Re: how dirty can the HTML be, and still be RDFa?

Hi Peter,

The RDFa snippet in the spec is informative, and you are free to tweak it
as long as the resulting RDF graph matches the WebID profile [1] and the
SPARQL ASK query [2]. We're using the current RDFa REC spec which is RDFa
1.0, but we'll soon switch to RDFa 1.1 once soon as it reaches a more
stable status (it should come to Last Call in a couple of months). I'm in
the RDFa WG so I'll make sure to report back. I posted a RDFa 1.1 snippet
earlier this week on the mailing list [3]. RDFa 1.1 has several
improvements leading to simpler markup, plays well with HTML5 and isn't
as restrictive as RDFa 1.0 re doctypes.

Steph.

[1]
http://www.w3.org/2005/Incubator/webid/spec/#publishing-the-webid-profile-document
[2]
http://www.w3.org/2005/Incubator/webid/spec/#verifying-the-webid-is-identified-by-that-public-key
[3] http://lists.w3.org/Archives/Public/public-xg-webid/2011Nov/0154.html

On Thu, Nov 24, 2011 at 8:06 PM, Peter Williams <home_pw@msn.com> wrote:

>  blogspot is free (like wordpress), and consumer grade. ost importanbtly
> to me, its part of the google family, and thusi works with a google IDP
> login (that is now mapped onto US realty logins, via Azure's openid/ws-fedp
> gateways).
>
> With one edit to a simple template, blogger did allow me to change the
> html tag's header (to comply with RDFa) and add some namespaces. And, it
> did not strip out the marked up material in the blog post that followed,
> which came from the current spec.
>
> But, the result is nasty, when tested using hte W3C validator. Its not
> that nasty however, as the webid test suite's tool chain shows:
>
>
> http://webid.fcns.eu/lookup.php?uri=http%3A%2F%2Fyorkporc.blogspot.com%2F2011%2F11%2Fnothing.html%23me&submit=+Lookup+&html=0
>
> Not suprisingly, uriburner got something useful
> http://webid.fcns.eu/lookup.php?uri=http%3A%2F%2Fyorkporc.blogspot.com%2F2011%2F11%2Fnothing.html%23me&submit=+Lookup+&html=0
>
> Now, the point is, regardless of the fact that it doesnt validate per the
> schema, 2 tools do seem to be happy. One (uriburner) is probably doing lots
> of guessing and intuiting data, and the other ill guess is simpler - and
> simply parsing the (dirty) HTML, per the standard -
>
> Now, I could go to my Microsoft CA and mint 1000 .p12 files whose certs
> have the relevant blogspot post URI, use the users password to encrypt the
> file, post off a download URI to the user's registred email address, and
> also machine post 1000 user profiles in RDFa to such each of 1000 such
> entrie s on that one blog site (creating 1000 "foaf cards" formally, each
> on their own URI, and each with hashtag of #, and the cert). But, is that
> kind of dirty HTML intended be acceptable and consumable by the typical
> webid validation agent?
>
> Im hoping the answer is yes. I need it really simple (and what I did above
> satisfies that rule).
>
> it really matters (to me) that I can use commodity web stuff, with sites
> powered by multi-vendor websso, works alongside Google Apps, hotmail, etc
> At some point, the keys in the webid profile will have to cooperate with
> the more formal CA-managed certs that realtors maintain (so they can submit
> signed PDF documents to the US govt realty sites). But, that can wait.
>
>
>
>
>
>