Re: One Ring to Bind Them

"Marshall Spight" <mspight_at_dnai.com> wrote in message news:aFAEc.3451$IQ4.1727_at_attbi_s02...
> "Dawn M. Wolthuis" <dwolt_at_tincat-group.comREMOVE> wrote in message
news:cbt4fk$90q$1_at_news.netins.net...
> > "Marshall Spight" <mspight_at_dnai.com> wrote in message
> > news:nnIDc.125667$Sw.113988_at_attbi_s51...
> > > > > Perhaps I misunderstand, but MV has only the one kind of
> > > > > relationship it is capable of understanding: containment.
> >
> > I should have read from the top of the topic down, but I now understand
what
> > you mean. As far as the database itself, without any triggers written,
nor
> > any application code, the only relationship "between relations" that it
> > understands is that of parent-child. --dawn
>
> So what do you do in the face of many:many relationships? I bet
> it's the same thing that OO does: you have links on one side and
> links on the other, and manage them in code.

yup

>
> Many to many relationships are one thing that the RM just totally
> nails. I bring this up not to run a whole "mine's bigger" thing but
> because I believe that if this entire years-long conversation has
> a use, it is to highlight the areas where each side succeeds, so
> that we may begin to work towards a new model that encompases
> the best of several existing systems.

Sounds good.

RM does do well with M:M, the most conceptually difficult for the user, but not in doing anything to simplify the presentation/ease of use for the user. Viewing books and their authors from one perspective and then authors and their books from another makes sense to a person. Viewing it as a many-to-many is not as helpful (as each book-author pair on a separate line so you don't have one row for each book, nor one row for each author). RM also has difficulty with multiple 1:M with the same 1 when there is a need for counting, summation, or other arithmetic and visuals/reporting against the same. STAR joins have helped a bit with that, I think.

> In programming languages, they are talkin more and more about
> "multiparadigm." I think we should follow their lead.

agreed. --dawn Received on Wed Jun 30 2004 - 18:50:20 CEST