Re: [XSLT 2.0] Conformance levels

>>>>> "Michael" == Michael Kay <mhk@mhk.me.uk> writes:

    Michael> Colin, You commented in

    Michael> http://lists.w3.org/Archives/Public/public-qt-comments/2004Feb/1237.html

    Michael> to the effect that you felt it should be legitimate for a
    Michael> conformant processor to implement some but not all of the
    Michael> features required in the schema-aware conformance level.

    Michael> The Working Group debated this. The WG understands your
    Michael> reasons, but felt that it was undesirable in the
    Michael> interests of interoperability to allow implementors to
    Michael> "cherry-pick" the features they liked from the spec,
    Michael> while leaving out those they didn't like or found harder
    Michael> to implement.

    Michael> For an implementor aiming to achieve full conformance but
    Michael> who has not yet reached that level, there are two
    Michael> options:

    Michael> (a) you can put a switch in your product that disables
    Michael> all features beyond the basic conformance level, and
    Michael> announce that the product, with this switch set, conforms
    Michael> at the basic level

    Michael> (b) you can simply announce to your users that you have
    Michael> implemented all the features required of a basic level
    Michael> processor and more, while avoiding a formal claim that
    Michael> your product is conformant.

    Michael> I know that this isn't what you wanted, but I have to ask
    Michael> if you will accept this response to your comment.

Having thought about it at greater length since then, I think it is
quite satisfactory.
-- 
Colin Paul Adams
Preston Lancashire

Received on Friday, 20 August 2004 13:40:29 UTC