Re: Data Modeling?
Posted in 2006
Topics: Server Administration, Triggers, Constraints & Referential Integrity
Urich Ann wrote: > Is it a 'norm' that DBA's use a data model (both logical and physical) > to design a database? > or > Are data models mostly used as a 'post-design' afterthought, to document > what is physically already implemented? > > In most shops that I have worked, we never used a model. And if they did > have a data model, it was never up-to-date. > The model was an after-thought for documentation purposes... It was done > after the table was implemented in Production. > > Is there a good modeling tool for Informix 10? > Is there any one 'data modeling' design class that is better then > others? > > I think you should look at Data Modeling from two perspectives. One view is for new work, the second for existing systems. On new work, I think the best Data Modeling tool is still pencil and paper, or a white board. This is a fun time, where you can actually design the data without any restrictions other than what the business people throw into it to screw it up. On existing systems, probably best to reverse engineer the schema with some kind of software tool, but even the best I've never used to go and make changes to the database with it. PowerDesigner does a good job, even Visio is adequate for that. But I've yet to see some kind of software product actually go beyond just linking primary and foreign keys. You can write your own column cross-reference tool to see column relationships, which is basically what visual tools are going to do, unless you can link some kind of documentation in the metadata describing what each table is and how it relates to the other tables. Good luck if you find something that doesn't cost a fortune, and actually does more than just draw lines connecting primary-keys and foreign-keys together. The evaluation of current products will be interesting, you should produce a white paper of comparison products, it might be quite beneficial for the rest of us.
Thanks to ALL, for the feedback and suggestions. If I am able to test many/most of the data modeling tools, I will write a white paper on the results. And, I'll add to the white paper, whether a 'formal' data modeling class teaches me anything new... ... or if it is about what I ALREADY know and am using -- but is 'labeling' it differently. :) Thanks. Ann -----Original Message----- From: informix-list-bounces@iiug.org [mailto:informix-list-bounces@iiug.org] On Behalf Of Double Echo Sent: Friday, December 08, 2006 6:36 AM To: informix-list@iiug.org Subject: Re: Data Modeling? I think you should look at Data Modeling from two perspectives. One view is for new work, the second for existing systems. On new work, I think the best Data Modeling tool is still pencil and paper, or a white board. This is a fun time, where you can actually design the data without any restrictions other than what the business people throw into it to screw it up. On existing systems, probably best to reverse engineer the schema with some kind of software tool, but even the best I've never used to go and make changes to the database with it. PowerDesigner does a good job, even Visio is adequate for that. But I've yet to see some kind of software product actually go beyond just linking primary and foreign keys. You can write your own column cross-reference tool to see column relationships, which is basically what visual tools are going to do, unless you can link some kind of documentation in the metadata describing what each table is and how it relates to the other tables. Good luck if you find something that doesn't cost a fortune, and actually does more than just draw lines connecting primary-keys and foreign-keys together. The evaluation of current products will be interesting, you should produce a white paper of comparison products, it might be quite beneficial for the rest of us. _______________________________________________ Informix-list mailing list Informix-list@iiug.org http://www.iiug.org/mailman/listinfo/informix-list