[cc-community] Is there a validator for CC tagging?

Paul Houle paul at ontology2.com
Tue Mar 9 12:59:19 EST 2010


Mike Linksvayer wrote:
> On Tue, Mar 9, 2010 at 9:10 AM, Paul Houle <paul at ontology2.com 
> <mailto:paul at ontology2.com>> wrote:
>
>        On the other hand,  if it can really be made as simple as dropping
>     in a <meta> element,  and if you can give people unambiguous
>     instructions that give them exactly one choice as to how to do it,  I
>     think you'd get a high take-up rate.
>
>
> For what purpose? <meta> elements are hidden to user and machines 
> don't need yet another place to look. If you're concerned about 
> validation (indeed even if you aren't), change
>
> <img src="http://ny-pictures.com/nyc/cc-icons/by.png" alt="by/2.0" 
> class="small_license">
>
> to
>
> <a href="http://creativecommons.org/licenses/by/2.0/" 
> rel="license"><img src="http://ny-pictures.com/nyc/cc-icons/by.png" 
> alt="by/2.0" class="small_license"></a>
>
> and add nothing else.
>
> That happens to work as RDFa, but validates and is also 
> http://microformats.org/profile/rel-license
>
    The fact that that "works" as RDFa is pretty scary to me,  because I 
don't know why.

    (i) How does it know what namespace "license" is in?  Where was 
"license" defined?  What else is defined in that source?  Is there one 
page I can go to and get a complete list of "rel"s that are valid?
    (ii) I'm assuming that it knows that the subject of the assertion is 
the document that contains the assertion because I haven't specified any 
subject; my understanding is that I could set a different subject by 
using a certain attribute on an enclosing element...  That would 
probably let me publish semantic metadata on a page like

http://ny-pictures.com/nyc/photo/topic/10421/Unisphere

    that makes assertions about individual pages visible,  but now all 
of a sudden a whole bunch of problems show up...
    (iii) I'm comfortable using
http://creativecommons.org/licenses/by/2.0/

    as a linked data link (although I'm not sure if it should have the 
"/" at the end.)  I can tell you,  with absolutely no uncertainty,  that 
that's a terrible link to send end-users to.  I have another photo 
collection site that causes me to get more than 10 emails a week from 
people who ask "can I use this picture?" even though I do send them to 
that link.  I don't know what it is,  but I know they look at it and 
their eyes glaze over and then they pound an email out to me.  (Often it 
takes multiple messages to try to explain "CC-BY-SA" so I usually say 
they can use "CC-BY" and they'll never understand "CC-BY-SA")

    I ~do~ know that there's A LOT more to the CC-Licenses that appears 
on the page (for instance,  the fine print about CC-BY-SA and 
collections) that I think would surprise many publishers and users,  but 
you're not going to find those from the license pages.

    Longer term I'm much more inclined to point people to my own pages 
per license,  which will have two things:  (a) a step-by-step checklist 
that answers the question of "can I use this and what do I have to do?"  
(I can't afford to keep answering these damn emails as my traffic 
increases) and (b) a navigational axis to see content that's under a 
certain license.

    Provide something that's meaningful to end users at that page and 
I'll change my mind. 

    That brings up a general issue I have with RDFa too,  which is that 
it ties together presentation and content.  In the case above,  for 
instance,  the decision of what to show users for the license (text,  
image,  flash,  whatever) and where to put it is a question for a web 
designer:  it's really a matter of usability,  how it looks,  etc.  
However,  if they make changes to the code that does that display,  it's 
also got an impact on my "linked data" presence.

    Take an example of "using enclosure relationships" to denote the 
subjects of triples.  Well,  I might have it all working just fine,  but 
then I decide that I want to make things look different,  so I move some 
content to a different <div>,  or rather,  my web designer does.  She 
forgets to put the right attribute on that other <div> so now I'm 
publishing the wrong triples.  It's a pretty easy mistake to make,  
because she's already dealing with a system that stacks templates inside 
of templates so I can reuse visible elements in different places.

    I'm willing to say that if you did a "usability test" of RDFa and 
picked 10 average webby people off the street,  you'd probably find that 
a large majority of people would misread the specs and do things 
incorrectly.  Compounding all this is the "open world assumption" which 
conflicts with the idea of validation at all:  I mean,  maybe you really 
want to assert "dc:licenes".






More information about the cc-community mailing list