Re: S.O.D.A. database Query API - call for comments

From: Carl Rosenberger <carl_at_db4o.com>
Date: Fri, 18 May 2001 19:02:12 +0200
Message-ID: <9e3kjg$qtj$06$1_at_news.t-online.com>

Daniel Dittmar wrote:
> > The discussion has become kind of dull.
>
> How about firing it up again by mentioning XML databases.

Hi Daniel,

would this be a theme that fits the original subject of the thread? I was hoping on constructive statements to our first design. http://www.odbms.org/soda/soda.zip

Instead it developed into a discussion that questioned the usefulness of object databases.
...very awkward...

> - language independent
> - complex queries (XPATH), both navigational and value oriented
> - normalization is up to the database
> - assembling of data into meaningful larger groups (master-detail) is done
> by the server

Be careful what you post here. Lee Fesperman tries to point out, that all the points you mention are negative.

<caricature>
- "language independant" ??? Language independance is not possible. You are using XML, aren't you?
- "both navigational and value oriented" are very bad because two different methods to access data add complexity.
- "normalization is up to the database" does not work because it does not allow reuse.
- "assembling of data into meaningful larger groups" ??? Now what is that? It this proved by relational theory? You are bound to get redundant and misplaced data. You are mixing metadata and data. </caricature> :-)

Personally I dislike XML since:

- it needs to be parsed.
- repetitive tags produce unnecessary overhead.
- it is very unefficient in representing complex linked structures.

The widespread use of XML only proves the inability of the IT industry to work together on a clean data representation protocol.

XML is an O.K. format for exchanging very simple documents. It's use as a database protocol is a waste of performance and efficieny.

Kind regards,
Carl

---
Carl Rosenberger
db4o - database for objects - http://www.db4o.com

Received on Fri May 18 2001 - 19:02:12 CEST