|
Re: [xml-dev] Microsoft FUD on binary XML...
|
[
Lists Home |
Date Index |
Thread Index
]
Dave Butler wrote:
Somebody was asking about tests
between ASN1 and XML.
I work in an industry where
resticted bandwith, memory and butget push us towards binary. When
dealing with mobile solutions particularly satallite based D+ or Fleet
costs really add up when sending large XML files. Tx rates are slow
and if the byte charges dont get cha the connect time will.
Here is a study by some guys a while
ago;
dave
It would be interesting to know what their data was: was it binary
payloads
encoded as Bin64, decimal numbers, dates/coordinates in some form?
I don't see the that the first complete paragraph on page 2 is accurate:
it misses out that there has always been a strong group of opinion that
adoption of
XML willy nilly for all data interchange is just dumb: "terseness
is of minimal importance" is not a statement about the universe
but a statement of the goals of XML: that XML is explicitly not
designed to be useful when terseness is important (though it may be.)
In particular, XML's tersness goal is a response to building in
clever methods of reducing file sizes using schemas (e.g. tag ommission)
which SGML allowed, while still being text. Using schematic
info to reduce file size is not new, it was XML's departure point
that it should be (left to) some later layer (fitting in with HTML's
capabilities, in particular.)
IOW, XML is a technology for the WWW. If your application
is not WWW then don't feel cheated if you have to transform
the data to meet pragmatic requirements. I don't see anything
wrong with having optimized native formats and public XML
bindings. If it makes design sense to have a very simple relationship
between the optimized binary format and the public XML, why not.
Cheers, and apologies for posting so often this week
Rick Jelliffe
|
|
|
|
|