From: mcstockX@Xenquery   
      
   "Hans Forbrich"  wrote in message   
   news:zfGdc.28850$J56.25383@edtnps89...   
   | /motten wrote:   
   |   
   | > English isn't my primary language so I need to make sure i understand   
   what   
   | > you're saying.   
   | > Are you saying my approach is reasonable or are you saying that both   
   mine   
   | > and my collgegaues   
   | > designs are defect ?   
   |   
   | My apologies for not being clear.   
   |   
   | I reserve the right to change my mind if given new or more detailed   
   | information.  However, based on what you have written:   
   |   
   | - I disagree with your colleague (for the reasons I gave); and   
   | - I agree with your approach.   
   |   
   | Defective design (to me) means it will not work.  I do not believe either   
   | design is defective.   
   |   
   | My comments about your colleagues design: Based on my experience, I   
   believe   
   | there will be an unacceptably high cost of maintenance and administration.   
   | And I suspect the performance will not be consistent, but that is   
   something   
   | that would need to be tested.   
   |   
   | /Hans   
      
   i have created systems that were designed to be extensible as new   
   functionality was added that could not be handled by the original data   
   model -- but the idea was to analyze the data well enough up front so that   
   the original structures were generic enough to handle new types and subtypes   
   without needing new data structures. and, any new data structures would be   
   created by developers, not by end users, and not dynamically. i share hans'   
   concerns with proliferation of tables -- likely many would be nearly   
   identical in structure -- and would certainly emphasis the major   
   complications you will encounter with reports, as well as interfaces in   
   general   
      
   in your survey scenario, it certainly should be quite possible to model   
   structures to handle almost any survey -- as long as sufficient analysis is   
   done. i would want to make sure that the preference for creating tables on   
   the fly is not due, at least in part, to insufficient analysis and design   
      
   get and study david haye's 'Data Model Patterns: Conventions of Thought'   
   (isbn 0932633293) for good insights into well-thought out relational data   
   structures   
      
   -{ mcs   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)