[tei-council] http://purl.org/tei/fr/3519866 (@rend datatype)

[Martin Holmes]

On 12-07-05 11:27 AM, Lou Burnard wrote:
> On 05/07/12 18:09, Martin Holmes wrote:
> s wrote:
>>>
>>>> I don't think that's the only solution. Surely @style (content = CSS)
>>>> would solve the problem for those who want to use inline CSS?
>>>
>>> bother me a bit that we special-case CSS with this; and it is also not
>>> clear
>>> whether you can have @rend _and_ @style, ie what are precedence
>>> and inheritance rules?
>>
>> That's up to the project, surely. Most would presumably only use one or
>> the other. If you're wondering what the default TEI stylesheets should
>> do, then that's another question; but the stylesheets already have to
>> guess at what any given value(s) appearing in @rend are supposed to
>> mean. I'd be inclined to say that if @style is present, then ignore
>> @rend, but you could also combine the two. Redundancy in an HTML @style
>> attribute is ignored, isn't it?
>>
>
>
> I can't believe we're seriously considering adding a THIRD possible way
> of specifying rendition!
>
> But if we are, did you mean @html:style or @tei:style?

As far as I know, there is no such thing as either of these. The @style 
in XHTML is a regular attribute, and therefore exists in the 
no-namespace namespace; and the same would apply to @style if defined in 
TEI:

"Attributes are never subject to the default namespace. An attribute 
without an explicit namespace prefix is considered not to be in any 
namespace."

<http://en.wikipedia.org/wiki/XML_namespace>

So TEI's @rend attribute is not @tei:rend (it's not in the TEI 
namespace), and HTML's @style is similarly not in the XHTML namespace. 
And as far as I know, there is no way to import an attribute (such as 
@style) from an external schema unless it has an explicitly-defined 
namespace, which HTML's @style doesn't AFAIK.

So there's no way to use HTML's @style anyway. We might define our 
@style to be identical to it, if that suits us, but since that is

style = CDATA [...] The syntax of the value of the style attribute is 
determined by the default style sheet language...

that doesn't help us much. I would prefer to define it as CSS, and add 
some Schematron rules to enforce the syntax. (We don't want to define 
constrain the actual ruleset content because CSS is a developing standard.)

> Presumably the former is always a possibility (thus giving us FOUR ways
> of saying 'output this in "Incredulous Bold"')

I think we have only @rendition and @rend so far. @style would make 
three, but it has significant advantages over @rend because it uses an 
existing and powerful standard, and is easy to handle in rendering.

Cheers,
Martin

-- 
Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)