RE: Are the static inference rules normative or non-normative?

Dear Murata-San

Since we have not yet finalized the conformance section, we have not yet
a full answer to your question.

The latest discussion was that the static analysis is normative as a
lower-bound and implementations could infer more precise types. What is
needed for conformance, I don't know yet.

I personally find the loss of type information on backwards-axes to be
problematic as well. If you would like to propose more precise inference
rules, I am sure we will look at them and see if we can/want to
incorporate them.

Thanks
Michael

> -----Original Message-----
> From: public-qt-comments-request@w3.org [mailto:public-qt-comments-
> request@w3.org] On Behalf Of MURATA Makoto
> Sent: Monday, September 15, 2003 11:32 PM
> To: public-qt-comments@w3.org
> Cc: mm
> Subject: Are the static inference rules normative or non-normative?
> 
> 
> Are the inference rules for static analysis normative or
non-normative?
> In other words, are all implementations of static analysis requierd to
> follow the static inference rules as specified in the formal semantics
> spec?  Suppose that I construct a static inference engine that takes
> full advantage of backward axes.  That engine would create types
> narrower than those created by the static inference rules in the
formal
> semantic spec.  Is that engine non-conformant?  This would be very
> unforunate.
> 
> --
> MURATA Makoto <mmurata@trl.ibm.com>
>