home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group02b.txt
/
000153_icon-group-sender_Mon Dec 9 08:31:09 2002.msg
< prev
next >
Wrap
Internet Message Format
|
2003-01-02
|
3KB
Return-Path: <icon-group-sender>
Received: (from root@localhost)
by baskerville.CS.Arizona.EDU (8.11.1/8.11.1) id gB9FV7k24216
for icon-group-addresses; Mon, 9 Dec 2002 08:31:07 -0700 (MST)
Message-Id: <200212091531.gB9FV7k24216@baskerville.CS.Arizona.EDU>
From: ernobe <ernobe@msn.com>
X-Newsgroups: comp.lang.icon
Subject: Re: Icon compiler
Date: Mon, 9 Dec 2002 15:23:50 +0200
X-Newsreader: MicroPlanet Gravity v2.50
To: icon-group@cs.arizona.edu
Errors-To: icon-group-errors@cs.arizona.edu
Status: RO
In article <3df45481$1_7@news.meganetnews.com>, aa@bb.cc says...
> Thanks for your answer.
>
> > The Icon-to-C (iconc) compiler can produce very fast programs. I recall
> once
> > using it on the Icon versions of uuencode / uudecode ("iiencode.icn" and
> > "iidecode.icn" in the IPL) and ran some timing checks. The iconc versions
> of
> > iiencode / iidecode were over 3 times as fast as the interpreted (icont /
> > iconx) versions. More surprisingly, I found that the iconc programs were
> > actually faster than native C uuencode / uudecode programs that I was
> using
> > at the time!
>
> Well ... I have this Fibonacci test for benchmarks. With MS C ++ it takes 8
> secs to calculate the result for 40 as input and for Icon it takes about 600
> secs. Now with even 3 times speed improvement, it'd still be 200 secs for
> Icon. Given that, I think it's very exaggerated to say Icon programs can be
> faster than C. However slow compared to C, I do recognise Icon's power and
> I'm considering adding it to my software toolchest.
>
>
Icon is meant to enable people to code more efficiently than in C. Given that
it is people who produce computer programs, not computers programmed in C, to
refer to it as simply another tool in ones toolchest seems rather derogatory.
In fact it would probably not be far fetched to say that if everything that is
done in C nowadays would be done in Icon, there would not be a noticeable
difference, and programs would have much more functionality and attractiveness
to users.
I don't know much about computers myself, I can't really tell what an Icon to C
translator is, but since there is a compiler which produces executables and is
itself coded in C, it would probably take just someone well grounded in C to
make a translator. Of course the reason it hasn't is obvious to those familiar
with the problems of implementing a low-level language across several platforms
and OS. Until such time as there is more agreement about such platforms and OS
to make generalized low-level language programming efficient, Icon programmers
are well advised to familiarize themselves with C or Assembly, and optimize
their programs themselves.
I think functions programmed in C can be called, I don't know if it is possible
to add machine code, but it sounds like something a-lot simpler to accomplish
than in most other computer languages. It is a great way to learn Assembly, I
think, if one could learn how to optimize generators.
--
my esoteric links:
http://www.costarricense.cr/pagina/ernobe/