Lists Home |
Date Index |
"Bullard, Claude L (Len)" wrote:
> I'm not reading intent. I see trends. Tim Bray et al's skunk document
> makes it clear enough that some see removing DTDs as the best
> way to put namespaces in the core. The ways DTDs have to satisfy
> namespace requirements now are clunky.
the way java parsers and dom implementations support namespaces is
clunky. which means that dtd support for namespaces can't be much
better. (if i might observe, schema support for namespaces in java is
but marginally better.)
> There are competing
> language efforts for the metainformation space which is a niche
> space. Some do intend to do away with DTDs, have for a long
> time, and they have been vocal about that.
> But that isn't the point. Until and unless something like
> the skunk document is accepted, DTDs are here to stay. Until
> and unless something like DSDL succeeds, they are here to
> stay in their current form. That makes them a formal antagonist
> to namespaces, formal in that they are in core and namespaces
> are not.
in point of fact, there is no difference between cl-xml based code which
pertains to namespace-aware dtds and code which pertains to
namespace-unaware dtds. everything is expressed in terms of universal
names available to the programm at (as early as) "compile" time.
is there anyone still reading this thread who would use namespace-aware
dtd-encoded document definitions but does not for lack of a java
implementation? by which i don't mean support for the
i wonder if the only way to respond effectively this thread is to port
the cl-xml code to java.
if there is interest, speak up about which parser/dom you are using.
> We can dance around it as much as you like, but
> for namespaces to be core, DTDs have to go or be reengineered.
> Given that, waiting is best.