Lists Home |
Date Index |
On Thursday 24 January 2002 10:12 am, Mark Baker wrote:
> Right. But I'm only interested in the serialized structure, not
> anything to do with activation and binding with that structure
> (Bento vs. Opendoc).
I think they're somewhat related though I understand the point you're
making. Bento was good primarily because it did/assumed very little. I
assume you're thinking along the same lines? If so, then this is
really just going to be another form of packaging. Anything beyond
this starts encroaching on the application space.
> That depends. If any subdocument acts as a container, and the
> element of that subdocument which does the containing is required to
> be processed (e.g. smil:skip-content="false"), then if a suitable
> compound processor cannot be constructed, processing must fail.
I think that would depend on the application though.
> > Defining a dispatch model is very good. A useful step, but it
> > ignores the problem of *building the application* and *packaging
> > the application*... and claims that "we'll just download the code"
> > don't really work, because you *still* can miss components.... or
> > you open yourself up to the well-known Trojan horse problem with
> > embedded downloadable code.
> It doesn't "ignore" those things. It says those are orthogonal
> issues; the "what" versus the "how".
I think these, in particular, are closely tied.
> I understand what you mean, but I'm not convinced of the value of
> it. Anyhow, that's a separate topic.
Perhaps so... I think being able to constrain processing to be an
important part of packaging applications.
> For packaging, I'm assuming a compound document with containment
> purely by value.
The point I was making is that in the context of a given application,
dispatch is usually a trivial part of the overall application.