- From: Bjoern Hoehrmann <derhoermi@gmx.net>
- Date: Fri, 17 Sep 2004 21:27:35 +0200
- To: Larry Masinter <LMM@acm.org>
- Cc: "'Roy T. Fielding'" <fielding@gbiv.com>, uri@w3.org
* Larry Masinter wrote: >We have a problem that is over constrained -- it is impossible >to accomplish all of the desired goals of clarity and compatibility >simultaneously. I do not think so. I do not think that any of the changes from RFC2396 in this regard make the spec more clear, rather more confusing and as with all changes necessarily less compatible. Only few people would expect that the term "URI" does not include "relative URIs" and only few people would expect that the term "relative URI" does not exist at all in the URI specification. I would rather say that the desired goals of clarity and compatibility can easily achieved simultaneously by not applying the proposed patch and changing the definition of URI to include relative URI references, for example. It is my experience that most people expect the term to be defined that way and I think that there is a lot of evidence for this assumption in the collection of W3C Technical Reports and many other resources. The same applies to the assumption that there are things we can call "relative URIs", there are tens of thousands of web pages that use it. I would suggest there need to be tremendous benefits from changes that are incompatible with previous specifications and and/or common usage; it is not clear to me from the message Roy provided as background for this decision what the actual benefits would be, considering all the terms like broken, fudge, awkward, nonsense, terrible, big mess, etc. in that message it rather seems to be a highly subjective point of view than a clear discussion of both the technical and social benefits. I do not really see such benefits deriving from the changes either. >If definitions of terms in specifications were required to >be "compatible", we could never make any changes to any specs >at all! That's commonly referred to as stability which I think is an important aspect of a stable specification. In fact, that's why they are called stable. But I might be missing your point here, I think of RFC2396bis as a Second Edition while you seem to imply that it is rather a new version of the specification, like URI 2.0. Then I do not think that there should be a new version of the specification without changing the core terminology from URI to something else or major new features.
Received on Friday, 17 September 2004 19:28:25 UTC