> From: Sebastian Rahtz
> On Mon, Feb 10, 2003 at 10:44:48AM -0000, Stuart Brown wrote:
> > But the TEI is extensible...
> quite. thats the assumption I started from, that the
> description belongs in the TEI, not SVG
Of course, I would never have accused you of anything else.
> > Would it not be possible to extend the TEI's network tagset
> to describe
> > this?
> you're a braver man that me, looking at that!
Well, it would be nice to get a good working example of it, albeit one which
perhaps stretches the original purpose.
> > Let us say I have declared x and y attributes on <node>
> (SVG-style y goes
> > down not up), and a type attribute on arc.
> I am wondering if you should not use SVG namespace on the attributes.
Nice idea, but is it necessary? Does an x or y coordinate lie solely within
the spiritual domain of SVG?, and expressing them as such are you not giving
an indication that SVG is the _only_ way to represent these vectors? My
argument is that SVG, for all its usefulness(ish), became really just (as
you implied) an (Adobe) XML graphics vehicle which happened to be expressed
in vectors, rather than a geniune vectors-expressed-in-XML format. The
expression of character strokes as vectors seems to me to be entirely within
the remit of the TEI's purpose, even if the exact mechanism is not entirely
there, and using <node svg:x="1" svg:y="2"/> implies that this physical
alignment lies outside of the TEI's domain. I could compare it with using an
<html:i> tag in place of <emph rend="italic">.
Maybe this is minor pedantry, but once again it keeps me occupied. It may
look like a bit of an anti-SVG rant, but it isn't, honestly. SVG is nice in
the same way HTML is nice: quick'n'dirty visualization, and damned fine for
that. Surely here (there) at the TEI, though, we are interested in Higher
Things with a Deeper Level of Meaning?
Also, of course, it could be tricky for a user who wasn't experienced in
multiple namespaces, particularly if they're working off the DTD.
 Am I out of line here (pun wholly intended)? Whilst no mathematician nor
a graphics guru, I find that SVG's path= solution to almost all conceivable
vectors is comparable to putting CSS in an attribute rather than using FO.
The vocabulary may more compact than a host of elements and attributes, but
the atomic values are not obtainable directly through the XML infoset (at
least until we have XPath2 in its full datatype-aware glory): it's just
another vocabulary in an XML wrapper.