Re: Modelling Considered Harmful
From: mAsterdam <mAsterdam_at_vrijdag.org>
Date: Sun, 08 May 2005 11:44:10 +0200
Message-ID: <427ddf69$0$164$e4fe514c_at_news.xs4all.nl>
>
> Hmm maybe it comes down to semantics then, because the way I look at it,
> the metadata describes a top-level model of reality, and the data just
> provides a more detailed model of reality.
>
> For example, consider a group of employees I want to model. For
> simplicity assume I'm using a single table.
>
> The metadata that descibes that table gives me a very broad model of
> reality: it tells me that I'm dealing with a group of people who have
> certain attributes.
>
> But it doesn't tell me who those people are or what values the
> attributes have for each person - for that I have to supply the row
> data. But to me that's still a model; it's just more detailed, giving me
> the names, departments etc. for each person. Each row is a model of a
> person, if you like.
>
> Maybe I'm belabouring an unimportant point though and it just comes down
> to the definition of the word "model" in different contexts.
>
> I agree that what we commonly refer to as "data modelling" is all about
> metadata, but I still think data itself is also about modelling.
Date: Sun, 08 May 2005 11:44:10 +0200
Message-ID: <427ddf69$0$164$e4fe514c_at_news.xs4all.nl>
Paul wrote:
> Bernard Peek wrote:
>
>>>I don't understand this point at all. To me it seems almost a tautology >>>that databases are models of reality (unless of course they're modelling >>>some virtual world that doesn't exist outside the database). >> >>The metadata in the database is a model of reality, and datamodelling >>concerns itself with metadata.
>
> Hmm maybe it comes down to semantics then, because the way I look at it,
> the metadata describes a top-level model of reality, and the data just
> provides a more detailed model of reality.
>
> For example, consider a group of employees I want to model. For
> simplicity assume I'm using a single table.
>
> The metadata that descibes that table gives me a very broad model of
> reality: it tells me that I'm dealing with a group of people who have
> certain attributes.
>
> But it doesn't tell me who those people are or what values the
> attributes have for each person - for that I have to supply the row
> data. But to me that's still a model; it's just more detailed, giving me
> the names, departments etc. for each person. Each row is a model of a
> person, if you like.
>
> Maybe I'm belabouring an unimportant point though and it just comes down
> to the definition of the word "model" in different contexts.
>
> I agree that what we commonly refer to as "data modelling" is all about
> metadata, but I still think data itself is also about modelling.
To the user of the data, yes, a model (if the database is any good). To the DBMS it's meaningless signs to be cleverly kept - no model. Received on Sun May 08 2005 - 11:44:10 CEST