Star vs Snowflake: Looking for good reading material

Hi all. I'm currently studying "The data warehouse lifecycle toolkit" by Kimball et al.
They are against snowflakes in general, and they have a number of good reasons.
However, with an OO and 3NF background, I'm having trouble accepting this :-)
I'm also still in a learning stage and may not have all the concepts down.

I'd like some more reading material, perhaps with some examples. Any suggestions
and links appreciated.

An example I can think of is the location portion of an address. We have city, county,
state, zip and so forth. All these are related and constrained. What is my grain? The
smallest chunk, e.g., zip? Can I allow additional tables to support this dimension
table?

What if I make my dimension table a view based on the "root" tables? I know I might be trashing performance due to loss of bitmap indexing, but is there anything wrong with this in principle?

Kimball et al. state that one of the big problems with snowflakes is complicating the user interface. I'm actually working via an OO-RDBMS mapping tool, so I can handle this. I think. Once again, any experience appreciated.

Thanks in advance. Received on Sun Nov 12 2000 - 20:46:13 CET