[Clipart] release tomorrow

[Jonadab the Unsightly One]

Matthew Gates <matthew at porpoisehead.net> writes:

> For example, 
> in ./openclipart-0.14/unsorted/lighthouse_matthew_gates_01.txt:
>
> Title:    Lighthouse
> Author:   Matthew Gates
> License:  Public Domain
> Keywords: HASH(0x87262e4)
>           HASH(0x8920e08)
>           HASH(0x8652a68)
>           HASH(0x85cb20c)
>           HASH(0x8721074)
>           HASH(0x8894264)
>           HASH(0x87d1774)
>
> I noticed some similar error in the submissions list on the site at:
> http://www.openclipart.org/clipart/statistics.txt

Hmmm, yes, and the new statistics.txt seems to be heavily afflicted
with this problem.  It appears, however, that the actual keywords in
the actual metadata in the .svg files are intact.  For instance, in
unsorted/lighthouse_matthew_gates_01.svg you see the following:
        <dc:subject>
          <rdf:Bag>
            <rdf:li>house</rdf:li>
            <rdf:li>boat</rdf:li>
            <rdf:li>navigation</rdf:li>
            <rdf:li>transport</rdf:li>
            <rdf:li>lighthouse</rdf:li>
            <rdf:li>building</rdf:li>
            <rdf:li>ocean</rdf:li>
          </rdf:Bag>
        </dc:subject>

So it appears that the metadata have not been dammaged by the
processing (including the authority control, which is the same script
that generates statistics.txt).  This is a very strange bug and is
probably related to the weird-filename bug (wherein, some images have
hash0xsomething in the filename instead of the author's name;
fortunately, these filenames are still fairly unlikely to cause
filename collisions, so it's not a huge disaster, but a bit odd,
somewhat puzzling, and a wee tad bit user-unfriendly).

It has been happening for a long time (especially in the keywords),
but it seems to be cropping up more *often* lately, and I'm starting
to think that I'm going to have to take a systematic approach to
hunting it down.

-- 
$;=sub{$/};@;=map{my($a,$b)=($_,$;);$;=sub{$a.$b->()}}
split//,"ten.thgirb\@badanoj$/ --";$\=$ ;-> ();print$/