Re: XML databases [ was: S.O.D.A. database Query API - call for comments ]

Philip Lijnzaad wrote:
> To be fair, this is frequently because advocates of OODBMSs, at least in
 the
> past, have tried to slight RDBMSs. This time and again proves an uphill
> battle. This is indeed not very helpful, because I too hope that both
 'camps'
> can learn from each other.

Since applications and usecases vary drastically, it is not possible to lead an objective discussion. Everyone has a different standpoint. The two technologies simply compete so a flame war is only natural.

I can switch sides very easy:
- Object databases match object-oriented languages better. They provide better performance and decrease development time. - Object databases lack a common standard, maturity and querying capabilities.

Depending on your usecase, either of the above arguments is sufficient for a decision, what to use in practice.

> Carl> - "both navigational and value oriented" are very bad because two
 different
> Carl> methods to access data add complexity.
>
> and power, methinks.

You quote me out of context. You missed out the "<caricature>" bit. If you read the thread that brought us here, you will find that I only brought up Lee's strange arguments, to use them against XML.

Of course a navigational path to data is a great practical advantage. You use the same method to navigate through your objects in your programming language. Transparent persistency ideally hides from the programmer that the database is involved to continue to reiinstantiate objects. In this case there is not even any complexity added. The programmer writes object.member.member ... in his program as he usually does.

> Carl> Personally I dislike XML since:
> Carl> - it needs to be parsed.
> Carl> - repetitive tags produce unnecessary overhead.
>
> yup. But surely we'll soon have some dedicated encoding which can cut this
> down considerably (and gzip already routinely compresses XML down to less
> than 30% of the original).

Zipping XML only makes the situation worse: - You need to unzip the entire document to retrieve content. - You loose the only little advantage that XML has: read- and editability with widespread text editors.

A more effective format would

- not be text-editor-readable from the start
- try to avoid redundancies by system
- allow queries for points of interest without parsing the entire document
- use up exactly the amount of information that the original data contains:
  - Integer, Float = 4 bytes
  - Long, Double = 8 bytes
  - Unicode would be possible

• no tags
• internal links would simply be pointers within the file
  • internal file offset

> Carl> XML is an O.K. format for exchanging very simple documents. It's use
 as a
> Carl> database protocol
>
> well ... protocol ... no. The XML hierarchical queries do seem to offer an
> interesting perspective not offered by RDBMSs. To me it's vaguely
 reminiscent
> of parts of OQL, with the added benefit that it is standardized and
> implemented.

O.K. sorry, XPATH seems to add a nice features: Queries return a tree structure.

Kind regards,
Carl

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