Cart
[Home] [By Thread] [By Date] [Recent Entries]
A smart system design degrades gracefully. That's just good practice. Arguing the customer is stupid is a bad way to create requirements. Again, tell your wife she is a result of "worse is better" requirements. We can argue a scale of reliability, but given that system means exist to declare IDs, and that persons are now suggesting a different means, unless they are just pursuing their own visions, there should be a sensible reason. XML has built in unreliable mechanisms but if we are to preserve unreliability as a system feature, we could go away and say done. It is. One could say "efficiency realized as a consequence of frequency of use of this type of information across the system". Maden and Bray's arguments about "the success of the web" are speculation, not fact or requirement-engendering. We need to avoid that kind of argument. len -----Original Message----- From: Champion, Mike [mailto:Mike.Champion@S...] Sent: Monday, November 05, 2001 9:51 AM To: xml-dev@l... Subject: RE: Re: determining ID-ness in XML > -----Original Message----- > From: Bullard, Claude L (Len) [mailto:clbullar@i...] > > ID resolution in this case should be reliable. Christopher Maden, and Tim Bray (repeatedly!) made the point that the success of the web was founded on the NON-requirement for reliability. Indeed, the whole point of the internet itself was to do as well as possible in a non-reliable networking environment. I respect the fact that Len (apparently?) gets paid to build designed-in reliable and secure solutions, but this cannot IMHO be a requirement for infrastructure such as XML. I suspect that Len profoundly disagrees and I'm gonna get my head handed to me We need a way to define ID-ness that works in an environment where the supplier and consumer of the XML both agree on the definition, but that degrades gracefully if one or the other do not.
|
