Lists Home |
Date Index |
Michael Champion wrote:
> So, XML succeeds for the familiar 80:20 reason -- it handles (well
> enough) a lot of familiar situations such as document-driven business
> processes with few impedance mismatches between the conceptual/end-user
> level to the physical/implementation-level technologies without
> demanding the scarce skills of effectively normalizing conceptual models
> into relations,...
I see it this way: We now have 80 % of data living in the messy horror
world of proprietary file formats and ad hoc structures inside Excel
sheets and the like. If those 80 % are taken over by XML, that's a big
step forward. But if XML takes over 80 % of the remaining 20 % of data
that today lives in relatively professionally designed relational
databases, that'd be a big step backwards. And the reason is not so much
a technical one. Of course nobody keeps XML people from thinking about
the consistency of their XML 'data models' and use hierarchical nesting
only where there's conceptually a containment relationship and refrain
from writing order dependent XSLT where there is no inherent ordering.
But if the motivation to use XML in the first place is to avoid to pay
for the scarce skills of people who are able to create a consistent data
model, then I'm pessimistic about the effects of the XML revolution.
It's not relational technology that demands normalisation. It's the
consistency requirements of some application domains.
What I see as the big opportunity of XML as a technology is that it is
possible to start with semi structured, denormalised models and
continually evolve them into a more strictly defined, more normalised
form (or even the other way around sometimes) as need arises and as
resources permit. Because that's what the relational model is really bad
at and that's the reason why we have this huge gap, technologically and
culturally, between the messy horror world of file formats and the
orderly (and indeed more manageable) world of relational databases.
What we should be concerned about is inhowfar the various XML related
specifications support this kind of continually evolving datamodels.