Re: Logical Attribute or not

From: David Portas <REMOVE_BEFORE_REPLYING_dportas_at_acm.org>
Date: 18 Mar 2006 04:30:29 -0800
Message-ID: <1142685029.625456.186170_at_i40g2000cwc.googlegroups.com>


sagaeta_at_yahoo.com wrote:
> So, let's say Fred is assessed. The assessor would need to code
> something for each disorder at the assessment. The physical
> implementation would include 4 rows of data as follows:
> Value Code
> ==== ====
> a 0
> b 1
> c 2
> d 0
>
> I am getting some push back due to the fact each code value is
> mandatory. The premise being that since each is mandatory that each
> should be a logical attribute. I am thinking that would violate 3NF.

Why would it violate 3NF? It isn't clear what the key is so I can't comment. Presumably the key identifies the subject being assessed in which case the key is a determinant for each of those attributes.

> And, what if 10 mood disorders were added later?

Add more attributes?

> Further, if this data was to be exchanged using XML, would each of a-d
> be a specific XML tag or would it be typified off of the parent XML
> root?

I don't think there is any generally accepted design theory for XML. Consider using some metadata standard. I'm sure that standard XML schemas exist for patient data.

-- 
David Portas
Received on Sat Mar 18 2006 - 13:30:29 CET

Original text of this message