Re: More versioning thoughts.
David J. Fiander (davidf@worf.mks.com)
Thu, 06 Jun 1996 08:25:19 -0400
> * It looks like different users are trying to push different capabilities.
> Although expected, I think we should agree on a common `goal' for
> versioning in HTTP. Are we trying to simply put versioning capabilities
> in HTTP, are we trying to put *a* versioning capability in HTTP, are
> we trying to seamlessly integrate versioning and non-version aware
> clients/servers,etc. It looks like the group needs an objective very
>much;
> this in itself could be an interesting exercise, because I think we
> will be able to identify multiple layers of functionality that can be
> built on top of eachother.
I agree entirely, which is why I've been recommending defering
discussion about the type text/config until we have some basic
versioning stuff laid down.
> * An a start (?), to me, putting versioning into HTTP means to put
> versioning
> facilities into HTTP, that allow version-aware clients/servers to
> implement a particular versioning strategy suitable to their needs.
> Thus, capabilities like LOCK and UNLOCK are needed as predicates, but
> do not necessarily need to be used. An RCS like implementation of
> versioning of web-pages would use it, David's approach would not.
Well, I also said in one message that we clearly need to add
methods. Right now MKS's "Integrity Engine" is using RCS to do
locking and unlocking, but we're not using new HTTP methods, for
historical reasons. What we're looking for from the group is a
specification that allows versioning-aware clients to interact
with versioning-aware servers, regardless of where you bought
each.
> * Thus, I like to make a distinction between versioning in HTTP, which is
> a communication protocal, and versioning of web-pages, which is a much
> higher level problem. It seems like the two are getting confused in some
> of the earlier messages.
True. Right now, in the immediate-term (even shorter than
"short-term"), Jim and I have proposed an HTTP URL syntax to
allow clients to specify a version number. This allows
versioning-clueless clients to at least GET old version of pages.
The server could then do something sensible if the user PUT a
page with a versioned URL.
> Before I start rambling too much (which I probably already did), I think the
> group needs to have a common set of goals, possibly layered in some
> hierarchical plan of attack, as well as a clear separation between versioning
> features in HTTP, and higher level issues.
>
I agree entirely. I'm starting to wonder at the silence that
followed my suggestion, which I made last week, about developing
a list of goals for the group. I guess everybody is heads-down
on getting ready for the IETF meetings in Montreal and working on
the HTTP specification.
- David