At 12:46 PM 4/16/03 +0000, Phil Shrimpton wrote:

>On Saturday 05 April 2003 23:42, Helen Borrie wrote:
>
>Hi,
>
> > One of the traps I see with locking XML storage into the engine is that we
> > get stuck with something that's certain to be deprecated pretty soon, when
> > the next wave of *ML becomes the currency of the day.
>
>I agree with this, but I don't think we need to bother with the storage side
>of things, as the the current BLOB side of things can handle those needs (it
>could be improved, but that would beifit all 'blob' storage).

Blobs are a storage mechanism, and have next to nothing to do with translation
between XML and structured data. I think it very important to keep in mind
that
XML (or *XL) is a data interchange format, and not a format for data
storage or,
gasp, manipulation. (Note: for those of you born after 1990, there used to be
absolutely dreadful database built around the notion of a tree of data. IMS
was a pretty convincing case that this was a bad idea.)

A properly designed ruleset language for mapping between structured data
and an ascii stream should work equally well for XML, HTML, and an
arbitrary future *ML.

> > Let's blaze the trail, be a Genuinely Modern Database Engine,
> > and focus instead on an excellent interface for plugging in optional
> > modules.

That's an embedded JVM. But XML (and HTML, et al) mapping is of sufficient
important to be considered in it's own right, at least in my humble opinion.

Jim Starkey