Lists Home |
Date Index |
At 02:33 PM 5/13/2002 -0400, Mike Champion wrote:
>5/13/2002 2:18:32 PM, Jonathan Robie
> >If XML is used as a serialization format for languages that use numerics of
> >fixed size, it is useful to know that you have an instance that will fit in
> >one of these numeric types.
>I'm not asserting that there is no rational reason for the distinctions made
>by the designers of the XSD types, I'm simply asserting that many are
>to the needs of the vast majority of XQuery users. The distinction
>between int, long, and short, for example, made perfect sense in a world where
>8-bit, 16-bit, and 32-bit machines were all commonplace. It is now rather
>quaint for most of us, and will be a bizarre anachronism in 10 years or so
>(when I assume that even cellphones will use 64-bit processors).
Yes, this is a valid point.
>Keep it in XQuery if you insist, but please don't perpetuate the distinction
>in a conformance level of XQuery that attempts to find the 80/20 point.
>And if that's not the question on the table, sorry.
XQuery implementations will not need the more limited types. There may well
be a place for them in XML Schema. Once someone can use them in a schema,
XQuery has to support them.