[tei-council] another High Noon proposal

[Syd Bauman]

LB> My recommendation, for the @url case, is
LB> a) remove @url from the att.media class b) define a new class
LB> att.resource (or something) which provides the @url attribute c)
LB> make the elements which need @url specify their membership in the
LB> new class

I like this idea and think not only is Lou's reasoning sound, but the
reverse is even more likely: to wit, at some point in the future
someone is going to want to point to the target of their <note> or
<ref> by some means which is not a URL.

I think att.resource is reasonable, given that the name of the
attribute stands for Universal Resource Locator, and it is defined to
be a Universal Resource Identifier. (And using att.identifier would
drive us nuts! att.locator might not be so bad.)


SB> ... what I'm worried about ... is ... a customizer being baffled
SB> by how to add the attribute back.
SR> they can't. its as if it never existed.
KH> Actually, I believe the way for them to "get it back" is to
KH> re-define the attribute for just that element. But this means
KH> they won't inherit any revisions to that attribute's definition
KH> as they would with attributes in other cases.

But as Lou and I have separately pointed out, this is confusing. (Why
do I have to define @bar of <foo> in my own namespace when <foo> is a
member of att.duck which (normally) provides @bar?)

So

a) I like the idea that if <attDef mode="delete"> is used in the
   Guidelines it should be used only in cases where we are quite
   confident that users will not want that attribute.

b) I like Lou's suggestion of att.resource meaning we don't have to
   decide this issue forevermore right now, nor do we have to figure
   out what the output of a tagdoc that has <attDef mode="delete">
   looks like for at least the next few months. So the good stuff
   happens right away and the hard stuff is deferred.

LB> but since we don't actually need to use this mechanism, what's
LB> wrong with my earlier suggestion of providing @url via some other
LB> class?
SR> nothing, that's fine. can you activate it?

If no one else wants to implement this, I'd be happy to. (Although
maybe I should read TCW 20 first?)