Re: Abbreviated Tag Names (ASN.1)

[Tim Bray <tbray@textuality.com>]

At 10:06 AM 19/02/01 +0100, Peter Jacobi wrote:
>IMHO there is clear divide between use cases which would profit more 
>from a "binary XML external form" (ASN.1 or not) and those better suited 
>for general compression:
>A) Data a) consisting mostly of numbers, enumerations, bit sets, 
>booleans and the like which b) conforms to Schema without #any will 
>shrink dramatically with a custom binary encoding.
>B) XML instances having large amount of "textual" data and/or some 
>#any will better do with gzip, as the custom binary decoding usually can't 
>do anything about the redundancy of natural language texts. (And can't 
>avaoid coding in the tag and attribute names in the #any part).

Nobody's opinion is worth much.  This area is full of surprises and
counter-intuitive results.  And the results might depend more on
the specifics of the language and the size of the messages and other
things you haven't considered.  Having an opinion isn't worth the
time and effort until you've done some measurements. -Tim