On 10-09-09 02:22 PM, Lou Burnard wrote:
>> On 9 Sep 2010, at 21:18, Martin Holmes wrote:
>>> My argument is that this very much IS the TEI's business; we spend so
>>> much time describing in painful detail (yet idiosyncratically) the
>>> appearance and rendering of text and objects that we should simply adopt
>>> the CSS standard natively
> I'm not sure I buy this. Certainly a lot of time gets spent describing
> various aspects of a text in a rich TEI encoding of it, but I would not
> agree that these are necessarily or exclusively concerned with its
> appearance or rendering.
> And didn't we spend quite a lot of time a year or so back defining ways
> of using CSS within a TEI document via the @rendition global attribute?
> Are you proposing to junk that, or to add a third (fourth if you include
> Sebastian's suggestion of using the HTML namespace) way of skinning this
> particular cat?
Using @rendition forces you (IIUC) to abstract the CSS into something
resembling a CSS class, which you then point to. In practice, we find
that most things we want to describe are not similar to other things,
and it's more straightforward to describe them in-place (which is more
like html:style). Once a large document has been marked up, it's
possible then to process it in order to discover items which share CSS a
ruleset, and then replace the inline CSS with a centralized reference,
but there are still lots of individual items left over.
What I'm suggesting is that @tei:css would be better than the current
@rend, because it uses a formal language for which validators and
parsers exist; that it could ultimately replace @rend, but could happily
coexist alongside it for as long as necessary; and that it's a natural
partner for the @rendition system, allowing inline styles alongside
centralized classes (just as we have in HTML).
University of Victoria Humanities Computing and Media Centre
([log in to unmask])