[
Lists Home |
Date Index |
Thread Index
]
Umm... PDF and Flash aren't painting to the
same device contexts are they, say one in which
a 2D circle is drawn alongside a 3D spinning
sphere? I get in over my head at this point,
but as I recall, the MS experience with Chrome
was less than efficient, in fact, a dog.
So, we can level the playing field but not
the players. In some sense, a composite
document is like an all-stars game in which
the problems of behavioral fidelity and
performance can only be met by the very
best players, and then only in a given
location on a given day. It is
rough enough to get all the components
to work realiably, the composite forces the author
to also consider if his audience has all
these players, is willing to download
what he doesn't have, or says, Pass.
So it becomes very quickly, caveat vendor.
len
-----Original Message-----
From: Jonathan Borden [mailto:jborden@mediaone.net]
> Reliability and behavioral fidelity are twins.
> This sort of thing really outs in real time
> 3D rendering systems. The abstract idea of
> namespaced composite vocabularies is great;
> I am wondering if it is a concept that works
> fine for data systems that do the sort of thing
> relational dbs do with tables, but falls apart
> the closer one gets to the renderer in the
> pipeline.
>
Optimization and caching. That is the job of the renderer. I assume a
renderer would not download a plugin each time a namespace qualified element
were encountered. At least I know that my browser does not download the PDF
or Flash plugin each time a document is encountered. Heck a really smart
renderer might even optimize across downloaded plugins.
I just want to create a level playing field in which namespace names can be
used as URIs and in which any inefficiencies so created can be optimized
away when something is demonstrated to be an actual rather than theoretical
bottleneck (remember that Ma Bell said the Internet would never work on
theoretical basis, perhaps it doesn't but nonetheless here we are.)
|
