Re: How to model searchable properties of an entity

From: mAsterdam <mAsterdam_at_vrijdag.org>
Date: Fri, 20 Aug 2004 13:34:17 +0200
Message-ID: <4125e1bc$0$78749$e4fe514c@news.xs4all.nl>

Laconic2 wrote:

> mAsterdam wrote:
>

>>... and hide the base tables.

>
> I think you are onto something here. The idea needs work, but it's
> potentially so powerful that it could result in a useful synthesis between
> object centered thinking and data centered thinking. I encourage you to
> develop this further.

:-)

Why do I still get warm and fuzzy feelings when I read such a comment? I'll take it as an excuse to go slightly off-topic.

>>There will be very weird searches possible.

>
> Yes. That's always true. But we should not focus
> all our design effort on the most unexpected searches.

Generic efforts tend to set up that trap.

Maybe that explains this mistery:

How come a theoretically problematic topic gets solutions over and over again in practise? Or, how can a problem that *has* to be, and *is* solved in practise every time can be so hard in theory?

To significantly change any system both the database and all applications need to undergo carefully synchronised major changes. We do that all the time in practice. Now how to do that in theory? Before you know it you are stuck in 'the impedance mismatch' and other discussions similar to 'which editor is best'.

> If the truly wierd (or perhaps "truly novel")
> uses of the data take more time and effort
> than the expected uses of the data, that's probably
> appropriate.

Exactly. That is the strategem we *can* apply in practise, but not yet(AFAIK) in theory. Received on Fri Aug 20 2004 - 06:34:17 CDT