Re: Any new thoughts on OTLT (One True Lookup Table)
From: Jerry Gitomer <jgitomer_at_erols.com>
Date: Sat, 27 Mar 2004 23:30:54 -0500
Message-Id: <pan.2004.03.28.04.30.53.304042_at_erols.com>
>
> Yes, it does. And that's precisely the problem with putting the code "PA"
> in the same column twice, with two different meanings. Making use of the
> "class" column,
> requires a programmatic solution.
>
> We've taken a step away from self describing data.
>
> How does the new user of the data, whether person or automaton, "know"
> that you have to qualify the "code" column with the "class" column? Is that
> inherent in the two columns? Is it written in the metadata? No!
>
>
> Is it inherent in the underlying subject matter? No!
>
> The new user "just has to know" that that was the intent of the designer,
> and program accordingly. And that's what's wrong.
Date: Sat, 27 Mar 2004 23:30:54 -0500
Message-Id: <pan.2004.03.28.04.30.53.304042_at_erols.com>
On Sat, 27 Mar 2004 08:12:44 -0500, Laconic2 wrote:
>> Doesn't OTLT have to include a third column, 'class' perhaps, with >> values 'state' and 'airline' respectively? >>
>
> Yes, it does. And that's precisely the problem with putting the code "PA"
> in the same column twice, with two different meanings. Making use of the
> "class" column,
> requires a programmatic solution.
>
> We've taken a step away from self describing data.
>
> How does the new user of the data, whether person or automaton, "know"
> that you have to qualify the "code" column with the "class" column? Is that
> inherent in the two columns? Is it written in the metadata? No!
>
>
> Is it inherent in the underlying subject matter? No!
>
> The new user "just has to know" that that was the intent of the designer,
> and program accordingly. And that's what's wrong.