RE: Fwd: [xml-dev] Data versioning strategy: address semantic, re lation

[Len Bullard <len.bullard@uai.com>]

http://books.google.com/books?id=1nRUIy-Z_0QC&pg=PA607&lpg=PA607&dq=input+co
ntrol+output+mechanism+second+order+systems&source=web&ots=70NIcTT41C&sig=Xx
f1e8vlf3TR2ctIXYyvdgpZSqc

Awful URL but.. this points to a page about adaptive control modeling from a
book entitled "The Industrial Electronics Handbook" (J David Irvin).  Roger
may find this interesting reading.  One has to extend the abstraction of the
control model to the versioning problem in terms of signal correction by
adaptive controls.

While I agree that the problem is one of identifying breaking changes and
that set modeling is part of the approach, it may be insufficient.  It is
also necessary to apply a corrective change, yes?  What if the corrective
model creates more instability?

I want to point to the models where phase transition models come into play
because even if the document (a model itself) is static, the process is not
and in fact, the model may not be sufficient to provide corrections because
it can go 'out of phase'.  Note the necessity of "tuning" and consider if
tuning is a useful model for data versioning in the face of changes to any
of the three domains (semantic, relationship, syntax). 

The paper cites work on third order systems for correcting drift.
Corrections themselves can create failures given a) insufficient information
given time to respond and b) unknown modes that occur in the instance that
are not part of the reference model.  

Consider these in terms of the validation model and the conformance model as
corrective over some set of language instances on the wire given some
frequency of update.  It may be possible to take Noah's sets and combine
these into a formula that includes the adaptive mechanism such as an analog
of integrator leakage, a means of reducing the sensitivity to error perhaps
known as planning some 'slop' in the system.

But get a beer first.

len
This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail.