Re: New Requirements Draft

> In so far as I am concerned a variant is simply "yet another version"
> and requires no special handling outside of the normal versioning
> mechanisms.

And the body of literature I just pointed you to says exactly the opposite: 
variants handling requires special constructs and variants are not "simply 
just another version".

> As for your assertion Andre, that if web sites had proper variant tools
> they would use them, this would contradict your previous statement that
> variant support is widespread through versioning systems. Most major
> versioning systems hook directly into the Internet, so apparently
> variant support isn't so terribly important.

No contradiction whatsoever. There are namely two ways CM systems hook into 
the web:

   * they have some applet or HTTP FORMS thingy that allows their CM policy to
     be enforced. This is totally orthogonal to what WebDAV is trying to do.
   * they manage a set of web pages by some simple extension to the URL
     mechanism. Given the limited power of URL's, variants handling is totally
     out of the question (and so are lots of other CM processes etc).

Thus, the problem of making variants explicit in HTTP, and adding it as a 
primitive to the protocol has not been solved yet, and that is what WebDAV 
should be addressing.

> Let us be clear on the issue, variant supports means creating mechanisms
> which allow for a client to instruct a server on how it should handle
> accept headers. I believe this to be a server configuration issue and
> out of scope for WebDAV.

Well, this is your opinion. I believe Judith, Martin, and I have a compeltely different definition of variants, and I believe WebDAV should at least address the issues we raise.

=== Andre ===

Received on Wednesday, 27 August 1997 15:39:34 UTC