Re: Too add a new table or not

From: Carlos Bromanski <cbroman_at_shpamcore.com>
Date: Fri, 9 Nov 2001 01:28:47 -0600
Message-ID: <3beb905c$0$35622$1dc6e903_at_news.corecomm.net>


what you're saying is:
cheap, fast, correct: choose any two

"Bernard Peek" <bap_at_shrdlu.com> wrote in message news:FHZwKzM1eF67EwUb_at_shrdlu.com...
> In message <NCVF7.14844$xS6.20479_at_www.newsranger.com>, Lennart Jonsson
> <lennart_at_kommunicera.umea.se> writes
>
> >I was more thinking about an initial design - and to make things
simpler - only
> >the logical model. Lets say I have identified the structures mentioned
earlier,
> >should I aim for 1 or more entities? I'm participating in a project where
we
> >have similar structures, thus interested in what is considered the
apropiate
> >thing to do.
>
> If your modelling system supports subtypes and you know they exist then
> it would be a mistake to leave them out of your logical structure.
>
> But entities and tables aren't the same thing. Once you start
> considering how to implement the database as tables you have left the
> entities of the logical structure behind. The same logical structure can
> be implemented in different ways and usually every choice is both right
> and wrong in different ways. You have to make some compromises and each
> implementation choice makes a different compromise.
>
> You want to optimise development costs, development time,
> maintainability, storage space and processing time. You can't have them
> all. Do you want it better, faster or cheaper? Make your choice. If you
> choose cheaper do you want it cheaper to develop, cheaper to maintain or
> cheaper to run? Make your choice.
>
> Once you have prioritised all of the constraints on the development
> process you need to consider how your implementation choices affect your
> costs. Should you front-load the costs, burning up developer time to
> make the long-term running costs lower, or should you throw something
> together in ten minutes and live with higher maintenance costs later? In
> practise the compromise is somewhere between the two. Exactly where
> depends on your priorities and only you know what those are.
>
>
>
>
>
> --
> Bernard Peek
> bap_at_shrdlu.com
>
> In search of cognoscenti
>
Received on Fri Nov 09 2001 - 08:28:47 CET

Original text of this message