[
Lists Home |
Date Index |
Thread Index
]
On Wednesday 05 February 2003 14:34, Dare Obasanjo wrote:
> Your list is one of the reasons I consider attempting to standardize on a
> globally accepted binary XML format as a fool's errand. Various
> applications have different requirements especially when it comes to size
> vs. speed and I've yet to see a binary XML proposal that satisfies both
> camps let alone the other interesting requirements that you site. I can't
> imagine that trying to please all those requirements would spawn a creation
> not unlike W3C XML Schema which attempted to satisfy too many conflicting
> requirements and ended up a satisfying them in an inadequate manner,
Ah, read your history! The same dilemma faced the designers of ASN.1 so they
defined a common data model with multiple encodings.
In the ISO OSI model, two communicating parties would (as part of the
protocol stack) negotiate which encoding to use based upon what each end
supports and prefers, so there weren't interoperability problems. Everything
supported the Basic Encoding Rules and as many other rulesets as applicable.
There wasn't a problem of "but what if I find some PER data lying in a file
with no metadata?" because in that case, you shove on a header with metadata
in. In a communications protocol that session state was negotiated
transparently on connection by the lower levels, like the window sizes and so
on in TCP and session keys in TLS, so it didn't need to go into the messages
themselves.
ABS
--
A city is like a large, complex, rabbit
- ARP
|
