From: Clifford Heath <no@...>
Date: 2007-03-28T20:55:04+09:00
Subject: Re: Object/Relational Mapping is the Vietnam of Computer Science

brabuhr@gmail.com wrote:
> Anyone have any comments / experience with "associative" databases like:
>    http://www.associativesolutions.com/relavance.php
>    http://www.lazysoft.com/technology_sentences.htm

Ooooh, Omnicompetent programming, that must be good... Application feature
partitioning... sort of like schema access rights... What a lot of mumbo-
jumbo.

Data is as useful as it is trustworthy and unambiguous. If you build a DBMS
that allows you to stash facts without having to think about what they mean
and under what conditions they mean it, and to avoid constraining them so
only such facts can be stored, it becomes a nice way to build a personal
filofax, or any other system where a human must do all the interpretation,
but it's a completely useless way to build a software application. If a
developer likes such a system and finds it works well, it's because s/he
believes that all constraints on the data structure should be hidden inside
the application code. I think you can make up your own mind about whether
that is smart....

That said, these tools are actually seeking after a holy grail... how to
allow us to store facts in the same structure in which we think of them,
but retain the efficiency and transaction reliability of traditional tools.

A direct approach to this problem is already well-understood in the form
of fact-based modeling. Austin would like it, as it's a hyper-normalised
relational form, yet it can be invisibly mapped to an efficient relational
and transactional storage structure, leaving the queries against the user
friendly fact-based schema. It truly is time that this technology leaves
the academic domain where three decades of work have yielded nothing but a
few CASE tools, and finds its home in the heart of a dynamic development
environment like Ruby... It's because I believe this that I've left my job
of 17 years, in a company I co-founded, to build the ActiveFacts project
for Rubyists everywhere. My last day is Friday :-). Really. I almost can't
believe it myself :-). I've allocated myself 6 weeks full-time to build a
credible first release.

If anyone wants more information on fact-based modelling, take a look at
www.orm.net, http://www.casetalk.com, http://sourceforge.com/projects/orm,
www.objectrolemodeling.com, or some of the other sites they point to. Or
join the new Yahoo information_modeling mailing list. Please Austin, find
the time to look at it. Terry Halpin's book "Information Modeling and
Relational Databases" will be proved to be one of this century's most
influential works.

Clifford Heath.