Re: Vocabularies for file data, content events, errors

2009/9/2 Toby Inkster <tai@g5n.co.uk>:
> On 2 Sep 2009, at 10:37, Niklas Lindström wrote:
>
>> * simple file data properties, describing:
>>  - checksum+algorithm (and/or direct properties for md5, sha1/-2 etc.),
>>  - filename/slug (unless dct:identifier is suitable enough?).
>
> foaf:sha1 exists, but that might not be much use if you if you want to...

There are some conceptual issues remaining with that property:

Q: do we consider it uniquely identifying? if so, then ... of what?
(ie. is it owl:InverseFunctional)

Even asideConsider two zero-byte files, created by different people,
on different machines, and
with different creation dates. The empty file will have the same sha1
hash, but the file itself (considered as a manifestation of that null
work) has other characteristics that differ. I'd like to have a more
careful model for this that brings in some elements of the FRBR
work...

Something else to look at (if only because it's inside millions of
JPEGs and PDFs :) is Adobe's XMP...

cheers,

Dan

Received on Wednesday, 2 September 2009 10:21:36 UTC