Re: Against the Grain: Pascal commentary about XML and databases

["Thomas B. Passin" <tpassin@home.com>]

[Alaric Snell]

>
> Relational databases won over the faster but less powerful (due to the
lack of
> arbitrary joins) hierarchical databases in Days Of Old, mainly due to data
> modelling issues IIRC.
>
> Does anyone know how/if XML databases differ from those old relational
database
> revisited, and if the arguments then still hold? Any historical recordings
of
> those debates?

Remember how hierarchical databases were generally designed around a
particular set of data structures and that these were usually tuned
carefully for your problem?  With relational databases, people learned to
concentrate on a good, general data model, which is the way to gain
flexibility; it also brings that adaptability that often makes up for
possibly slower responses from relation databases.

XML and xml databases will share many of the advantages and disadvantages as
hierarchical databases did (still do, some are still around) as long as the
xml structures are developed without good data modeling methods.  Why would
it be different?  My main concern here is that most of the books on the
shelves about xml and xml databases largely overlook proper data modeling.
So it looks like back to the past.  Let's try to move forward, not back -
work out how to build on al those years of data modeling lessons instead of
learning them all over again.

Cheers,

Tom P