Hi Folks,
Here is a summary of last week's discussion:
http://www.xfront.com/xpath20/best-practice/graceful-degradation/index.
html
Last week David Carlisle raised an excellent point: "it helps to keep
[data validity] separate from your program logic."
In the summary I wrestled with this:
(a) Keeping data validation and program logic separate, versus
(b) XPath is shipped around and executed in different, unanticipated
environments where it may not be safe to assume that its environment
provides schema-validation, and so program logic must address
validation.
I am interested to see what you think: has the summary struck the right
balance between validation and program logic?
/Roger
