#6 interface --> final class

[Marin Dimitrov]

the following interfaces may be better off as final
classes, since it is highly unlikely that anyone will
provide an alternative implementation:

org.wsmo.common IRI, Identifier, Namespace
org.wsmo.validator ValidatorWarning, AttributeError

[holger]

Logged In: YES
user_id=1125735

I disagree here. Jacek made already an own implementation of
IRI for his versioning. Furthermore I exepect multiple more
specific interfaces / implementations for the Warning an Errors.

Or porbably I misunderstand your intention here?

[Marin Dimitrov]

Logged In: YES
user_id=548066

>I disagree here. Jacek made already an own implementation of
IRI for his versioning.

I'm still not convinced this was necessary. Can you elaborate?

>Furthermore I exepect multiple more
specific interfaces / implementations for the Warning an
Errors.

I didn't say there won't be new Errors and Warnings - I said
they should be classes that subclass the top Error/Warning
class. Just like Java exceptions do

[Marin Dimitrov]

Logged In: YES
user_id=548066

besides, Jacek could have used a class + subclassing to
provide an alternative IRI implementation, since providing
an alternative implementation is something more people won't
do anyways

[Nobody/Anonymous]

Logged In: NO

Jacek: I don't technically need a different IRI
implementation, so I can live with this change.