RE: Versioning goals doc

>From: "Babich, Alan" <ABabich@filenet.com>
	
>I'm also of the school of thought that it would not
>be feasible to have a design based on additive diff's
>for all formats, not even for all "text based" types:
>I use a well known word processor that has a proprietary 
>format, and that WP itself can't deal with diff's 
>correctly in some cases. And the proprietary format of 
>the WP changes from release to release.


None of these objections, yours or Jim's, goes to the heart of the point
that I was actually (trying) to make. Additive diff processing is obviously
not going to be implemented by every server -- and that's OK. However, 
DIFF/PATCH method allow efficient client update protocols for large
documents and also for useful end-user features, and (for some servers, an
some media-types) also allow the support of additive or change-set oriented
versioning). Files that contain checksums can be DIFFed by a straight binary
diff on octets, even if those DIFFS can't be added sensibly. That's OK.
There are surely media-type specific formats that can be developed for such
data types if there is value in it (and a lot of people find that there is).

So the fact that you can't do everything with every kind of data is a more-or-
less completely irrelevant fact. You also can't search textual content in GIFs
of page-images. That may affect your choice of data formats, but it doesn't
mean that you have to eliminate full-text searching, because it won't  work for
GIFs!

I'm not proposing that WEBDAV define additive versioning for arbitrary file
formats, but thgat the protocol not prevent interoperable server
implementations of such capabilities. There are lots of things that make sense
for a server to do for some media-types and not for others, and that's not
the issue her.

>One could defined an additive diff format for straight
>ASCII text (and people have done that), but a design
>that works for that one type and no others is not
>adequate.

As I said, I can define a format for intensional versioning of arbitrary
binary content, and for representing changes to integrate into that
representation. I'm not pushing my particular format because I'm not religious
about that, and defining such a format is not rocket science, just plain
old work.

  -- David

------------------------------------------+----------------------------
David Durand                 dgd@cs.bu.edu| david@dynamicDiagrams.com
Boston University Computer Science        | Dynamic Diagrams
http://www.cs.bu.edu/students/grads/dgd/  | http://dynamicDiagrams.com/
                                          | MAPA: mapping for the WWW
                                          | http://dynamicDiagrams.com/minimapa

Received on Friday, 2 October 1998 19:59:47 UTC