Re: Application dispatch for application/xml

Hi Martin, thanks for the feedback.

On Sun, Jan 25, 2004 at 01:35:48PM -0500, Martin Duerst wrote:
> Your web page says that you want to adapt this to current
> behavior of user agents. This wasn't clear in this mail.

Sorry, I thought that was implied.

> I'm also not sure that we already have enough experience
> with this; one indication is that the only namespace that
> you seem to be testing is XHTML.

I have to profess to being rather an (X)HTML bigot.  If somebody would
like to offer their assistance with setting up tests for SVG, MathML,
etc.. that would be very much appreciated.

As for "enough experience", I don't know if we do or not, but is
that important?  I mean, what I believe we need to do is simply document
current practice, and there's *always* a current practice, no? 8-)

> Also, you write:
>  >>>>
> Firebird appears to use namespace dispatch with a fallback (for
> unrecognized or absent namespaces) to tree view. This seems an
> obvious bug, as the semantics of such an interaction are a function
> of the capabilities of the recipient, rather than being visible to
> the world at large.
>  >>>>
> 
> It is totally unclear to me why you would want to call this a bug.
> It seems to be the most natural thing if we want to have any kind
> of dispatch at all (where we very clearly have to assume that not
> all users will have the same capabilities installed) and want to
> make use of the fact that we know it's XML.

Hmm, I'm sure there was a reason I wrote that, but I can't think of
it right now.  Your explanation seems quite sensible.

Mark.
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca