[Clipart] PR Up - Please edit
Jonadab the Unsightly One
jonadab at bright.net
Mon Mar 7 06:33:31 PST 2005
Jon Phillips <jon at rejon.org> writes:
> find . -name "*.svg" | uniq | wc -w
As someone else pointed out, this doesn't strip out the paths. There
is a count-unique-images.pl script in the tools that is designed to
count unique filenames. This is still not exactly right. What we
really need to do is count unique author/title/filename combinations.
Perhaps in a future release we can get this worked out.
> In terms of names from the packaging process, there are around 250
> artists it seems...so actually more.
I am sure a few of those are duplicates (e.g., due to whitespace
issues (you'll note newlines in a couple of the names, and leading and
trailing spaces), pseudonyms, the inclusion or not of a middle initial
or email address, ...), but yes, I think there are nearly 250
contributing authors. But I think it sounds better to say "over 200"
than "nearly 250". When we think it's cleanly over 250, then we can
say so. I think it will not be very many releases before we can say
with confidence that there are over 250 contributors.
> I just did line count of the generated statistics.txt file. I see
> that Jonadab or someone else linked that file to the upper left hand
> link of amount of clipart and then number of artists.
Yes, although it actually only shows artists, not number of images, so
really it should only be linked to the authors part.
The numbers by each author are, as Nicu suspected, effected by the
same double-counting issue that makes it tricky to count the number of
unique images. I believe I know how to correct that, but it involves
counting somewhat differently, looking at both filenames sans paths
and the author and title metadata. The good news is that the same
counting script could report both an image count and author submission
statistics. I think for this release the counts in statistics.txt are
the best collated information we have on author submissions.
split//,"ten.thgirb\@badanoj$/ --";$\=$ ;-> ();print$/
More information about the clipart