RE: A question re: using XML validation for ensuring proper unitsused in instance document

Dear Deborra,

 

Sorry that I didn’t reply sooner. I’ve copied the
OASIS UnitsML TC members with my reply because others in the TC are more
familiar with implementing UntisML; I’m more on the development side.
Here are my comments on this issue:

 

1)    
For those developers who wanted to restrict their users to a
specific list of units, we always anticipated that a custom “database”
as you describe would be maintained by the developers. This “database”
could be incorporated into every XML file, or it could maintained separately at
a fixed URL.

2)    
We’re in the process of developing a Units Database
(UnitsDB) which will eventually include code lists provided by various
communities that are “linked” to units that are in UnitsDB. E.g.,
if the military uses “KM” for kilometer, then we would maintain a
list of all the military codes for various units and it would include the code “KM”
linked to the unit km in UnitsDB. A user could then request all of the military
units output in UnitsML. There is an element in the UnitsML Schema that allows
for providing the code value for the unit in a specific code list. Conceivably,
we could also output the code value as the unique identifier for each unit; or
the user could make that substitution.

3)    
Unless there is a great demand for it, we cannot guarantee that
UnitsDB would be available 24/7, which is why we would expect that developers
would maintain their own “database.” However, NIST does have
experience with providing networked  time data on a 24/7 basis to the financial
community, so we would consider providing that service for UnitsDB if we
thought it was essential to our user communities.

4)    
Regarding your question below, I don’t have a definitive
answer (others on the TC might provide more input). I’ve attached a
sample XML output from UnitsDB. For each unit, there is at least one
QuantityReference available that could be used to guarantee that a user, for
example, only uses the unit meter to designate a length and not a weight.

 

We’re having a teleconference TC meeting tomorrow. Ordinarily
only TC members are allowed to participate in the TC meetings. However, if you
wanted to attend this meeting and make a few remarks or ask a few questions, I
would allow that. If it seems that the conversation is extensive, we may need
to set up a separate teleconference meeting to address this issue, or we may be
able to handle it through email messages.

 

Details for the TC meeting:

The next UnitsML TC teleconference meeting is at 10:00am –
12:00 noon EDT Wednesday, Sept 16, 2009.

Call Info

Toll Free: 888-946-7619

Toll: 212-287-1632

PASSCODE: 13284

 

I hope you find this information to be useful.

 

Sincerely,

Bob Dragoset

 

 

Robert A. Dragoset, physicist

Chair of the OASIS UnitsML Technical Committee

Physics Laboratory

National Institute of Standards and Technology

NIST

100 Bureau Drive, Stop 8400

Gaithersburg, MD 20899

Phone: 301-975-3718

Email: 

Website: http://physics.nist.gov

 

 

 

 

From: Deborra J Zukowski
[mailto:] 

Sent: Thursday, September 03, 2009 4:05 PM

To: 

Cc: David Asano

Subject: [unitsml-comment] A question re: using XML validation for ensuring
proper units used in instance document

 

Greetings,

My group at Pitney Bowes is investigating how
best to use unitsML for descriptions of mail items. We are building an XML
schema that represents the physical dimensions of a mail item, including width,
height, thickness, weight, etc. One of the things we want the schema to ensure
is that those who use the schema use one of a set of proper units for a given
dimension, e.g., length dimensions can be in, cm, m, yd but not lb or oz. 

 

Our approach for this is to create our own
units "database," i.e., an instance document that conforms to unitsML
that uses a pattern for the identifiers of the units. This way, we can use
standard XML validation for ensuring that appropriate units are used. For example,
the identifier for a specified inch unit is "distance-in" and one for
a cm unit is "distance-cm." We create specialized measurement types
that restrict the identifiers used in a unit attributes to the pattern. This
approach means, as far as we can tell, that we will need to always have our own
"database." 

 

Question: is there a better way in unitsML
where we can use the XML validation feature to ensure that the proper units
(i.e., one from a set) are used for a given quantity? If not, is there a way
that we could have a "sameAs" description (URL/IRI and units
identifier) in the UnitType (e.g.) where we could connect our description of a
unit to your standard description of a unit, and so enable applications to
leverage information, like conversions, etc, that look to be a part of a bigger
and more complete database?

 

Thanks,

Deborra Zukowski

Pitney
Bowes

 

UnitsDB_output.xml