A couple of versioning personal blog entries

I wrote up a couple of personal blog entries on validation by
projection.  This seems to be a useful way of achieving forwards and
backwards compatibility without relying upon schemas that have wildcards
or open content models.
 
>From the TAG's definitional perspective, I'd characterize validation by
projection as an architecture where the schema(s) define a Defined Text
Set and an Accept Text Set that is equal to the Defined Text Set, then
the process of projection is the creation and validation of the text
against a generated Accept Text Set that has the original Accept Text
Set plus all possible extra undefined elements and attributes.  It could
also be said that a system with validation by projection has an Accept
Text Set that contains the Defined Text Set plus all possible extra
undefined elements and attributes, and validition is done by generating
a new text by taking the original text and removing all the extra
undefined elements and attributes then applying validation.  I'm not
sure which way I like better.  Of some note, I list Henry's V2S and UBL
in the implementations.
 
http://www.pacificspirit.com/blog/2007/12/12/validation_by_projection_in
troduction
http://www.pacificspirit.com/blog/2007/12/12/validation_by_projection_im
plementations
 
If any of this material was to go anywhere in the finding works, I'd say
in the compatibility strategies work.  But I don't think we currently
have the collective energy to be adding material to the versioning
documents.  Your thoughts?
 
Cheers,
Dave