RDFa example
Bruce D'Arcus
can see an example here:
<http://www.citeulike.org/user/egonw/article/7812280>
So in their case, they're just adding the triples to the document head, a la:
<meta property="dc:date" content="2010-09-17T12:00:26-00:00" />
<meta property="prism:publicationYear" content="2010" />
Bruce
Bruce D'Arcus
> Bruce,
>
> 2010/10/4 Bruce D'Arcus <bda...@gmail.com>:
>> So CiteULike has recently added RDFa support (though not BIBO). You
>> can see an example here:
>> <http://www.citeulike.org/user/egonw/article/7812280>
>> So in their case, they're just adding the triples to the document head, a la:
>
> don't currently support PRISM, so we're better off using the
> purpose-built CiteULike translator for now anyway.
>
> My understanding was that RDFa was implemented by included links to
> RDF description of the page, or of specific content on a page.
>
> It would be good to add support for PRISM
> (http://www.idealliance.org/industry_resources/intelligent_content_informed_workflow/prism/specifications)
> and other metadata standards to the Embedded RDF translator, but that
> hasn't yet happened.
The point is, this is W3C standard RDFa; it just happens to be
embedded in the header, instead of the body content.
So on the translation end, you could imagine a kind of two-stage
process: generic RDFa + the specific vocabulary mapping. Is there
really much point long-term in having a specific translator for CUL?
But I was also just pointing out that CUL has RDFa support now, and
zotero.org doesn't ;-)
Bruce
Avram Lyon
2010/10/4 Bruce D'Arcus <bda...@gmail.com>:
> So CiteULike has recently added RDFa support (though not BIBO). You
> can see an example here:
> <http://www.citeulike.org/user/egonw/article/7812280>
> So in their case, they're just adding the triples to the document head, a la:
This is already supported by Zotero -- we call it Embedded RDF. We
don't currently support PRISM, so we're better off using the
purpose-built CiteULike translator for now anyway.
My understanding was that RDFa was implemented by included links to
RDF description of the page, or of specific content on a page.
It would be good to add support for PRISM
(http://www.idealliance.org/industry_resources/intelligent_content_informed_workflow/prism/specifications)
and other metadata standards to the Embedded RDF translator, but that
hasn't yet happened.
Avram
Avram Lyon
> embedded in the header, instead of the body content.
>
> So on the translation end, you could imagine a kind of two-stage
> process: generic RDFa + the specific vocabulary mapping. Is there
> really much point long-term in having a specific translator for CUL?
I would like for the Embedded RDF translator to handle more expressive
vocabularies. PRISM and BIBO are good examples. So, perhaps, is Zotero
RDF. I don't see why these couldn't be folded into "Embedded RDF.js".
As for the long-term fate of the CiteULike translator, we hope that it
and all site translators will become obsolete, as all sites start to
provide detectable and useful metadata, as well as pointers to
alternative representations for download (PDFs, etc.). That,
unfortunately, is unlikely to ever happen, so some translators will
always be necessary.
> But I was also just pointing out that CUL has RDFa support now, and
> zotero.org doesn't ;-)
Ah, yes. Some day. Maybe when Endnote starts using citeproc-js?
- Avram
Dan Stillman
Slightly tangential, but while we're on the subject, it seems to me that
the most interesting part of RDFa is the ability to overlay
human-readable page content with computer-readable metadata. I don't
know if there are any existing tools that actually take advantage of
that, but you can imagine such tools doing interesting things. So while
it's more difficult technically, my inclination is to embed our RDFa in
the XHTML content rather than just in the document head.
Dan Stillman
> On 10/4/10 11:29 AM, Bruce D'Arcus wrote:
>> So CiteULike has recently added RDFa support (though not BIBO). You
>> can see an example here:
>>
>> <http://www.citeulike.org/user/egonw/article/7812280>
>>
>> So in their case, they're just adding the triples to the document
>> head, a la:
>>
>> <meta property="dc:date" content="2010-09-17T12:00:26-00:00" />
>> <meta property="prism:publicationYear" content="2010" />
>
> Slightly tangential, but while we're on the subject, it seems to me
> that the most interesting part of RDFa is the ability to overlay
> human-readable page content with computer-readable metadata. I don't
> know if there are any existing tools that actually take advantage of
> that, but you can imagine such tools doing interesting things.
(Tools like, say, Zotero.)
Bruce D'Arcus
Yes, good point. There are some generic RDFa plug-ins for FF that
will, for example, highlight all RDFa content on a page. Here's an
example, though I haven't tried it:
<https://addons.mozilla.org/en-US/firefox/addon/3886/>
I also blogged about some work that Jeni Tennison was doing awhile back:
<http://community.muohio.edu/blogs/darcusb/archives/2009/03/26/rdfa-for-scholarship>
Bruce