Re: Style Issue

Hi Graham,

We looked into Fresnel, but it seems much more heavy-weight than we need.
 Also, we're not styling the RDF itself, but the resources that are part of
the RDF, so the semantics aren't quite correct either.

Rob


On Thu, Jan 3, 2013 at 5:35 AM, Graham Klyne <graham.klyne@zoo.ox.ac.uk>wrote:

> A long shot here, but does Fresnel help at all?
>
> http://www.w3.org/2005/04/**fresnel-info/manual/#**csshooking<http://www.w3.org/2005/04/fresnel-info/manual/#csshooking>
>
> #g
> --
>
>
> On 02/01/2013 20:16, Robert Sanderson wrote:
>
>> On Wed, Jan 2, 2013 at 12:56 PM, Tim Cole <t-cole3@illinois.edu> wrote:
>>
>>  **1.       **Probably no with regard to gss:style. Was tied to some of
>>>
>>> the contemporary CSS work, which has moved on, while GSS seems entirely
>>> dormant. And as you say, it never reached the level of W3C
>>> Recommendation.
>>> But it does give some reinforcement to the approach you've proposed.
>>>
>>>
>> Agreed. It's also unclear if it has Literal or Resource as its range.  In
>> other words, does it point to the Style resource, or does it have a
>> literal
>> which is interpreted relative to some CSS.
>>
>> ****
>>
>>>
>>> ** 2.       **With regard to whether the XHTML attribute class could be
>>>
>>> considered as an RDF predicate (i.e., xhtml:class), a few years back in
>>> the
>>> development of RDFa this was considered:
>>>
>>>    http://www.w3.org/2006/07/SWD/**track/issues/3<http://www.w3.org/2006/07/SWD/track/issues/3>****
>>>
>>>
>>> So probably too much baggage to try and use xhtml:class. Too bad. Unless
>>> you read this discussion differently.
>>>
>>>
>> Also agreed, unfortunately.
>>
>>
>>
>>  So I guess oa:styleClass it is, unless someone has another take on how to
>>> avoid.
>>>
>>> ** 3.       **But there will not, I assume, be an oa:styleID predicate?
>>>
>>> We would limit ourselves to CSS class selectors?
>>>
>>>
>>
>> That's a good question.  I'm not sure that there's a meaningful
>> distinction
>> between them, given the differences between RDF and [X][HT]ML.  Either the
>> ID would need to be globally unique, which would be pointless, or not, at
>> which point it would be the same processing model as class.   So my
>> thought
>> would be no, but if someone has a rationale for including it, then it
>> would
>> be coherent.
>>
>> On that front, a way to avoid the Specific Resource requirement would be
>> to
>> allow predicates as elements in the CSS selector slot.
>>
>> eg:
>>      oa:hasTarget { color : red }
>>
>> Would mean take all of the targets and apply the CSS block.  This would
>> apply to all multiple bodies/targets, but when there's only one would
>> avoid
>> the need for the Specific Resource.
>>
>> Rob
>>
>>