Cart
[Home] [By Thread] [By Date] [Recent Entries]
Yes. That which can't be decided once and for all but is recurring is negotiated. Protocols rule in the case of recurring pareto suboptimality. The system is dynamic and will rebalance itself as necessary. The early bridge systems seem to survive the longest according to http://www.rotman.utoronto.ca/~baum/workshop/Not_All_Equal.pdf and that is interesting. The reinvention of SGML in XML was successful. Attempts to replace SGML weren't. Augmenting XML with smaller networks of locally optimal versions of itself that remain compatible may produce some perturbations to the equilibrium of the XML membership, but could strengthen the utility value of the Internet. Cool. We'll make more money. ;-) len From: 'Liam Quin' [mailto:liam@w...] On Thu, Jan 06, 2005 at 03:10:21PM -0600, Bullard, Claude L (Len) wrote: > I suspect the answer will be classes of binaries of which, one > or two dominate the runway. It's possible, in which case the next question whether we (W3C) could (or should) develop a framework and/or best practices guide. A framework might end up being ways to use HTTP content negotiation, and ways to parse XML from an HTTP session saved as a local file, to give a wild example, with the goal that even if a processor couldn't handle a particular document, at least it'd know why, and maybe could delegate conversion in a predictable way. That sort of decision is beyond the mandate of the current Working Group, but (as Rich pointed out) they are providing the necessary background information.
|
