>> The fact that none of the nine schools listed by Jay Martin use Ada...
                             ^^^^
					Kenneth Almquist

Re: No top schools use Ada

[Fernando Mato Mira]

In article <3n10sv$i97@gopher.cs.uofs.edu>, beidler@guinness.cs.uofs.edu (Jack Beidler) writes:

> And in how many of those "top schools" is the course taught by
> grad. assistants? 
Hm. There are some _undergrad_ students that know more than some professors..

> So, what do you expect, hacking in C, not
> how to program!  

Using C for this in a `programmer factory' would be normal,
but at a _university_, it is really perverted..

-- 
F.D. Mato Mira           http://ligwww.epfl.ch/matomira.html                  
Computer Graphics Lab    matomira@epfl.ch 
EPFL                     FAX: +41 (21) 693-5328