Re: namespace node implementation

On Wednesday, Oct 22, 2003, at 11:22 Europe/Berlin, Kay, Michael wrote:

> > Unfortunately, XQuery has a few places like that where the compiler
> > needs to carry the compile-time namespace bindings through to the
> > runtime.  It looks like this area of XQuery is still unsettled.
>
> The worst case of this was casting QNames to strings, which will 
> disappear in the next draft (for this very reason). There are still 
> cases where the namespace context is needed at run-time, but they are 
> all situations where you can tell at compile time that you will need 
> the namespace context at run-time (if you follow me!).

the worst cases which i observed were those where it was possible to 
specify the name of a function in terms which included a prefix to be 
resolved at runtime.

this is the logical equivalent of doing

(apply (symbol-function (intern some-symbol-name (find-package 
*some-package*))) argument-list)

which, i hope, one would agree, is not the kind of thing which one 
would intend to elevate to a general principle for the implementation 
of large scale applications.

...

Received on Wednesday, 22 October 2003 06:54:48 UTC