[
Lists Home |
Date Index |
Thread Index
]
Bill de hÓra wrote:
> Mike Champion wrote:
>> Yeah, this gets into the "XML: Syntax or Infoset" permathread, but I
>> agree that "lossless translatability" to the 1.0 syntax is the most
>> important criterion for determining whether something has "XML nature".
>
> Well it's not lossy that worries me so much (the Infoset is quite clear
> on that); rather having to have bInfoset codecs to hand. Today all I
> need is an XML parser. With bInfosets I have no idea how many I'll need.
> All that value Tim Bray talks about re blame allocation is eroded when
> someone tells me I need version X.Y.Z.spK of their Infoset codec.
That is at the heart of the issues that the workshop needs to discuss.
The usefulness of bInfosets increases when they can host
vocabulary-specific codecs (eg for SVG path data) but interoperability
demands that these be available.
Several groups have devised solutions to that problem, with varying
degrees of success. It is certainly an interesting area of research.
--
Robin Berjon
|
