Re: A Normalization Question

> At one point, I thought I could work around this difficulty by 'normalizing'
> the data -- for instance, by creating just one table for 'things.' But in
> trying to do this, I discovered that I still could not 'normalize' the data
> sufficient to eliminate duplicate entries in tables -- to which the only
> solution seems to be dramatically increasing the number of tables.

The relational model is fundamentally limited. Its limitations have pluses and minus. On the plus side, it can represent/manipulate fairly structured data (ie tables) efficiently. On the minus side, it is either impractical or impossible to represent/maniplate/normalize some scopes of data.

> Is there any database scheme that is set up in this manner, and that could be
> used to eliminate the necessity of entering the same 'entity' in multiple
> tables?

XDb's data model (TDM) is more general than RM. On the plus side, it can normalize any data. On the minus side, it is less efficient under a limited scope of data (ie tables).

XDb1, an experimental db, is a limited implementation of TDM and can normalize to a greater extent than RM over certains scopes of data.

XDb2, under development, is a more complete implementation of TDM and is able to normalize nearly all scopes of data. Received on Mon Jun 28 2004 - 20:05:27 CEST