[sc34wg3] TMCL and MAX_INT

[Lars Heuer]

[...]
>> Yes, "-INF" and "INF" are valid xsd:float / xsd:double values:
>> <http://www.w3.org/TR/xmlschema-2/#float>
>> <http://www.w3.org/TR/xmlschema-2/#double>

> One drawback is that you'd have to switch from xsd:int to xsd:double
> (or xsd:float) and you have to check that the user does not use
> senseless values like

>   min: 1.23
>   max: 2.49

Another idea: We keep everything as it is (CTM supports just
xsd:decimal and xsd:integer natively) and introduce -INF and INF as
natively supported xsd:double / xsd:float values (no other values like
2.0E2 are supported natively). If the CTM parser sees INF as value of
an occurrence it translates it to "INF"^^xsd:double.

The TMCL spec stays at it is (only accepting xsd:integer) with the
extension that "INF"^^xsd:double and "-INF"^^xsd:double is allowed as
well. So, the TMCL processor has only to check if the value represents
infinity or not. It does not have to round double values or check if
the fractional part of a double is not zero etc.

Would that work or is it too strange?

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>