Re: CVS commit: pkgsrc/security/py-py-bcrypt

[rodent%NetBSD.org@localhost]

Just because we've been doing something a certain way for some time doesn't
make it right nor the best way to proceed.

Our idea that we know how to name a package better than upstream is wrong.

When users search for a particular package in our system, they will use what
upstream names it, not what we think it's supposed to be called. In the latter
case, the search might (if at all) return a result that is not the same
package as desired, the user would be confused upon using it or have to dig
through a list of other packages to determine which one is "right", which will
differ from whatever we said (POLA) was "right". There are several examples of
this in our tree, from rubyzip to python MySQL wrappers, etc.

Asking users to magically know our "better" naming system is wrong as well.

The simplest way to proceed is to slam php/python/ruby prefix on the beginning
of the package name - even if that produces funny package names - because it
will cause less confusion overall to users, minimise the chances of
conflicting package names in the future (again, see rubyzip or others), and
require less thinking on the parts of everyone involved.

We could continue discussing minutia of little value too. If anyone is
offended egregiously by this package name, they are welcome to move it.