"Version" Aversion

[Bob Perez <bobp>]

I've used "global identity" to refer to the kind of underlying 
platonic expression you refer to, but after thinking about this 
issue quite a bit (and discussing it with MarcS) I've concluded 
that this only confuses the issue even more. As MarcS correctly 
notes, what really drives the distinctions among the various 
things that people call "documents" is their *intent* with respect 
to those things. 

Attempting to reduce the complexity of intent-driven document 
creation and editing to a simple division between specific instances 
and their relationship to a single underlying platonic expression 
is far too simplistic for the real world, and anyway misses the 
point that matters most to the community of users most interested 
in any particular document: the purpose of that document. 

I like to think of it as a strength of our system that it provides 
direct support for this notion of intent-driven creation and 
editing (e.g., by creating a new Bert for me when I want to work 
on a variation for a new audience), while providing no direct 
representation for the "cloud document" or "global identity" 
notions.

I like Ravi's "edition" suggestion a lot, and plan to incorporate 
it into my work. And while "variation" doesn't deliver quite 
the persusasive impact that "edition" packs, I'm reasonably happy 
with it in the absence of a better suggestion. However, I agree 
with you that we should try and avoid using the term "version" 
to reference both "editions" and "variations". I say "try" because 
I have very little confidence in our ability to moderate the 
habitual abuse "version" receives, either internally or externally.

So, let me summarize my understanding and see how well it maps 
to yours. Berts are the soul of a document and represent the 
discrete, persisting identity behind the changing representations 
of that identity. As changes are made, a document visits different 
states (represented by new Stamps), and whenever a particular 
state is blessed with a Bert hopped from a predecessor Stamp 
(thanks, Nick) we refer to the result as a new "edition" of the 
original document.

When a document is replicated a new Bert is created representing 
a new and separate identity; we refer to this as a "variation" 
on the original document, anticipating a distinct intent by providing 
a distinct identity and state/edition history capability (note 
that the term "variation" is somewhat presumptuous, but not overly 
so since it fits our reasonable expectation of actual changes 
representing a new intent). 

It doesn't seem to add anything but confusion to think of the 
various relationships possible between editions spawned from 
different variations. The act of variation is itself an assertion 
that significant distinctions exist between the resulting variations 
to warrant a new and separate identity. We should honor this 
assertion by abandoning any attempt to draw cloud patterns or 
global identities around the two.

Are we getting closer?

-- bobp