Fw: Entity membership Jonathan S. Shapiro (shap@eros-os.org)
Tue, 6 Jun 2000 22:55:57 -0400

This should have gone to the list.



From: "John C. Randolph" <jcr@idiom.com> To: "Jonathan S. Shapiro" <shap@eros-os.org> Sent: Tuesday, June 06, 2000 6:53 PM
Subject: Re: Entity membership

>
>
> Begin forwarded message:
>
> > From: "Jonathan S. Shapiro" <shap@eros-os.org>
>
> > I've tripped on a problem in the merge logic. Here is the question.
> >
> > If I say "merge that thing and this thing", the logic finds the nearest
> > common ancestor and does the merge. This follows Josh's logic -- I claim
no
> > credit.
> >
> > The question is whether we ever need to be able to get the branch and
> > version of the common ancestor. Is it important to be able to say what
the
> > common point was?
>
> Sure, but that should be easily derivable. Each object should know its
> immediate ancestor, and we should be able to simply follow those back
until
> we hit a match.
>
> > If so, a back pointer needs to exist in the entity data structure. I'm
> > trying to figure out if this is necessary.
>
> I'd say it's necessary. What's *not* necessary, is for an object to know
> anything at all about its descendents.
>
> -jcr
>
> "The right to be heard does not include the right to be taken seriously."
> - Hubert Humphrey
>