- From: David G. Durand <dgd@cs.bu.edu>
- Date: Mon, 28 Sep 1998 14:51:51 -0400
- To: WEBDAV WG <w3c-dist-auth@w3.org>
At 12:04 PM -0400 9/28/98, Jim Whitehead wrote: >One of my major concerns with a change-oriented approach is that it tends to >assume the system has a lot of content-type-specific knowledge, and it tends >to assume you're dealing with text-like objects. This is contrary to one of >the underlying design principles of the Web, which is that operations are >media-type independent. I have three reactions to this: One is to agree with Sankar's points. (But I'll actually belabour them a bit in my own words) The second is to point out that any object stored in a computer can be treated as a sequence of octets, and changes to those sequences are _always_ a possible, if sometimes far from optimal, way to represent changes. So, on the face of it, your assertion is (trivially) false. "Text-like objects" is a term that makes a fundamental fact of computer encoding look like special pleading for a favored data format. The third is to say that the notion of sequence of characters is in fact a significant special case deserving of meaningful treatment (as, for instance the MIME standard's attempts to provide useful rules for the text/* types). So, even if you don't believe my argument that there is no special pleading for text, I think that there's a very good claim that such special pleading is legitimate anyway. The possible need (or desire) for content-type-specialized change formats is a separate issue, and has exactly nothing to do with the accomodating change-set style versioning, especially given the existence of a lowest-common-denominator chnage format. There are many possible ones that would work for arbitrary octet sequences (as for instance, the as yet unreleased VTML 2.0). Even for text, there is a lot to be said for an octet-sequence approach, given the variety of modern character encodings. I can think of single-octet character codes, varying length codes like UTF-8, 16-bit codes, 16-bit codes plus the use of "surrogate characters" to map pairs of 16-bit characters to a larger portion of the ISO character space. And someone somewhere will evantually want some other variation for some reason... -- David _________________________________________ David Durand dgd@cs.bu.edu \ david@dynamicDiagrams.com Boston University Computer Science \ Sr. Analyst http://www.cs.bu.edu/students/grads/dgd/ \ Dynamic Diagrams --------------------------------------------\ http://www.dynamicDiagrams.com/ MAPA: mapping for the WWW \__________________________
Received on Monday, 28 September 1998 14:44:29 UTC