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

Re: Notes on accessing past versions

Antti-Juhani Kaijanaho wrote:
> On 20010301T165336+0100, Benjamin Fallenstein wrote:
> > One catch: Cell c is a normal cell, cell v is from an old version.
> > Suppose you do:
> >
> > c.connect("d.1", v)
> >
> > After that, wouldn't c.s("d.1") give v, but v.s("d.1", -1) not? That
> > violates the ZigZag principle, which is a bad idea.
> Hm, yeah.
> I think I'll need to add a way for a cell to veto a connection :-)

Well, the connections can't really be the cells' responsiblity; it's too
much of a drag to ensure bi-followability that way. I'm more for
introducing a special case into the ordinary dimensions for this. If,
OTOH, we generally want spaceparts to be able to use the ordinary
dimensions, one possibility would be:

* Normal cells are by default not connected on the ordinary dimensions.
* For spacepart cells, the default is set by the spacepart.
* However, you can break and remake connections as you like in both cases.

As another option, a connection made by a spacepart could be made
un-breakable, but edges not connected in spaceparts could be freely modified.

- Benja