[cc-metadata] non-web embedding revisited

Tim Olsen tolsen718 at gmail.com
Fri Dec 23 17:37:00 EST 2005


On 12/23/05, Mike Linksvayer <ml at creativecommons.org> wrote:
> A long time ago there was some discussion concerning license metadata in
> files disassociated from the web, eg
> http://lists.ibiblio.org/pipermail/cc-metadata/2003-June/000123.html
>
> In retrospect the conclusion to shove a simultaneously semi human
> readable and semi machine parseable statement in the "TCOP" field was
> dumb.  It isn't very readable, isn't exposed by programs if it were,
> isn't translatable, and is hideous from the perspective of machine
> parsing.
>
> We're going to release recommendations for embedding similar web-backed
> notices in several other file formats over the next year, so now may be
> a good time to correct the original mistake, create a new recommendation
> for MP3/ID3, and replicate that recommendation elsewhere.  (There isn't
> that much properly marked content out there using the existing
> recommendation, and only a few programs that look for compliant embedded
> metadata, all of which I suspect would be glad to comply with a saner
> standard.)
>
> The new recommendation would probably be to use two fields specifically
> intended to contain URLs and only URLs.  For ID3v2, those might be
>
> WCOP (pointing to a license URL)
>   Definition: http://www.id3.org/id3v2.3.0.html#WCOP
>   Example value: http://creativecommons.org/licenses/by/2.0/
>
> WOAF (pointing to a web page describing a file, potentially including
> license information/metadata)
>   Definition: http://www.id3.org/id3v2.3.0.html#WOAF
>   Example value: http://example.com/a_song_sung_by_me.html
>

I'd say go with WCOP.  This way a publisher has the option of having a
different link for WOAF.

Where will the verification url go?

> Note that these map to current use of licenses and metadata on web pages
> -- the license notice points to a CC license URL, and describes the
> current page (the current page is its own "official page").
>
> Comments?  Suggestions for implementation in other formats?  I'm
> particularly interested in EXIF, MP4/QT and OGG at this point.

Ogg has a LICENSE field which they say can be used for (among other
things) a URL pointing to the license.  See
http://www.xiph.org/vorbis/doc/v-comment.html

MP4 has a Copyright Atom.  See pages 36 - 37 of
http://mediaxw.sourceforge.net/files/doc/MPEG%204%20System.pdf
It's exact format is not specified in that document.  I wonder how, if
at all, any MP4 files are using that atom.

It might be difficult to persuade every file format owner to add a tag
that is specifically for a license (and/or verification) url.  The
nice thing about the semi-readable/parseable license statement is that
we can use any field that relates to copyright or licenses.

Furthermore, I don't think the current license statement is so hideous
for a machine to parse.  Only the content after the words "verify at"
is constrained - at least that's my reading of
http://creativecommons.org/technology/mp3 and
http://creativecommons.org/technology/ogg.

What can be hideous though, is if some programs are becoming dependent
on what may be becoming a de-facto format:  <year> <copyright holder>.
Licensed to the public under <licenseurl> verify at <verification url>

A more specific specification as to what a "CC-compliant"
publisher/reader should or should not depend on may help here.  (a la
RFC-style)

-Tim


More information about the cc-metadata mailing list