Re: Feedback on candidate recommendation

Noah writes,

> First of all, it is known and intentional that equivalence classes (now 
> called substitution groups) do support interactions across namespaces.  It 
> is further known, and I _think_ this is the essence of your concern, that 
> a reference to the head of such a group represents a disjunction (a 
> choice) across all the elements in the group, and that the content model 
> can therefore be seen to "change" as you dynamically aquire schemas.

I think the lesson of Volker's work is that it is not obvious from the
specification, and as a result he picked an implementation style that
did not work well.  Hence, it might be useful if somewhere was
recorded the advice that viewing a substitution group as a disjunction
can lead to this implementation problem.

In fact, this particular problem is not my greatest concern; I simply
relayed Volker's concerns as he expressed them to me.

My greatest concern would be the difficulty that Volker had in
implementing the rather ad hoc rules to test when one type is a
restriction of another.  With substitution groups and imports, you
can argue that the complexity arose from a bad decision about how
to implement the specification.  With the ad hoc definition of
restriction, I would argue that the complexity is inherent in
the specification and all implementations will suffer from it.

Cheers, -- P

Received on Thursday, 18 January 2001 09:53:28 UTC