XSL & Stuff

[Joel Meulenberg]

Wow!  I didn't expect such a flurry of comments to come out of the
questions I asked.  I was actually more interested in your
opinions/experience with using "trees" (strictly hierarchical structure
which allows only 1-to-many relationships between entities) as the
fundamental model for your data as opposed to using relations as the
fundamental model for your data.

I'll ask a couple more questions about that below...

> >What do you mean by an "XML database"?
> 
> Here's the deal.  Use a nifty, propietary way to store
> data in a compact format.  Create a process to turn a
> single item (product, set of products, whatever) into into
> XML using MS Tools.  This is done on-demand, so
> we don't need to store the data in XML format.
> 
> Use XSL to transform that XML into HTML.
> 
> This allows us to separate the data storage layer (proprietary)
> from the business logic (make XML) from the presentation
> layer ("Stock" HTML pages that are filled dynamically).
> 
> If you think about it, you can create several different XSL
> "templates",
> and apply a different Template to get a different view of the data.

I dig.  Using XSLT to "repurpose" the data, etc.  (Two transformations
seems a little heavy, but I'm sure it's not that big of a deal.)

I'm wondering how the "make XML" step embodies "business logic"?

I'm also wondering how one would query the proprietary data store?

> >If it's more like a data file, then I'm curious about your
> >opinions/experiences with storing the data in independent
> hierarchical
> >data structures (i.e.- the XML documents)
> 
> Too much disk space.
> 
> >in contrast to storing data in relations (tables) in a DBMS?
> 
> Too much disk space.
> And too expensive.  (Or hard to maintain, if you're an open-sourcer)
> And takes too many clock cycles.

I'm not so interested in physical storage aspects, etc.  I'm more
interested in how you like working with a hierarchical model of data as
opposed to a relational model of data.  *Both* hierarchical databases
and a relational databases (in the abstract sense - I'm not referring
to any particular SQL product (none of which are quite relational
anyway)) *can* be stored compactly and can perform well given the right
implementation.
(Indeed some of the followup comments of others suggest that has been
their experience.)

It's your feelings/experiences with the fundamental abstractions (i.e.-
tree versus relation) that I'm wondering about.

The shortcomings of always trying to shoehorn data into a strictly
hierarchical structure were recognized decades ago and that's why the
idea of hierarchical databases was essentially abandoned in the 1980's.
 (e.g.- The hacks that people had to use to handle a many-to-many
relationship in a model of data that only allowed one-to-many
relationships were horrible.)

Have you felt any "impedance mismatch" between your data and the
strictly hierarchical structure imposed by XML?

I could ramble some more, but I've been told my emails get too long and
then no one wants to read them, so I'll stop here.  : )
Thanks for the clarification so far.

+Joel


__________________________________________________
Terrorist Attacks on U.S. - How can you help?
Donate cash, emergency relief information
http://dailynews.yahoo.com/fc/US/Emergency_Information/