XPath 1.0 comparison of NaN and any other value

Question

I'm learning about XML and XPath in my university course (XPath 1.0 to be exact) and I stumbled upon something. F.e. if I write an expression:

NaN < true()

it returns true. Which is very weird. When I do.

NaN < 1

it returns false. AFAIK XPath 1.0 comparisons (without node-sets) first convert the values to numbers and only then compares them. So, in the first expression it goes like this

NaN < number(true()) -> NaN < 1

and in the second

NaN < 1

I'd like to get to the bottom of this, so any information would be helpful. I didn't find anything concerning this situation in the docs. Maybe I missed something.

Thanks in advance!

Answer 1

The token NaN does not represent the double value NaN, it is short for child::NaN which returns an empty node-set (unless of course your XML document contains an element named NaN).

I'm a little surprised that a university should be teaching XPath 1.0 which has long been superseded by new versions. Of course lots of people are still using it, but I would have thought a university would teach a newer version.