[COAS-List] RE: Units Quantities and Measurements (UQandM)

[Cory Casanave <cory-c@dataaccess.com>]

Enclosed is a "value framework" that is an -example- of what could be
the base for a complete framework of units.  This framework does not
wrap elemental types (such as hours) as value objects, something that
would be required for full support of units.  To really support units,
every value has to have a type (assertion: the word "unit" is just a
specialization of the word "type", "unit" being used for value types
representing a measure).  What this does show (for time) is how you can
bring all units together in common abstractions.  Doing so allows for
understanding the relationships between various units and the
information required for automatic conversion, type safety, formatting
and (to some extent) automated calculations.
Cory

 
	----------
	From:  Tim Brinson[SMTP:tim@protocol.com]
	Sent:  Tuesday, July 14, 1998 7:52 PM
	To:  Richard Dixon
	Cc:  Timothy Slidel; Pablo Madril; Jeanette Breton; Harold
Solbrig; Evan Wallace; Eric Neumann; Edward Barkmeyer; David Schramm;
David Benton; Cory Casanave; Chris Nelson; coas@cs.fiu.edu
	Subject:  Re: Units Quantities and Measurements (UQandM)

	I never responded to Richard's message before but I offer strong
support
	for it.  As Richard mentioned CORBAmed will be dealing with this
NOW
	(well next week and the week after in Helsinki) for the Clinical
	Observations Access Service (COAS).  While the COAS RFP final
submission
	deadline is for late August the submitters currently think it
should be
	post poned a few meetings.

	We are already looking at models for Measurements from various
	healthcare efforts.  Please forward any ideas on modeling
measurements
	by Monday 20 July so we can consider them at next week's meeting
in
	Denver.  On Monday 27 July (in Helsinki) the COAS submitters
will be
	meeting all day and this subject is in scope for that meeting
too.  You
	are invited to attend.  

	We know this will not be the end all for modeling measurements
but we
	need something now that can be massaged over the next few
months.  Our
	experience with them can also provide feedback to any larger
	standardization effort.  The COAS effort will also be concerned
about
	how measurements can be passed efficiently between remote
systems.


	Regards,

	Tim Brinson


	Richard Dixon wrote:
	> 
	> Hi everyone,
	> 
	> Before the Orlando meeting, I raised the issue of UQ&M with
Harold and I
	> know there was some discussion on the topic.  I tried to get
those
	> interested in this together briefly at Orlando, but as it
turned out  most
	> of the interest was from CORBAmed - maybe because folks didn't
get chance to
	> come/see the notice or maybe because the HealthCare domain has
more demands
	> on this than others.
	> 
	> A brief summary of specific points before and at Orlando:
	> 
	> * Should measurements be part of Common Domain Objects (CDOs),
independent
	> of the BOCA?
	> * Should a "measurement" type be included within further
BOCA-related RFPs?
	> * Should some form of Measurements or Quantitative
Observations RFP be
	> issued?
	> * Can we come up with a 'common' spec for measurements and
keep all Domains
	> happy?
	> 
	> If I have this correct (Jeanette, please correct me if I
haven't), the
	> MES/MC Working Group of the Manufacturing DTF took the
following
	> assumptions:
	> 
	> "In addition to being requested for specifics within domain
RFPs, types such
	> as 'measurements' should alos be specified as part of the BOCA
extensions as
	> BocaValueTypes in the BOCA II RFP in the BO DTF."
	> "Some group of interested folks need to do a 'test case'
definition of a
	> type that is clearly useful within and outside of business
objects
	> ... If there are methods that would not make sense outside of
BOCA we should
	> define the subset that was common and thereby be compatible
with the BOCA
	> superset"
	> "Specific domain efforts could ask submitters to reply to
specific domain
	> reqs on types such as Measurement' ..."
	> 
	> Tim (Slidel) raised the question:
	> "Is there a mechanism by which domain specifications can work
either
	> independently or within the BOCA? ... It is certainly a
repeating problem
	> that as soon as something appears to be more globally useful
it
	> gets pushed towards the BODTF..."
	> 
	> Several people expressed the view that they felt
'Measurements' and other
	> CDOs should be developed independently of the BOCA.
	> 
	> Ed Barkmeyer mentioned ISO 10303-41:1994 and another spec in
	> pharmaceuticals.  I know someone from Stats (Chris?) gave me a
further ref
	> at Orlando (if only I could find where I wrote this down).
	> 
	> >From the CORBAmed perspective, something is needed asap for,
among other
	> things, the Clinical Observation Access Service (COAS) work.
I have models
	> for Units, Quantities and Measurements that I will certainly
be happy to put
	> into the pot and I know others (Pablo for one) have other good
models and
	> work done on this.
	> 
	> Since, as we know, the issue of UQ&M crops up in many domains,
it seems IMHO
	> that - in parallel to what individual DTFs have to produce
themselves - we
	> should try to pool what everyone is doing and see if we can
indeed define
	> Common Domain Objects.
	> 
	> Now, we could carry on discussing exactly whether the results
should be part
	> of the BOCA or not or how generic the results can be etc. for
a long time,
	> but IMHO it does not matter (at least at this stage).  What
matters is that
	> we try to get the model/work together - we can decide what to
do with the
	> result later or in parallel, but let's get down to the work.
	> 
	> So, in order to try to do this, I am quite happy to volunteer
to take a lead
	> (hence this email) unless anyone else wants to.  Can I in
doing so, request
	> that anyone who has or knows of a specific paper, RFP (or
specific part
	> thereof), standards doc (or part of), in-house model, etc.
that we can take
	> as input , please respond with an email with either the
document(s) attached
	> or a reference to where and *how* to get hold of them or a
person to
	> contact.
	> 
	> I have been putting a draft together of many of the issues
involved with
	> UQ&M which I could get tidied up and circulate if anyone wants
to see it.
	> 
	> Anyway, look forward to hearing from you,
	> 
	> Best regards,
	> Richard
	> 
	> ---------------------------------------------------
	> Dr. Richard Dixon
	> Medical Informatics Group (MIG)
	> University of Hull
	> Hull HU6 7RX
	> UK<<File: vcard.vcf>>

ValueFramework.PDF