- From: Ivan Herman <ivan@w3.org>
- Date: Wed, 16 Nov 2011 09:11:50 +0100
- To: Gavin Carothers <gavin@carothers.name>
- Cc: "Eric Prud'hommeaux" <eric@w3.org>, Manu Sporny <msporny@digitalbazaar.com>, W3C RDF WG <public-rdf-wg@w3.org>, Jeni Tennison <jeni@jenitennison.com>, Sandro Hawke <sandro@w3.org>
On Nov 15, 2011, at 22:41 , Gavin Carothers wrote: [snip] > Please note the change in email address gavin@topquadrant.com is no > longer a working group member, gavin@carothers.name is ;) > > On Tue, Nov 15, 2011 at 11:19 AM, Ivan Herman <ivan@w3.org> wrote: >> >> On 15 Nov 2011, at 19:36, Gavin Carothers <gavin@topquadrant.com> wrote: >> >> [snip] >> >>> >>>> >>>> However. I wonder whether it makes sense for the surrounding RDFa content to have some effect on the turtle portion. Namely: >>>> >>>> - base setting in HTML (which is also the base for the generated RDF from RDFa) would be a @base for the encoded turtle. AFAIK we discussed that at some point, but I have not found it in [1] >>> >>> There is a need for specific language. But I'm not sure that >>> supporting using the HTML base the best way to go. The UA would need >>> to support <base> @xml:base and the base URI DOM API in order to be >>> compliant with the HTML5 notion of base. And exactly how that >>> interacts with @base would also need to be defined. This would also do >>> some odd things to copy and paste safety. >>> >>>> - maybe more importantly: if RDFa sets a bunch of prefix declarations (and in RDFa there are even some defaults, eg, for rdf or foaf), I wonder whether those prefix declarations should not be valid as @prefix declarations in the embedded turtle. I think that would really be useful for HTML+RDFa authors. >>> >>> Allowing the use of RDFa prefix declarations which can come from >>> xmlns, prefix attributes, and vocab attributes would in my mind >>> needlessly complicate the consumption and authoring of Turtle <script> >>> fragments. Again it would greatly reduce copy and paste safety. >>> >> >> First of all, I would consider only xmlns and prefix or, possibly, prefix only. Not vocab, that does not define a prefix. > > Okay time to go deeper into this issue. The prefixes defined in RDFa > are not the same as the prefixes defined in Turtle. RDFa prefixe > resolution uses CURIES and in Turtle PNames. The syntax is not > identical. Clearest case of that is in the Facebook Open graph > vocabularies which use :'s in the local part of the name which is not > allowed in Turtle. I'm not totally opposed to aligning Turtle with > RDFa, but others were as it breaks alignment with SPARQL which isn't a > great idea. > >> >> I understand the issue of copy paste. On the other hand, if I author an RDFa file, where I define a load of prefixes, and then I have to repeat the whole thing again is also error prone and certainly a paini the neck. Ie, I am not convinced the balance is again the reusage of prefixes. The same holds for base, referring to the previous issue... > > One thing this does to Turtle is create the notion of an initial > prefix map that may exist before parsing starts. This is a reasonably > large change to Turtle. A <script> Turtle document may not be a valid > Turtle document without the HTML it's in, not thrilled by that idea. > This might be simpler to solve with tooling that just reads an HTML > document and outputs the correct @prefix lines. > Sigh. You do have good arguments, although at present it is a pain to repeat all the prefix declarations. Ie, there would be no really smooth ways of using both turtle and RDFa in the same page, although I do see very legitimate usage for this (that is why I like it!). It has happened several times to me that the RDF I wanted to generate from an HTML file included a bunch of statements that are really not for display (this does happen with more complex vocabularies) and the current trick is to use a <div style="display:none> for this, and then encode things in a series of <span> which then resembles RDF/XML:-). Adding turtle in HTML is a very elegant way of solving this issue... Ie: I am torn! Ivan >> >>>> - SVG already has a way to add RDF/XML as metadata, as well as the possibility to add RDFa statements[2]. More interestingly, it also has a script element[3]. I think the Turtle syntax should allow for the same style of turtle embedding for SVG, too. >>> >>> The script element (with the added CDATA directives) would work the >>> same way as it does in XHTML. >> >> Right. What I am saying is that the Turtle document should refer to SVG alongside HTML. > > Okay, I'll take a stab at some text to that effect tonight. > >> >> >>> >>>> >>>> B.t.w., I think it would be good to publish a Turtle draft soon with those features. This Turtle-in-HTML would be an important addition to the current approaches of embedding RDF data into HTML... >>>> >>>> Thoughts? >>> >>> I would like to publish a new draft before the of the month. Exactly >>> what is included in it should be decided by the RDF WG. (N-Triples? >>> Turtle in HTML? xsd datatypes?) >>> >> >> I think turtle in html is important and good, and I would be in favour keepig in. > > So am I :D Just not sure about the rest of it. > >> >> Ivan >> >> >>> --Gavin >>> >>>> >>>> Ivan >>>> >>>> >>>> [1] http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-turtle/index.html#in-html >>>> [2] http://www.w3.org/TR/SVG11/metadata.html >>>> [3] http://www.w3.org/TR/SVG11/script.html#ScriptElement >>>> >>>> ---- >>>> Ivan Herman, W3C Semantic Web Activity Lead >>>> Home: http://www.w3.org/People/Ivan/ >>>> mobile: +31-641044153 >>>> FOAF: http://www.ivan-herman.net/foaf.rdf >>>> >>>> >>>> >>>> >>>> >>>> >>>> >> ---- Ivan Herman, W3C Semantic Web Activity Lead Home: http://www.w3.org/People/Ivan/ mobile: +31-641044153 FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Wednesday, 16 November 2011 08:09:25 UTC