[
Lists Home |
Date Index |
Thread Index
]
At 11:35 PM 12/4/2002 +0000, Sean McGrath wrote:
>I see it is really simple terms:
>
> <Example>
> <Weight>123.456</Weight>
> </Example>
>
>Q. What is the value of the element Weight?
>A. It is the *string* "123.456".
>
>Q. Is it not actually the floating point number 123.456?
>A. Nope. I might interpret it that way. It depends on how I am processing the
>document. I make the decision in my processing - it is not up to XML to
>tell me how
>to interpret the text.
As far as I can tell, XQuery and XSLT will both allow you to process data
in this way. Programmers need flexibility to choose from a variety of
strategies, and this is one perfectly acceptable approach for many kinds of
applications. If someone puts datatype information in a document, does it
get in your way when you try to program using your preferred approach? If
so, can you show me an example that illustrates the problem?
Some people are saying that it is wrong for an application to add datatype
information to a document stating that this data is a floating point
number, or that it is wrong for XQuery to support the interpretation given
by datatypes in addition to the purely lexical processing of data that it
supports. Does XQuery limit you when you try to process information using
your preferred approach? Again, concrete examples would be helpful.
>Q. Is it a good idea to bind the concept of "floating point number"
>tightly into the processing
>of this XML document? Would this not be wonderfully convenient?
>
>A. No, because interpreting it as a floating point number is an
>*interpretation* of the text, not
>a fundamental part of the text. My process will do that interpretation,
>thank you very
>much.
And nobody is constraining your process from doing that. On the other hand,
some people seem to be arguing that also supporting datatypes causes
problems. Do you agree with them? Can you provide some examples that
illustrate the problems?
Jonathan
|
