RE: [xml-dev] Lesson Learned: Use namespaces for both markup anddata

["Glidden, Douglass A" <Douglass.A.Glidden@boeing.com>]

Okay, perhaps I should have worded that a little differently:  Using QNames in data dangerously de-atomizes the data.  I used the word "metadata" before because a namespace prefix by itself can never really be considered data; outside of an XML parser it is completely useless, and within the parser it functions only as a pointer to a namespace identifier.

Doug Glidden
Software Engineer
The Boeing Company
Douglass.A.Glidden@boeing.com

---

On Wed, Aug 12, 2009 at 4:20 AM, Michael Kay <mike@saxonica.com> wrote:

>
> I definitely don't agree with Roger's idea of using QNames in
> data, either, though-IMHO that dangerously mixes metadata with data.
>

But the notion that metadata and data can be separated is itself deeply
flawed. If I have two lines of business, "product sales" and "services", are
these two things data or metadata? You'll find some systems in which they
are data (rows), others in which they are metadata (columns). Similarly, if
I have three kinds of personnel, "salaried", "hourly-paid", and
"contractors". Or if I have 23 kinds of asset on the books, or 592 kinds of
equipment, I will sometimes want to treat these as types and sometimes as
instances. Mixing metadata and data is something we have to live with.
Functional data models rightly eliminate the distinction.

Regards,

Michael Kay
http://www.saxonica.com/
http://twitter.com/michaelhkay