RE: "Binary XML" proposals

["Christopher R. Maden" <crism@maden.org>]

At 15:42 10-04-2001, Al Snell wrote:
>On Tue, 10 Apr 2001, Christopher R. Maden wrote:
> > As you noted, any binary XML format must contain exactly as much
> > information as the text form.
>
>Wrong. Exactly as much *useful* information. The binary format can miss
>out a whole load of redundant cruft that helps fleshlings follow the plot,
>but has to be located and discarded by a computer.

This is what "information" means, in the entropic or information 
sense.  Compression science is about trying to find the smallest 
representation of the same information (or, in domain-specific cases like 
JPEG or MPEG, determining what information is OK to lose when the end 
processor will be a human brain).  What you're working on comes down to an 
efficient compression algorithm... and those already exist.  You can 
implement compression on a filesystem or over the wire regardless of 
format; it's not clear to me where the win is in precompressing.

-Chris
-- 
Christopher R. Maden, XML Consultant
DTDs/schemas - conversion - ebooks - publishing - Web - B2B - training
<URL: http://crism.maden.org/consulting/ >
PGP Fingerprint: BBA6 4085 DED0 E176 D6D4  5DFC AC52 F825 AFEC 58DA