On Sun, Mar 28, 2010 at 11:48:00AM -0500, Randy Kobes wrote:
> Has some sort of disk quota system for CPAN author accounts ever been
> considered?
There are authors with 100 distributions. There are authors with just
one distribution.
There are authors with big distributions, and authors with only tiny
distributions.
I'd not be in favour of anything like that, which would impose burdens
on authors (prolific authors - the most prolific being pumpkings: perl
lives in their PAUSE directories - would have to contact admins to get
their quota increased) and on the volunteer admins (who would have to
decide whether to increase someone's quota or not).
OK, so I have a vested interest: my CPAN directory is, in terms of size,
number 37 out of 4900-something, because I have two *really* big
distributions. For both of those I delete older versions when I think
it appropriate. However, the load on rsync servers doesn't really come
from the size of files - no matter whether you use rsync or some other
protocol, they still have to serve those big files out at some point,
once to each person who mirrors from them. The real load is the
*number* of files, and hence the number of stats they have to do when
someone asks rsync for changes.
If you really want to reduce the load, how about getting rid of the
CHECKSUM files and all the extracted blah.readme files in authors'
directories? I'm kinda tempted to say the same about the .meta files as
well, although I imagine they're more useful to some downstream reusers
of the archive.
--
David Cantrell | Nth greatest programmer in the world
When one has bathed in Christ there is no need to bathe a second time
-- St. Jerome, on why washing is a vile pagan practice
in a letter to Heliodorus, 373 or 374 AD
