On Sun, 16 Jun 2002, Jonathan M. Hollin wrote:

> I have been thinking about a reorganization of the Apache/Perl modules for
> a while, and have come to the conclusion that it would probably be a good
> idea. Please tell me what you think about this proposal.
>
> Per Einar, I have cut most your email only for convenience...
>
> I agree with you.  It would be great (necessary?) to reorganise the mod_perl
> modules on CPAN ready for mod_perl/apache v2.  Users of v1 would also
> benefit from a more descriptive name-space.
>
> However, the major problem, as I see it is, is simply that people already
> KNOW modules by a specific name and changing them is probably going to lead
> to confusion and possibly even some bitterness.  If I'm setting up a new
> server I'm going to be mighty pissed to have to unlearn what I know about
> the mod_perl module namespace...
>
> If you think this can be overcome (symbolic links on CPAN anyone?), or if
> you see major support for your proposal, then you also have my vote.  I
> think your idea is sound and logical.  Nice one.

It is a really good suggestion ... The existing modules not
following a naming convention do pose a problem, though, and
possibly future ones who don't/can't follow such a convention for
various reasons (historical, ignorance, multi-purpose, ...).
And then one might get back to having to manually edit the module
list to add these ....

To try to automate somewhat the idea of creating Apache
subcategories, and to accommodate existing/future modules, what
about doing something like PAUSE does in how it creates top-level
categories? That is, have a form which presents a list of some
number of subcategories of Apache (eg, some variation of those
Per Einar had: authentication, logging, database, ...., ), and
the author chooses one (or more) subcategories she/he would like
her/his module to appear in. For each subcategory a number of
subsubcategories are then automatically created, based on the
module name, and the name of the distribution then appears in the
subsubcategory. In this way existing modules can be categorized
without having to change their names (some of which probably
never will, even with modperl-2). As well, it can make allowance
for relevant distributions (eg, Embperl, HTML-Mason, SOAP-Lite)
for which one might not even think to look under an Apache::*
namespace, or those which may fall into one or more
subcategories. Finally, manual editing of the list will be kept
to a minimum, as the author will be able to do most things ...

best regards,
randy kobes

Reply via email to