[exslt] Extension candidates

[Dimitre Novatchev]

>  Ok, I think I see what you mean.  There is on the one hand the
> reference mechanism (referencing return an item of a specific type
> that is able to be dereferenced) and on the other hand the following
> rule: "the processor will recover of a typing error by using
> dereferencing when this is possible."

Yes!

>
> That is a lot of
> work I think to specify exactly under what rules recovering must take
> place, can take place, must not take place, discover and treat the
> ambiguities, etc.

No, the principle, as you yourself formulated is quite simple, it is
not necessary to enumerate long lists of specific cases.

>
>  This rule is quite ambitious and I am not sure it is really possible
> to implement it as an extension,

Why not? The three different developers of XSLT 2.0 processors know
well their products and know exactly where to add this.

Actually, it is interesting to hear what they would say.

Dr. Kay, Colin,  and Altova developers, will it be too difficult to
implement the following principle:

  "The processor will try to recover of a typing error by using
   dereferencing when one of the ill  arguments is a reference".

> but that's interesting!

Definitely! Why, otherwise, would we be doing it?


Cheers,
Dimitre