Re: transitive closure

John wrote:
> Marshall Spight wrote:
> >
> > The transitive closure issue seems like a bigger issue than
> > the missing values issue. What are the possible solutions?
> > What's good and bad about them?
> >
> I like writing procedural code for the computation of transitive closure
> and for similar tasks like tree traversal. I would much rather store
> clean and simple information about tree structure (ie a transitive
> reduction) than mess around pre-computing the information and worrying
> about keeping it correct and current.

Okay, I suppose that's one approach we could use today, but I was more asking about an ideal solution. That is to say, a solution in theory, rather than a workaround for today.

One weakness of your proposal is that your procedural code will have a query (involving a network roundtrip to the db) inside a loop.

> "Caching is for the middle tier!".

Caching, phooie! It has been the bane of my existence. Caching means you get scores of application programmers thinking they don't have to pay any attention to their queries. So they write horrific queries and push them to production.

Marshall Received on Mon Jul 18 2005 - 16:49:53 CEST