[tei-council] xml:space= -- harmful to our health? (was "Re: comments on ST")

[Syd Bauman]

> None of these however is concerned with text. 

That depends on your definition of text, I suppose. (Remember "text
is not a special kind of data, data is a special kind of text" :-)


> >     * DocBook -- only "preserve" on certain elements
> this is the closest to TEI, so the fact that it does seem to want
> xml:space is interesting...

Indeed. Remember first and foremost, though, that DocBook is about
*authoring*, not capturing extant texts. But the point I was making
is that even DocBook doesn't make this attribute globally available
or permit the value "default". DocBook only makes this attribute
available on 8 of its 390+ elements: <address>, <classsynopsisinfo>,
<funcsynopsisinfo>, <literallayout>, <modifier>, <programlisting>,
<screen>, and <synopsis>. And while xml:base=, xml:id=, and xml:lang=
are in the set of ~20 "common attributes", xml:space= is not.

I am disappointed, though, that there is almost no documentation at
all about its use in _DocBook: the definitive guide_. (At least, not
mentioned in the index or glossary.)


> >     * DHQ -- only on <eg>
> isnt that TEI-related in some small way?

Like many other schemas in the world, it is based on TEI, but is not
TEI. (At least currently.)


> >     The only languages I know about that seem to have it declared
> >     more generally are out of the W3C:
> >     * XHTML 
> >     * SVG

> Both rather crucial to the TEI community methinks? SInce we
> actually recommend using SVG, it doesnt seem unreasonable to at
> least anticipate some users for it? In which case...

In which case what? Because a TEI user can find xml:space= available
on <svg:definition-src> she would expect it to be available on
<catRef>, too? (Yes, they are both empty elements, so xml:space=
seems nonsensical.)

SVG has unthinkingly included xml:space= on every element, and they
look foolish for it. I'd like us not to make the same error.

XHTML, like DocBook, has used it sparingly and with some thought. It
is available on <script>, <style>, and <pre>, and they also do not
permit the value "default".


I realize you just slammed the door on changes to the schemas, but I
fear that putting xml:space= in att.global is an egregious error, but
not one that is easy to fix in 1.1. (It would be much easier to put
xml:space= in where it is needed in 1.1 than to remove it from
global, due to backward compatibility concerns.)