discussion of the Creative Commons Metadata work

["Nathan R. Yergler" <nathan AT yergler.net>]

Is it less of a kludge/fudge if we limit the futzing to examining the  
embedded RDF, and only change the deed if they've specified the  
optional content type?  That way there's no inference, we're just  
reflecting what they say.

Nathan

On Aug 31, 2005, at 10:32 PM, Lucas Gonze wrote:

> I do similar referrer magic with http://webjay.org/playthispage.   
> It's a
> kludge but works pretty well.
>
> However, I think that linking to a CC license is an assertion by the
> linker, and fudging around with the end result might weaken the  
> assertion.
>
> - Lucas
>
> Mike Linksvayer wrote:
>
>
> > Pretend that http://example.com has a link to
> > http://creativecommons.org/licenses/by/2.5/
> >
> > When someone clicks on that link, typically the user agent sends
> > http://example.com as the referrer.  The CC server could get the  
> > content
> > at the referrer and alter deed presentation based on that content
> > (including metadata found there), e.g., if it looks like primarily  
> > audio
> > content, the deed could be tweaked accordingly -- I had to  
> > hardcode one
> > such tweak, see
> > http://creativecommons.org/licenses/by-nc-nd/2.0/deed-music -- among
> > many other possibilities, I want to avoid ever creating a custom deed
> > url like that again.
> >
> > Is there something horribly wrong with the idea of getting the  
> > referrer
> > url?  With potentially changing the deed based on content found  
> > there?
>
>
>
>
> > The obvious downside is that a slow/dead host pointed at by the  
> > referrer
> > will slow down deed rendering, but I don't know if that is a serious
> > issue--if slow, it's "just" a continuation of an already bad user
> > experience, as the user's browser already presumably slowly retrieved
> > the content that links to a CC deed, while the dead host case is
> > unlikely, as the user's browser never got the page linking to the  
> > deed
> > in the first place.
> >
> > I'm sure someone will say this is a vector for denial of service
> > attacks, but then wouldn't directly attacking the linking page be  
> > more
> > effective?  The CC server will of course have to do a little  
> > caching and
> > throttling.
> >
> > Thoughts?
>
> _______________________________________________
> cc-metadata mailing list
> metadata AT creativecommons.org
> http://lists.ibiblio.org/mailman/listinfo/cc-metadata