Phillip,
If the problem depends on a truly random number then many of the
implementations of loop/break here would fail. Most prngs operate as Icon's
(mind you most don't have the dual sequence characteristic either).
I think it import to keep in mind for these that the main purpose is show how
to code problems in different languages. Far different from production
considerations.
I sympathies, a friend worked on a project years ago to do random sampling for
a national survey. The chap who implemented the selection code used a very
broken random number and they had to re-sample and survey.
I'd love to see examples that are ugly in Icon and better in Unicon. HTTP is a
simple example of one.
David
________________________________
From: Phillip <[email protected]>
To: David Gamey <[email protected]>
Cc: [email protected]
Sent: Tue, April 6, 2010 11:51:08 AM
Subject: Re: [Unicon-group] Rosetta Code - Unicon / Icon consistency edit - fyi
David,
Thanks for making the changes in the loop/break program. In my opinion, if
this were a production program, the program does not run correctly under
Icon. Admittedly, more comments and samples of output would have made the
problem clearer.
We probably need some consensus about whether we need to $ifdef some
programs so that they perform identically under the two languages.
Some years ago when in Australia, I had a physicist friend who was about
ready to present a major paper. Just before the conference, he reran his
program and realized that the random function in his language needed a
random seed. When he fixed the problem (as one must under Icon), his
splendid results disappeared for most seed values.
I hope we can soon get some clarity on how to do the Icon/Unicon code
examples. This morning I tried to figure out how many of the Unicon ones
had been given over to the Icon section and gave up after about 3
attempts: just too much "see other_language" work. A bit peevishly, I
admit, I was thinking I'd like to put some things under Icon that read:
"Can't be done; try Unicon."
Phillip
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Unicon-group mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/unicon-group
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Unicon-group mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/unicon-group
