🚀 go-pugleaf

In <1605@newsserver.cs.uwindsor.ca> arindam@cs.uwindsor.ca (Arindam Das) writes:

>Hi !

>I am implementing an automated mask layout generator for dynamic CMOS
>circuits.

[Text Deleted]

>Now, In my program i need to do complex traversal of the graph.
>I being at ease in using Prolog (Quintus Prolog), am using it
>comfortably and i am finding that i can develop the code pretty
>quickly.

I think you had just give the reason why you should use Prolog 
instead of C. It is easier to develop the code.

>However I have to defend the use of Quintus Prolog instead of C.

Aside from the fact it is easier to write in Prolog, the resulting
code is easier to maintained even by other persons.

I could even add that I would rather have a slow and correct program
than a fast code that I am not really sure of. But are Prolog programs
uniformly slow across the board. An article in the Janaury 1992 issue of
the Computer magazine (of IEEE Comp Society) has something to say along
this line.

Some might say that Prolog is a nice toy language suitable for small
eligant toy problems. The book "Practice of Prolog" might offer some
counter arguments. [Edited Leon Sterling, MIT Press 1990].

=======

Some closing comments :

I believe that Prolog, like other programming languages, has its
good and bad points. It just sad that misconception and prejudice 
ambivalence towards Prolog is still wide spread these days.

Why should a Prolog programmer has to justify using Prolog instead of
language X? 

In article <gerry.718237883@coral> gerry@coral.cs.jcu.edu.au (Gerry Trinidad) writes:
>In <1605@newsserver.cs.uwindsor.ca> arindam@cs.uwindsor.ca (Arindam Das) writes:
>>Now, In my program i need to do complex traversal of the graph.
>>I being at ease in using Prolog (Quintus Prolog), am using it
>>comfortably and i am finding that i can develop the code pretty
>>quickly.
>
>I think you had just give the reason why you should use Prolog 
>instead of C. It is easier to develop the code.
>
>>However I have to defend the use of Quintus Prolog instead of C.
>
>Aside from the fact it is easier to write in Prolog, the resulting
>code is easier to maintained even by other persons.

This is a special case of the general problem, how to justify use of
programming language X instead of programming language Y.

Things you might appeal to:
  - Ease of expressing the algorithm
       (which reduces program development costs,
        as well as future debugging and modification costs)
  - Efficiency of execution
       (That's right, if you're using backtracking and unification then
        your program will probably run as fast under Prolog as under any
        other language.  We once ported a simple Prolog list processing
        routine into Pascal as a test, and got a factor-of-14 *slowdown*!)
  - Portability
       (Prolog is available for a wide range of architectures and is about
        to have an ISO standard.  Although not as portable as C, Prolog
        is, I would argue, comparable to COBOL and much superior to Pascal
        in portability.)

Argument you might have to refute (my tongue goes in my cheek here):
  "All programs should be written in C for efficiency/portability"
(Translation: "C is the only language I know."  In previous years this
same argument was heard with Fortran, COBOL, or assembly language in
place of C.)
-- 
:- Michael A. Covington     internet mcovingt@uga.cc.uga.edu 
:- Artificial Intelligence Programs       phone 706 542-0359
:- The University of Georgia                fax 706 542-0349
:- Athens, Georgia 30602 U.S.A.          amateur radio N4TMI

In article <1605@newsserver.cs.uwindsor.ca>, arindam@cs.uwindsor.ca (Arindam Das) writes:
> Hi !
> 
> I am implementing an automated mask layout generator for dynamic CMOS
> circuits.
> 
>    ...
> 
> However I have to defend the use of Quintus Prolog instead of C.
> 
> Could somebody in the net give me some advice and tips and points
> for defending the use of declarative prolog instead of imperative C ?


You might look in the proceedings of the conference on practical applications
of Prolog (April 1992).  There are lots of concrete examples there; one of the
more general papers is Peter Reintjes's "Elegant Technologies".  I suggest you
try to get a copy of the proceedings; if you can't get it, I have
a copy of Peter's paper (send me a note, specifying whether you want
it in LaTeX, DVI, or PostScript).  

For the conference proceedings, try:
	Al Roth
	PO Box 137
	Blackpool
	Lancashire FY2 0XY, UK
	Tel:	+44-253-58081
	Fax:	+44-253-53811
	Email:	alroth@cix.compulink.co.uk

-- 
	Peter Ludemann  +1-415-813-3800  ludemann@quintus.com

In article <1605@newsserver.cs.uwindsor.ca>, arindam@cs.uwindsor.ca (Arindam Das) writes:
> Now, In my program i need to do complex traversal of the graph.
> I being at ease in using Prolog (Quintus Prolog), am using it
> comfortably and i am finding that i can develop the code pretty
> quickly.

> However I have to defend the use of Quintus Prolog instead of C.

You just gave the best possible defence.  It wouldn't matter how wonderful
Prolog is if _you_ couldn't use it effectively for _your_ application.
Tell whoever is hassling you that you are "developing an executable
specification" and that you are willing to implement the final version
in C or whatever _IF_ there is a need for a really fast implementation
and _ONCE_ you have measurements on a working executable specification
which tell you what C data structures would be appropriate.  (I have
just pointed out to my 2nd year data-structures-in-Pascal students that
using the building blocks I have given them they can put together over
256 distinct implementations of graphs, and that's only counting versions
based on adjacency matrices/lists.  You need to know what operations your
_whole_ program performs and how often it performs each before you can
choose the best building blocks.)  Point out that it will be easier to
recode to the "right" C implementation from a high level specification
than from a low level one, so that if and when you _do_ need a C version
you are likely to get a better one than if you had started hacking right
away.
-- 
RMIT policy requires the use of Gender Inclusive Language in all publications
and in all information conveyed to students, staff, and potential job
applicants.  Lapses into English are the sole responsibility of the author.
The Constitution of Australia does _not_ guarantee Freedom of Speech.