DB modelling, is it finally going to happen?

It was (and is) so easy. Every single soul at Microsoft considered it best practise to directly connect to a database, bring table and columns names right up to the GUI layer, have a framework that mimics your database in memory (and every dev always try to load the whole database into memory).

Strange thing was however that the Visual Studio IDE always lacked any descent support for database modelling. I have tried Visual Modeller (in the VB6 era) but that didn't made me happy. The usage of Visio doesn't need to discussed here I guess. You always had to fallback to other very good tools like PowerDesigner (Sybase) or ErWin. Also Enterprise Architect is good rated.

Now it looks like Microsoft is striking back. The Visual Studio team System for DateBase professionals. That looks promising. Can't wait till June, 11th for giving that tool a test-drive.

[UPDATE]: download here.

Published 06-01-2006 8:34 AM by Rene Schrieken

Comments

# re: DB modelling, is it finally going to happen?

Drawing in a tool? That's the same as drawing pictures of what you want to produce in code later. Like little square boxes for your classes or something, and then create those classes in code later. If you start doing that, you completely lost every touch with reality.

;-)

Thursday, June 01, 2006 2:44 AM by Dennis van der Stelt

# re: DB modelling, is it finally going to happen?

Were are using enterprise architect extensively in our project and indeed. You can create class models to generate sql script. It works quite ok but is really doesn't have the quality like a good ERD tool should have. OO just doesn't map to ERD's :)

Thursday, June 01, 2006 3:56 AM by Ramon Smits

# Do it right

Create a nice clean database, after all applications exist only to flip bits and modify data, the database is the supreme artifact in any project and the sole purpose or requirements gathering. Once you have your requirements (again, requirements service data) then you generate a killer layer, better than most folks could right on their own and allow for regeneration without code loss when the data model changes.

Thursday, June 01, 2006 10:14 AM by Mike Griffin

Leave a Comment

(required) 
(required) 
(optional)
(required)