Lists Home |
Date Index |
I agree. The blobs, the texts, the otherwise NOTATIONs,
are the challenge to search systems that mine. By
releasing that code, IBM advances the practice.
Markup, as the name alludes to, is not about normalization
although from the computer's point of view, that is a
good thing to achieve. Should one markup only to
the advantage of the computer? It's an easy trap to
From one point of view, it is about capturing the
structure and meaningful names organically present.
In our younger days when some of us talked about using
markup for modeling, some thought it a means to create
a model, others a means to capture one. Practically,
XML doesn't care, but historically, I think the latter
approach dominated until XML made markup a favorite
among computer scientists.
But..... if one tries to 'fix' an organically derived model,
one may lose some of the meaning. So even as we
proseletyze good practices for structuring, we come
back to naming as the practice to cultivate because
idioms are meaningful.
From: Peter Hunsberger [mailto:email@example.com]
> So while a relational database may be a rigorous
> example for structure, I don't think that is what
> the article cited is really about. If one is looking
> for patterns, relationships, intentions, meaning, is
> it easier to get that from a relational database or
> from an XML instance? It depends, in my opinion,
> on how predigested the content is, not the structure
> although the structure is useful for finding an
> answer where one already knows the question and
> the structure. Questions are harder.
It seems to me that since both relational DBs and XML allow "escape to
BLOB"; either of them can have as much or little structure as one
cares to enforce. IMO the largest difference is the larger body of
best practices for building up good structure in relational DBs. Eg.
normalization theory for XML is still relatively immature by