From: Austin Ziegler <austin@...>
Date: 2002-12-16T22:19:16+09:00
Subject: Re: [OT] RE: help -- persuade my boss to adopt ruby

On Mon, 16 Dec 2002 21:58:26 +0900, Russ Freeman wrote:
> 
>> The problem with a hierarchical data format is that it pretty
>> much requires that your data be able to be represented as a
>> hierarchy, and that enforces a single data query path (a la
>> LDAP); you can't query efficiently in the opposite direction, or
>> from the middle.
> I don't see what the problem is as long as you have ways to refer
> to the same tag. That way you can have multiple, overalapping
> heirachies.

This isn't efficient with XML, unfortunately, because the data is
stored with the hierarchy. The optimal way to do a hierarchy is to
use an efficient random-access data format (often a relational
database) with a well-designed data model and describe the hierarchy
in terms of links to that random-access data format.

> The thing that gets my goat about XML is that it is just a way to
> persist structured information, whether it be data or meta-data.
> Big deal.

Well, yes.

> It seems a whole mass of the development community have just woken
> up to the beauty of object models and meta-models. DOH!? It's just
> a shame that they haven't chosen a better way to communicate them.
> Give me UML-on-a-whiteboard any day. Anyway YAML + Ruby is the
> future!

Unfortunately, before anyone can do any good as an object modeler,
they need to know something about data models. 99% of developers out
there don't know the first thing about data modeling, and thus will
NEVER be good object modelers. They will always think of data in
terms of their object hierarchies, and data models and object models
are NOT the same, nor should they be -- an object model is an
*instance* of a data model for a particular problem. A good data
model, represents the data being modeled -- not the data being
modeled for a particular problem.

-austin
-- Austin Ziegler, austin@halostatue.ca on 2002.12.16 at 08.13.40