Versioning terminology

>>Thus, the language in the specification, from Section 9.2 "Versioning Data
>>Model" should be used in the requirements.
>
>My problem with this is that the spec's usage seems counterintuitive.  I
>think it will confuse people.  To me it seems natural to say that a node in
>a version tree is a version (not a versioned resource).  A versioned
>resource is a resource that has versions -- it's not any one of the
>versions.  So I guess the closest thing to a versioned resource is the tree
>handle.

Well, while I agree with you that the language in the spec. is probably not
intuitive, I'm not fond of just using the term "version" for a node in a
version tree.  Since it took the group awhile to agree that a node in a
version tree is also a first-class resource, I'd like our terminology for
the node in a version tree to reflect this.

I'm also not fond of saying that a node in a version tree is a "version of
a resource."  This is because we may allow an individual resource to be a
member of more than one version tree, and because it may imply (for people
used to file-based versioning systems like RCS) that the version is not a
first-class resource.

I'm tempted to bypass this terminology thicket and coin a new term, like
"movert," "movet," or "MVT," which is just an acronym for "member of a
version tree."

- Jim

Received on Thursday, 20 February 1997 13:14:33 UTC