- From: Sandro Hawke <sandro@w3.org>
- Date: Fri, 29 Nov 2013 12:54:53 -0500
- To: Markus Lanthaler <markus.lanthaler@gmx.net>, semantic-web@w3.org
On 11/29/2013 09:29 AM, Markus Lanthaler wrote:
> On Friday, November 29, 2013 12:18 PM, Andy Seaborne wrot:
>> On 29/11/13 09:04, Dan Brickley wrote:
>>> Whatever we do, let's do it in RDFa this time. Humans are at least as
>>> important consumers of schemas as computers.
>>>
>>> FWIW I think some of the adoption we saw back with the FOAF work
>>> (apart from being there early) came from having namespace URIs
>>> de-reference to (more or less) human readable documentation. Far too
>>> many schema URIs point at a ridiculously unreadable XML file that
>> just
>>> gets saved to disk and can't be opened with any useful tooling.
>> Use content negotiation. Make at least HTML, RDFa, Turtle, JSON-LD
>> available.
> +1000
Why RDFa? What would that be good for?
Let's assume the master is in a database. Then I can serialize that in
Turtle, JSON-LD, and RDF/XML easily enough. I'm working on the
conversion to HTML. I don't see much value in having the HTML include
the triples, embedded. What value there is seems like it would come
from having the embedding be tight (markup around all the bits of the
HTML, like search engines line), and that's a lot of work. It would
be relatively easy (but slow down the page loading) to embed it as a
separated block of RDFa, and also include <script> sections with the
Turtle and JSON-LD, ... but I don't see much value in that.
-- Sandro
Received on Friday, 29 November 2013 17:54:59 UTC