Questions from soc.college

Thread View: comp.lang.functional

3 messages

[T] Back to Threads [V] Tree View [G] Group Articles

3 total messages Started by Charles Lin Wed, 19 Aug 1992 05:30

#160

Author: Charles Lin
Date: Wed, 19 Aug 1992 05:30

73 lines
3482 bytes

   Hi, again, folks.  Ignore this if you wish to read something
deeply serious about lambda lifting (whatever that is).

   There's been some talk in the group soc.college about
selecting computer science as a major.  Someone had
been wondering what languages some places might teach,
as well as the usual questions of whether research
or teaching is more important to a professor.

   Some folks have been arguing a more practical point of
view.  That is, languages like Pascal or Lisp is pretty
much useless in the "real world", and universities should
use some sort of real world language and deal with real
world projects.   Specifically, one person felt that
universities that taught Scheme-like languages as first
languages were too "theoretical", and that wasn't going
to get jobs.

    Now, I know that a quick response would be to say
that a course shouldn't focus on a language, and any
other languages can be easily picked up when the
concepts of one are known.  One programs into a language
not in a language (Gries?).

   So, the question is this.  While people are looking
to functional languages for their own research purposes,
is there a general feeling that these languages ought to
be the foundational basis of all work done, or at least
most work (in the same way that C is) done?  I doubt there
is any kind of academic push to make functional languages
the medium of instruction on general, or the language
for general "real world" use.

   Is it a danger that most people have to be "converted"
to learn functional languages?  Is it harder to learn?
Are people satisfied that most users of functional languages
seem to be purely in academia?   And even then,  it seems
that only researchers in functional languages use any
of the newer languages.  For example, never once have
I heard our group say "You know LISP is really not
such a great language, let's go to ML or Miranda".  

   I think it's implicitly understood that there is
so much support for LISP and that most of the AI
community understands LISP, and that it could take
a considerable amount of time to get proficient in,
say, Haskell.

   Obviously, one problem is currently support for
the language, in the sense of debugging tools.  Another
is availability.  But languages like PL/1 have been
around for ages, and have never really caught on.

   So, how compelling should it be for people to
learn functional languages or program primarily
in it?   Strangely enough, C was probably never meant
to be the language for everyday use, but since so
much software (for UNIX and related utilities) were
written, it already had lots of people using it and
maintaining it.

   Thoughts on the matter may be e-mailed or posted
if somehow it is felt that the volume in this group
is too low.

   By the way, I tries asking similar questions 
about logic programming as I did a few weeks ago
here.  No one bit, and the topic went away.

--
Charles Lin
clin@eng.umd.edu

Re: Questions from soc.college

#162

Author: Reinhard Eppler
Date: Wed, 19 Aug 1992 16:03

34 lines
2235 bytes

In article <1992Aug19.053005.28328@src.umd.edu>, clin@eng.umd.edu
(Charles Lin) writes:

|>   So, the question is this.  While people are looking
|>to functional languages for their own research purposes,
|>is there a general feeling that these languages ought to
|>be the foundational basis of all work done, or at least
|>most work (in the same way that C is) done?  I doubt there
|>is any kind of academic push to make functional languages
|>the medium of instruction on general, or the language
|>for general "real world" use.

My opinion is:

 o most of the concepts needed (at least I need) are present in LISP
 o we need another syntax for LISP
 o list comprehension, pattern matching, soft typing (and ...?) 
   should be added to LISP
 o a pure functional language might be too difficult to use
   (if we had started with pure functional languages, someone would have
   invented global variables to keep the number of parameters small.)
   -> we need multiparadigm languages, e.g. assignments for some cases 
 o by now functional languages are too slow; we need better optimization
   techniques. This also holds for 'pure' OOP.
 o Haskell probably has too much concepts (this is the case with LISP, too)
 o I regard the work on Haskell and similar languages as very useful
   for research and some day someone has to put the good things into
   a "real world" language.

--
Reinhard Eppler
eppler@cognac.informatik.uni-kl.de

"Would you like to end up with all work done?"

Re: Questions from soc.college

#164

Author: Thorsten Altenkirch
Date: Wed, 19 Aug 1992 18:51

38 lines
2368 bytes

In article <1992Aug19.160326.27474@uklirb.informatik.uni-kl.de> eppler@bordeaux.informatik.uni-kl.de (Reinhard Eppler) writes:

o most of the concepts needed (at least I need) are present in LISP
o we need another syntax for LISP
o list comprehension, pattern matching, soft typing (and ...?)
should be added to LISP
o a pure functional language might be too difficult to use
(if we had started with pure functional languages, someone would have
invented global variables to keep the number of parameters small.)
-> we need multiparadigm languages, e.g. assignments for some cases
o by now functional languages are too slow; we need better optimization
techniques. This also holds for 'pure' OOP.
o Haskell probably has too much concepts (this is the case with LISP, too)

Why don't you use ML?

I do not want to start a Haskell vs. ML debate here but I hope that
everybody will agree that ML is much more mature. Sure it does not
contain some concepts like type classes which are worthwhile to
explore but particulary for teaching this may be an advantage.

A survey which was recently posted here shows that there is a large
number of places where ML is used as an introductory programming
language. From what I have seen here in Edinburgh it seems to be well
suited for this.

To come back to the original question: I believe that people should
learn C (having hacked in it for a while myself) but not as a first
language. It is bad for your style as is playing an instrument the
wrong way in the beginning. It is hard to get away from bad habbits
you have aquired in the beginning although not impossible.

T.
--
************************************************************
Thorsten Altenkirch alti@dcs.ed.ac.uk
LFCS, Edinburgh
Scotland

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads