STMML units
Peter McCartney
peter.mccartney at asu.edu
Wed Oct 9 14:16:39 PDT 2002
I agree...highly relevant. however, the problem Chad and I discussed is a
serious shortcoming that needs to be fixed. Our interest was in a solution
that allows automated semantic resolution of any defined unit down to
universally understood concepts (dimension bases and some operators). the
current markup does not support that. changing the name of unitType to
quantity won't fix the example that Chad provided. Their email didnt
indicate that we had simply overlooked something and it seems like it will
take more than a minor change to solve the problem. Chad's suggestion is
probably the most minimal change that would solve it, although not my
preffered one. I would rather see the entire lineage down to fundamental
dimmensions nested into each unit defintion just as they are in the
spatialReference model. That way the content model can be inlined and you
dont have to juggle so many name pointers (which cant resolve outside the
document anyway).
So.....Im' calling into question the rationale of asking eml2.0 users to
follow the stmml.xsd to define custom units. Its a lot of work for something
that in the end has to be read as text in order to understand its meaning.
The documentation is not ready for primetime and the schema is not
constrained enough to force the production of a useful definition - a valid
stmml file can consist of no more than: <unit id="myunit">. Period. I think
we're safe by just expanding and committing to the enumerated units that
weve defined in rc2 and using the dictionary merely as a tool for generating
the text definition of them via a style sheet (ie, distancing it somewhat
from the "core" of eml, at least for now). I suggest we simply let custom
units be unconstrained, perhaps with a description attribute. Its obvious
from their email that the stmml schema is going to change and the unit
dictionary will have to be regenerated if it is to be valid. Becuase of the
flaw in identifying the dimensions of units that are instances of unitTypes,
there will be no way to automatically convert existing stmml content to the
new version, so i think we should minimize how much of that we build up now.
Peter McCartney (peter.mccartney at asu.edu)
Center for Environmental Studies
Arizona State University
480-965-6791
> -----Original Message-----
> From: Matt Jones [mailto:jones at nceas.ucsb.edu]
> Sent: Wednesday, October 09, 2002 1:23 PM
> To: chad berkley
> Cc: Peter Murray-Rust; rzepa at ic.ac.uk; Peter McCartney;
> h.rzepa at ic.ac.uk; yong Zhang
> Subject: Re: STMML units
>
>
> Hi everyone,
>
> Actually, we discovered it through an announcement Peter made
> on XML-DEV
> a while back. I thought it was highly relevant then, and
> even more so
> now.
>
> Matt
>
>
> chad berkley wrote:
> > Hello,
> >
> > We discovered it via google I think. We had been looking
> for a markup
> > language to define units in.
> >
> > Let me know if I can be of any assistance in interpreting
> the example I
> > sent.
> >
> > chad
> >
> > On Tue, 2002-10-08 at 15:56, Peter Murray-Rust wrote:
> >
> >>At 12:10 08/10/2002 -0700, chad berkley wrote:
> >>
> >>>Dear Sirs,
> >>
> >>Greetings and many thanks for your interest!
> >>
> >>
> >>>We have been interrested in your work on marking up units in a
> >>>structured manner for some time now. We are currently
> using your STMML
> >>>schema in our project (see
> http://knb.ecoinformatics.org/software/eml)
> >>>for unit definitions.
> >>
> >>I am offline but will look asap. I am very interested in
> collaboration.
> >>
> >>
> >>> We have recently run into a problem with parsing
> >>>unit definitions and we were wondering if the schema is
> lacking part of
> >>>the content model that we need, or if we are simply missing some
> >>>important point.
> >>
> >>There are some minor errors at our end and we have been
> passing the schema
> >>through various parsers. We hope to have this completed
> literally in the
> >>next 2 days.
> >>
> >>
> >>>The main problem is that we do not see that there is a way
> to relate the
> >>>dimension type in a unit definition back to the dimension in the
> >>>unitType. Here is a simple example:
> >>
> >>After the paper was published we realised that "quantity"
> is a much better
> >>name that "unitType".
> >>
> >>We are looking to see what minor revisions are possible at
> this stage. I
> >>will look at your example in detail and reply. During CML
> we found that a
> >>few design faults crept in and we are just about to release
> CML2. as a
> >>Schema - which relies on STMML.
> >>
> >>For our interest how did you discover STMML?
> >>
> >>Look forward to further discussion.
> >>
> >>Best
> >>
> >>Peter
> >>
> >>
> >>
> >>>This is how we currently have units marked up:
> >>>
> >>><unitType id="speed" name="speed">
> >>> <dimension name="length" />
> >>> <dimension name="time" power="-1"/>
> >>></unitType>
> >>>
> >>><unit id="meter" name="meter" unitType="length"/>
> >>><unit id="feet" name="feet" unitType="length"
> multiplierToSI=".3048"
> >>>parentSI="meter"/>
> >>><unit id="second" name="second" unitType="time"/>
> >>><unit id="metersPerSecond" name="metersPerSecond"
> unitType="speed"/>
> >>><unit id="feetPerSecond" name="feetPerSecond" unitType="speed"/>
> >>>
> >>>Now, what we would like to do is have an automated process
> understand
> >>>that feet and meters can easily be converted and hence, so can
> >>>metersPerSecond and feetPerSecond. The way the markup is
> now, there is
> >>>no way to correlate metersPerSecond and feetPerSecond,
> except to know
> >>>that they are both a length divided by a time. It seems
> to us that the
> >>>markup could look like the following:
> >>>
> >>><unit name="metersPerSecond" type="speed">
> >>> <dimension name="length" value="meters"/>
> >>> <dimension name="time" value="seconds"/>
> >>></unit>
> >>>
> >>><unit name="feetPerSecond" type="speed">
> >>> <dimension name="length" value="feet"/>
> >>> <dimension name="time" value="seconds"/>
> >>></unit>
> >>>
> >>>where the dimension tag inside of unit relate the base dimensios of
> >>>feetPerSecond and metersPerSecond back to the unitType.
> >>>
> >>>So, the question is, are we simply missing something that
> is already in
> >>>STMML or, do we need to extend the schema to do what we want to do.
> >>>Note that the speed example is a very simple one. We have
> some very
> >>>tricky units to deal with. I've attached our stmml
> instance document if
> >>>you would like to look at it.
> >>>
> >>>If the schema needs to be extended, and you want
> collaboration, we would
> >>>be happy to work along with you on this project.
> >>>
> >>>Thanks in advance for any help.
> >>>
> >>>Chad Berkley
> >>>
> >>>-----------------------
> >>>Chad Berkley
> >>>National Center for
> >>>Ecological Analysis
> >>>and Synthesis (NCEAS)
> >>>University of California, Santa Barbara
> >>>735 State St. Suite 300
> >>>Santa Barbara, CA 93117
> >>>berkley at nceas.ucsb.edu
> >>>-----------------------
> >>
>
>
>
> --
> *******************************************************************
> Matt Jones jones at nceas.ucsb.edu
> http://www.nceas.ucsb.edu/ Fax: 425-920-2439 Ph: 907-789-0496
> National Center for Ecological Analysis and Synthesis (NCEAS)
>
> Interested in ecological informatics? http://www.ecoinformatics.org
> *******************************************************************
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/eml-dev/attachments/20021009/810f88d8/attachment.htm
More information about the Eml-dev
mailing list