RE: Abbreviated Tag Names

[Don Park <donpark@docuverse.com>]

Tim,

You are right if application engineers had complete control over the
transport.  More often than not, they have to work within the bounds
drawn by others.  Whether it is WAP or i-Mode, containts are rather
severe.

Taking WAP as an example, lets say you want to send some XML fragment
to the client.  Embedding arbitrary XML fragment into a WML page will
not work with most WAP browsers and, lacking DOM support in WMLScript,
they are inaccessible.  XML fragment can be sent as hidden WML form
value with reserved characters escaped.  This is where tag name abbre-
viation helps.  If we compress the XML fragment at this point, result
will have to be BASE64 encoded which will negate most of the compress-
ion.  I am pretty sure a similar problem exists under cHTML used by
i-Mode.

Talking to WAP Forum is like talking to old Soviet Politburo committee
about problems of Communism.  You can throw 30K to join the ranks of
engineering-bureacratic-politicians from carriers, handset manufact-
urers, and mCommerce wannabes, but you will be better entertained in
a mental institution.

mCommerce cramps my style, but all the market hypes and VC blinds are
forcing me to chase it like a dog chasing ambulances down the street
to where dot.coms went.

*Woof*

Don Park
Docuverse

> -----Original Message-----
> From: Tim Bray [mailto:tbray@textuality.com]
> I'm sure others have asked, but I don't seem to have seen it; this
> whole approach seems questionable to me.  Tags are after all text,
> highly repetitive text at that, and so is what goes in between them,
> and between the quote marks of attributes.  Wouldn't you come out
> way ahead just applying LZ or equivalent to the whole chunk of XML?
> Compressing is a little CPU-pricey but decompression is nearly free,
> and ought to present no greater software load than the code that
> knows about magic abbreviated tags.
> 
> Or am I missing something? -Tim
>