xsi:null Feedback

I was pleased to see the xsi:null feature make it into XML Schemas and I
have had no reason to change my mind.

Before having first viewed XML Schema draft specs, I had created my own
XML document types with an "isNull" attribute as a key feature for the
reasons enumerated in the draft spec.

It should be noted, however, that this concept is a general programming
concept, not just a database concept.  Java, C, C++, Pascal, etc, all
have a notion of a null or nil value.  Hence xsi:null should be boon to
mapping between data from these languages and XML.  [My "isNull"
generation happens to be driven by a Java structure traversal
encountering a 'null' Java variable, which is of course firing
appropriate SAX events.]

Also, the "isNull" attribute is necessary if nothing else to distinguish
between a string specified as empty ("") and an unspecified string.  In
some cases this can be signaled by the absence of an element, but this
is implicit rather than explicit and can have other implications (e.g.
efficiency of some generic XSLT/XPath-based processing algorithms
decreases if one cannot assume element presence/ordering).  This could
also be done with another element or attribute, but the concept is so
standard that there is no reason to do something special / ad-hoc.

In short, please keep xsi:null.  I don't know of anything it is missing
or any reason not to keep it.  I won't die if you remove it but I'd hate
to have a million similar but different ways of representing this
fundamental concept!

--
Jess Holle

Received on Thursday, 28 September 2000 17:07:05 UTC