Re: more on delete from join

From: paul c <toledobythesea_at_oohay.ac>
Date: Thu, 27 Aug 2009 01:26:24 GMT
Message-ID: <4hllm.40846$Db2.36998_at_edtnps83>


Kevin Kirkpatrick wrote:
...
> In short, before delving into, "*how* should the DBMS handle view
> updates?", I'd like to see a discussion about the question, "*should*
> the DBMS handle view updates?".

Wow, that's far out. To want a dbms never to update views might be something an ostrich would want but not I. Better to want a logic that permits view updates. Lots of pragmatic reasons, security, archiving, optimization and data re-arrangement are just a few off the top of my head. Without view updating db design gets very hard in those two areas at least, not to mention implementation. Also, without logical independence, I suspect any progress in a much-ignored area, optimization, will be much harder, especially constraint optimization. .Personally, I think the only sensible way to provide definitions to users is to always provide them with views, never with base definitions, much easier to correct some mistakes that way,. Denying updates closes all kinds of doors. Received on Thu Aug 27 2009 - 03:26:24 CEST

Original text of this message