- From: Michael Kay <mhk@mhk.me.uk>
- Date: Fri, 14 Jan 2005 12:49:42 -0000
- To: "'Ashok Malhotra'" <ashok.malhotra@oracle.com>, "'Michael Rys'" <mrys@microsoft.com>, "'Colin Paul Adams'" <colin@colina.demon.co.uk>, <public-qt-comments@w3.org>
> The IEEE spec spells out the lexical forms as well.
>
You can't write
a := NaN
in Java or Fortran or any other language that supports IEEE floating point.
It would only be possible if NaN were a reserved word (or a predefined
constant) in the language. The only requirement is that you can cast from a
string having that form:
a := double("NaN")
and of course we allow that.
Michael Kay
Received on Friday, 14 January 2005 12:50:18 UTC