[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]

Re: [zzdev] Re: [zzdev] Subspaces & ZZ versioning



Rauli Ruohonen wrote:
> On Wed, 27 Dec 2000, Antti-Juhani Kaijanaho wrote:
> > On 20001227T115303+0100, Benjamin Fallenstein wrote:
> > > For general use in versioning etc., I would not use the closuring: for
> > > example, you'd want to be able to take the whole tree starting at the
> > > "AllViews" cell as a subspace, but it isn't a closed set in any way that
> > > makes sense.
> >
> > Well, just define an empty set as the set of hard dimensions.  Then there
> > will be no closures.
> 
> But then you have to specify all the cells and dimensions explicitly.
> Using semiclosures (in the sense of Documentation/Spec/zzspec) would seem
> better. It is more general, covers the "AllViews" case and seems "natural".

Yes, it does cover the AllViews case, because we have a tree there, but
that's hardly a general solution. Take the Holm Family Demo, and assume
we want to have a subspace containing one core family. Or some other
"strange" structure.

I don't see the problem. I mean, we don't need to specify the cells and
conns by hand; we can always write an algorithm that does that for us,
and it's expected a lot of these will be written. For example, you want
to be able to select a whole Nile stream at once. If we need a
mathematical definition, we can do that easily. If we say "Z' is a
subspace of Z with respect to a set of cells X and a set of connections
C iff...", then we can say "Let Z' be a subspace of Z' with respect to X
and C, where C is the set of connections in Z between elements of X
along a dimension which is element of D, and X is..." -- that kind of thing.

-b.