Revolution speed sucks?

[Brian Yennie]

> While the interest in fairnes is appreciated, if you're using fields 
> instead
> of variables it tilts away from Rev unfairly, as Perl has no rich-text 
> field
> objects to contend with.

I'm not using fields at all: I'm writing directly to text files. I also 
ran both in the script engine (no GUI at all) and in the GUI 
application with no GUI interaction (script gets compiled and gains a 
speedup).  The scripts don't have any major differences like this that 
I can see: they read and write the same files, use the same loops, etc.

> But even if algorithmically identical, one of the factors in choosing 
> a tool
> is to exploit its unique strengths.  For example, for many tasks simple
> chunk expressions are far faster than the more generalized RegEx, but 
> chunk
> expressions are only found in xTalks (definitely an "unfair" advantage 
> of
> Rev over even some 3GLs <g>).

Completely agreed. There are absolutely tasks for which each tool 
shines. If this script was full of "last char of item 4 of line -3 of 
myVariable", the Perl would be virtually unreadable.

I wouldn't even attempt a Perl-based GUI app, and I doubt I'd write a 
complex pattern-matching server-side script in MetaTalk either. Of 
course, both tasks are possible.

In the end, none of this should be considered to be anything 
approaching an exhaustive comparison of the two engines: just a direct 
comparison of a particular task with a particular algorithm. I'd love 
to see more, and I invite others to investigate the scripts in question 
and/or do the same thing with others.


-------------------------------------
Brian Yennie
Chief Technology Officer
QLD Learning, LLC
(904)-997-0212
briany at qldlearning.com
--------------------------------------