Defending the network model (was: Re: Relational Databases and Their Guts)

From: Carl Rosenberger <carl_at_db4o.com>
Date: Mon, 23 Jun 2003 19:37:58 +0200
Message-ID: <bd7d5m$jaj$01$1_at_news.t-online.com>


Bob Badour wrote:
> The network model raises the bar on suckiness.

What's "suckiness"?

I fully agree that the power of the Java language model appears to be inferior to other language models because it does not directly support:
- multiple inheritance

Nonetheless, the absence of features has led to:
- common coding practices

  • reduced education requirements
  • a very fast just-in-time compiler
  • excellent memory garbage collection
  • the availability of cheap and very powerful libraries, that get the job done for you

Simplicity rules.

Following Clinton's "It's the economy, dummy" slogan: It's the time-to-market, dummy!

> Application specific data models raise the suckiness bar even higher.

And this equally gos for application specific data models.

If you have one application that you need to persist, why not use it's model? It saves time and produces additional stability and allows better refactoring.

The lifecycles for some applications are very very short. Scale them to the needs!
A clean rewrite very often is a lot cheaper and less work than trying to develop a monster for 30 years in advance.

Kind regards,
Carl

--
Carl Rosenberger
db4o - database for objects - http://www.db4o.com
Received on Mon Jun 23 2003 - 19:37:58 CEST

Original text of this message