On Mittwoch, Aug 13, 2003, at 00:41 Europe/Zurich, yoy wrote:
Brian,
Thanks for sending me your script and everything. I'll study your work and
hopefully learn from it, once I figure out how to use the Revolution command
line interface. No doubt it's in the docs. That was surprising to learn.
Debating perl vs. Revolution speed in OmniLotto's case was good for me. In
reality, anyone who uses OmniLotto wouldn't likely play 100,000 games. For
$10 worth of lotto tickets, the speed of OmniLotto is certainly an
"in-blink-of-an-eye" performer!
It's also nice to get back a mature xtalk since hypercard was canned.
BTW, today I purchased the Studio edition of Revolution. It's a very feature
rich environment that builds for many OSs and I really want to get gui apps
out there for those who want my offerings.
There's so much to learn, again! ;-)
All the best,
Andy Burns
----- Original Message ----- From: "Brian Yennie" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Tuesday, August 12, 2003 5:38 PM Subject: Re: Revolution speed sucks?
Did your revolution script read in and write out to text files as the PERLotto script does?
Yes. It uses your config file and writes out to "results.txt".
If so that may explain the 10% Perl speed advantage. Did you read in and write out to a field? Did you read the field into a variable and then write the results back to the field? Did you leave out the textcolor commands out, etc.?
I ran the script from the command line. You may not be aware of this (and many aren't), but Rev also has command-line engines (i.e. no GUI, just runs a script like Perl). I ran with the Mac 'Darwin' version: there is also a windows console engine available.
The behavior is as close as possible to the Perl script, since I wanted
a fair comparison of the engine alone: it reads from a file and writes
to one in the same manner as the Perl script. It could be sped up by
using variables and writing the file once, but then it wouldn't be a
1-to-1 comparison anymore.
I modified OmniLotto to read the field into a variable "games" and did
all
the calculations there and then spit the variable games into the field
"gamelist" on a line by line basis (with lock screen until done).
To process 100,000 games in Perl took 14 seconds while Revolution took
over10 minutes to process only 10,000 lines.
The more lines in the field, no matter whether the data is manipulated
in
the field directly or in an array increases the time to process by some
order of magnitude. At least that's my perception, in this type of data
manipulation. Perl does not.
Would you care to share (and I'm humbled that you took the time to translate my perl code in the first place) your .rev script with me for inspection? I would be most curious and "sharing the magic" is never a bad thing, right? ;-)
I'll send you a copy off-list. You're welcome to use it for anything you like, share it, modify it, etc.
I think this has been a good exercise for all of us!
------------------------------------- Brian Yennie Chief Technology Officer QLD Learning, LLC (904)-997-0212 [EMAIL PROTECTED] --------------------------------------
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
