Slashdot Mirror

← Back to Stories (view on slashdot.org)

Evolutionary Database Design

Posted by Hemos on from the jurassic-database-park dept.

Andre Mermegas writes "Check out this article by everybody's favorite object mentor Martin Fowler on database design. Be sure to take a peek at his wonderful books as well."

2 of 171 comments (clear)

  1. woohoo by jon787 · · Score: 0, Redundant

    I actually get to read an article BEFORE it is knocked off the net by the slashdot effect!

    --
    X(7): A program for managing terminal windows. See also screen(1).
  2. Re:relational databases, woo hoo by Tablizer · · Score: 0, Redundant

    Ok, I want to store 1 million documents with metadata and I want to extract sections based on heading levels (ie, H1, H2...). Consider that it's all simple HTML markup.

    What do you mean "based on heading levels"? The query gives heading levels or the results? What is an example (English) query?

    Hierarchy in a relational database is done using pointers and parentID and, well, you get the idea. Assembling a simple tree is a hard thing for a database

    I will agree that if *all* your queries are tree-based, then (existing) RDBMS will probably be slower. However, if queries come in all kinds of shapes and sizes, trees being just one, then things are not so clear cut. RDBMS are decathlon athletes, not necessarily event athletes. I often wish I could do non-tree queries on my files. Trees are easy for users to grok (up front), but often don't reflect the true shape of the world when they grow non-trivial in size, in my observation.

    --
    Table-ized A.I.