Re: [xml-dev] How to develop a common schema for data exchange

["Stephen Green" <stephengreenubl@gmail.com>]

Hi Fraser

If anyone puts you off attempting they they might be doing you
a favour. It takes a tremendous amount of time and effort to not
just get the design right but to agree with all stakeholders on
either the semantic content or just to agree which effort is the
'authoritative' one to be doing the work. Probably best to avoid
a lot of the hard and technically controversial work by using an
existing set of what some groups call 'naming and design rules'.
One example is the one I got involved with - Universal Business
Language (UBL) Naming and Design Rules - which is royalty
free and open under OASIS but has special requirements to
follow the UN/CEFACT umbrella standard for interoperability
and harmonisation of business schemas called CCTS and that
puts a bit of an overhead on it to say the least. Still if you want
to give it a perusal you can find public draft #1 'NDR' (naming and
design rules) for UBL at
http://docs.oasis-open.org/ubl/prd-UBL-NDR-2.0.htm
and another draft is in progress.

Generally the UBL project put a lot of thought and work into how
your needs might be met by a recognised group (or even a private
team though that wasn't exactly thought to be something to
encourage) and came up with a methodology (and ongoing set
of methodologies) which a fair bit of looking up on Google under
Universal Business Language or UBL should bring to light - all
royalty free and designed for industrial strength solutions in the
main but not particulalry 'highend' ones.

But, as I say, it's a lot of work to get something usable and agreed.
Not to be undertaken lightly unless you've a lot of spare time and
no other hobbies :-)

All the best

Stephen Green


On 21/04/07, Fraser Hore <fraserhore@hotmail.com> wrote:
>
>
>
>
> I'm working with humanitarian organizations that do a lot of data collection
> in the field.  There is a tremendous amount of redundancy of data collection
> (surveys etc.) across different organizations and units because they don't
> have common data standards.  I would like to help them create a shared
> schema for common data such as nutrition and health data.  I would be
> grateful if anyone has any advice or can direct me to resources that
> describe how to go about this.  Intuitively I imagine that there is a
> process whereby a user group is formed to develop an agreed set of schema
> definitions for the data.  Some advice on how to design these schema
> definitions would be great.
>
>
>
> Thanks,
>
>
>
> Fraser
>
>