ietf-xml-mime
[Top] [All Lists]

Re: Announcement of an I-D

1999-09-30 08:43:41


MURATA Makoto wrote:

Chris,

I think that this is a very interesting issue.  Let me ask some questions.

Q1.  If you introduce a media type "image/svg",  you can certainly introduce
SVG-specific fragment
identifiers. 

There are two issues here.

Firstly, since it is defining a media type, it needs to define what the
fragment identifier syntax is. We use XPointer, but in this release
restict it to pointing to IDs. This is because we want a spec that is
fully implemented very rapidly.

Later versions will probably remove this restriction, once XLink and
XPointer are more established.

Secondly, since XPointer wisely moved toa prefixed syntax, ie xptr(), it
becomes possible to define other syntaxes. For example, a means of
linking to particular time periods, or a way of linking to particular
geometric regions, and so on. Thisis sonmething that SVG might be
intereste din later, but we have not done this for SVG 1.0

But are you willing to give up referencing into SVG embedded
into XHTML?

I don't see how I am giving this up. Such a file would not be served up
as image/svg (or indeed as image/svg-xml) - it would be served up as
application/xml, or application/xhtml-xml, or something else.

Q2.  How do you feel about making fragment identifiers namespace -sensitive?
For example, "if the
root element is of the namespace for SVG, the fragment identifier is
interpreted as ...."

That might be possible, but at the moment they are dependent on the
media type only.

Q3.  If we really want to make fragment identifiers namespace- sensitive, we
can even interpret
different portions of a single fragment identifier differently.  For
example, the first two conditions
are interpreted as XPath and the third condition is interpreted as SVG
fragment identifier, and

You seem to be under the impresion that SVG has some special fragment
syntax. It has a proper subset of XPointer.

the forth condition is intepreted as XPath (i.e., references into XHTML
embedded in the SVG).
How do you feel?

I would like to understand what you mean by "SVG fragment identifier"
before being able to answer how I feel. I need to understand your
question better first.

--
Chris

<Prev in Thread] Current Thread [Next in Thread>