Lists Home |
Date Index |
At 03:31 PM 7/6/2002 -0400, Mike Champion wrote:
>7/6/2002 2:01:33 PM, Jonathan Robie
> >These are precisely the reasons I want to be able to have type safety when
> >I need it. If schemas evolve in a way that changes the name of the 'price'
> >element to 'list-price' for some instances but not others, then my
> >function will always computed 0 as the total
>OK, well, it's kindof the opposite approach that WorldCom and Enron took to
>accounting and might go over well in the modern world, but most companies
>I'm familiar with would be a tiny bit upset to get a check for $0.00 because
>they sent an invoice with the new tag for the price field rather than the
>old one :~) And most CEOs get a tiny bit upset when the IT folks
>throw away revenue opportunities because someone sent them a PO with an
>extra <p> tag or whatever ...
And if you use the static typing, this error will not occur. (I know *you*
caught this, Mike, but if someone is just reading this snip they might miss
it, so I'm being painfully clear).
>But seriously, I can understand where you're coming from and understand that
>XQuery does not *force* one to "eat their spinach" and do the strongly
>typed Right Thing in order to use it. Still, it would be more effective
>in evangelizing XQuery to emphasize that while you have invested heavily
>in making strong typing / PSVI / etc. work, XQuery does not impose this
>on end users.
Yes, I agree.
>It does impose heavily on implementers, which is why I fear that we will see a
>repeat of the SQL (92 and 99) situation in the XML world -- no single
>vendor will support
>all the spec, at least in the same way as others, for some time to come.
I also agree here.
We are working on defining conformance levels, including one level that
does not require static typing, and one level that does not reven equire
support for schema imports.