[tei-council] rendition, rend, and style

[John A. Walsh]

On May 9, 2007, at 9:11 AM, Sebastian Rahtz wrote:

> I think that anyone who wants the equivalent of HTML's @style  
> should literally
> use html:style, and it would only be used on born-digital stuff. So  
> lets consider
> just HTML's class vs rend.
>
> If rend is to be used for what we _see_, we'll
> also very often use it for how we render. So <hi rend="italic">
> will want to rendered as <span class="italic">, usually. We
> don't want to force people to say <hi rend="italic" style="italic">,
> so it makes sense for "italic" be a link to a place in <rendition>
> where it is mapped to a) an expansion of what we say (did
> we differentiate italic from slanted?), and b) a rendition class
> (HTML, FO, whatever).
>
> So I am wondering whether we should not just keep the single
> rend, but change its datatype to be a pointer to a composite
> source/ rendering pair?
>

Sebastian,

There seems to be wise agreement not to duplicate @html:style with a  
new @tei:style.  For @rend, what you describe above-- at rend as a  
pointer to <rendition>--is what I was planning to document if there  
is agreement to move @rend more in the direction of @html:class.   
@rend would be different from @html:class because of its (potential)  
linkage with <rendition>.    I think this proposed change to @rend is  
a good idea (and something I've used for years in my P4 extension  
files and P5 ODD files). It helps prevent sloppy use of @rend (e.g.,  
@rend="i", @rend="italic", @rend="italics", @rend="italicize", etc.-- 
all in the same document).  It potentially allows easy automatic  
generation of CSS (or XSL-FO) styles from information documented in  
the <rendition> elements of the TEI Header. But changing the datatype  
of @rend would break lots of stuff, e.g., rendition ladders.  It's a  
fairly radical change to a widely used global element.

Is such a change worth pursuing as part of the Guidelines, or is an  
ODD and documentation about how to use @rend in this manner something  
for a separate paper?

John