>>>>> On Thu, 9 Nov 2000 22:28:02 +0000, Dave Cross <[EMAIL PROTECTED]> said:
> I've mailed this list a couple of times before about my Sub::Approx
> module, but have had no response. We've had a bit of discussion on
> hte subapprox mailing list and have decided that we will rename the
> module to Approx::Sub. This is because the we plan to work on
> Approx::Array, Approx::Hash, etc (in fact one of the list memebers
> already has a prototype of Approx::Scalar).
Just three days since the Sub:: namespace has been officially
introduced:-/
Dave, in your first note
http://www.xray.mpe.mpg.de/mailing-lists/modules/2000-07/msg00222.html
you say, this "is a completely stupid think" and now you decide to
write separate modules for Subs, Scalars, Arrays and Hashes? I regret
I didn't hear your talk, apparently I'm missing something.
My thinking is (I wouldn't mind being correted here): if you have good
approximate matching code, talk about its merits, compare it to
others, like Match::Approx or IR techniques and make it a module or
three, no? You instead say nothing about the matching code and create
namespaces for each and every entity that can be matched
approximately. This seems backwards to me. Feel free to just send me
your talk if it answers my qualms.
> I have therefore just uploaded the first version of Approx::Sub to
> my CPAN directory.
> I've also abstracted the glob-walking ocde that is fundamental to
> Approx::Sub into another module called GlobWalker - this is also in
> my CPAN directory.
Have you ever looked at Devel::Symdump? How does your GlobWalk differ?
--
andreas
