home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group95b.txt
/
000057_icon-group-sender _Mon Jun 19 18:09:00 1995.msg
< prev
next >
Wrap
Internet Message Format
|
1995-09-18
|
3KB
Received: by cheltenham.cs.arizona.edu; Tue, 20 Jun 1995 13:23:49 MST
Message-Id: <9506191810.AA27336@ns1.computek.net>
Mime-Version: 1.0
Content-Length: 2659
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: Mon, 19 Jun 95 18:09 CDT
From: gep2@computek.net
Subject: Re: ICON and Data Conversion
To: icon-group@cs.arizona.edu
X-Mailer: SPRY Mail Version: 04.00.06.14
Errors-To: icon-group-errors@cs.arizona.edu
>>....One of the requirements is always to take mainframe information and
convert it into some format that can be loaded into the new system database,
usually UNIX based but I suspect I'll be working with NT soon. When I perform
this work, I require the delivery of a plain ascii text file. I then use KSH
with PERL or AWK to parse the file and create the file(s) for loading. Some of
the conversion processing involves creating relationships with several other
files or database tables for referential integrity or looking up descriptive
data, etc. I am always looking for good tools that can make my work easier and
of better quality. I would like to know if ICON is suitable for this type of
work.
>ICON is certainly suitable for this type of work, but even though I have ICON
available for the extensive amounts of data conversion that I must frequently
perform, I was surprised to discover that AWK seems to be the language that I
most frequently choose for most of these tasks.
I think that while you gravitate back to AWK (ugh) that's where I would
gravitate back to SNOBOL4+... but I think S4+ is far superior to AWK/Perl, has
the concise and elegant pattern description (almost as concise, and FAR more
powerful, than REs).
>It's nothing against ICON. I still use it a lot, but it just seems like AWK has
about the right combination of simplicity, power, and ease of programming for
doing data conversions. I understand that AWK was designed with such tasks in
mind.
AWK and Perl are basically "Poor Man's SNOBOLs", a logistical substitute
suitable for implementation on a "tiny machine" that was too small to implement
a more sophisticated language like SNOBOL4. It's certainly understandable that
they would be more inclined to implement a simpler language than S4... there
have only been a handful of truly complete, totally new implementations of S4
ever written. (Nearly all of them have been ports of the SIL macro-based
implementation).
>I find that such functions as split, (g)sub, and printf, plus the
regular expressions are indispensible most of the times I have to
convert data,
I consider REs a miserably poor substitute for SNOBOL4 patterns, a throwback to
UNIX's teletype origins where every keystroke required a punch on a paper tape
or a character strike on a 10cps teleprinter. Nowadays that we all work on
video displays and with orders of magnitude more processing power, sticking with
something as primitive as REs is simply pathetic, as out-of-date as just echoing
a backslash or something for each press of the backspace key.
Gordon Peterson
http://www.computek.net/public/gep2/