Re: Modeling question...

From: paul c <>
Date: Fri, 24 Oct 2008 13:40:09 GMT
Message-ID: <ZekMk.3608$%%2.3567_at_edtnps82>

Walter Mitty wrote:

> "paul c" <> wrote in message 
> news:wLJLk.3213$fF3.1733_at_edtnps83...

>> Roy Hann wrote:
>> ...
>>> And before I leave this alone, there is no such thing a
>>> "semi-structured" data. That term makes as much sense as
>>> semi-understood knowledge. The concept that people using that term
>>> might be struggling to convey is "semi-shared business model", or to put
>>> it another way, "(only) some of us know what (only) some of this
>>> means". My attitude to that is fine, just don't expect me to know
>>> what any of it means.
>> Heh, in other words, semi-understood data?
>> Ironic how "not-invented-here" so often actually means "invented here".
>> (Letting "semi-understood data" proliferate might be chaotic. Maybe in
>> such a regime, to echo Walter M, it would be prudent to ensure that it be
>> kept "semi-shared". Eg., amongst the EAV protagonists and their cronies.
>> It usually seems to me that when these EAV proposals come up, the question
>> is not that the organization needs new organization-wide "entities", for
>> want of a better word, but additional attributes for existing relations.
>> So, I'd think it might be okay from an integrity viewpoint to let them
>> define their own tables which are partly based on organization-wide
>> tables. At least everybody could stick with the usual relational ops.
>> Not sure if I've ever seen this tried, though.)
> Perhaps it would be sufficient to categorize the outputs of the database as 
> "semi-correct".
> Let the users think about that one for a while!

Maybe the key problem to be avoided is inadvertent semi-redundancy which could pollute the db as a whole. If there's a way to avoid that I'd think there is no basic problem even though various pet niceties such as performance targets might suffer. Received on Fri Oct 24 2008 - 15:40:09 CEST

Original text of this message