[eml-dev] annotating log-transformed units

[Margaret O'Brien]

Hi Carl -
With regard to the term "standard unit":
as it's used in EML, I tend to think of it as a "pre-defined unit". The 
word 'standard' has too many other connotations, and none of them really 
applies here.

There are measurements that reduce to nearly unitless, but where it's 
important to the scientists to know what the reference units were; 
specific uptake rates are some we deal with. When this is the case, I 
create a customUnit to maintain that meaning. So in your case it could 
be "log centimeter". If you went with "dimensionless", you'd have to 
make it obvious that the reference unit was centimeter in one of the 
text fields (attribute/definition or attribute/name), but neither of 
these is as obvious as in the unit name itself.

Margaret

-----------
Margaret O'Brien
Information Management
Santa Barbara Coastal LTER
Marine Science Institute, UCSB
Santa Barbara, CA 93106
805-893-2071 (voice)
http://sbc.lternet.edu

On 11/26/13 10:59 AM, Carl Boettiger wrote:
> Yes, I agree entirely that it would be nice if researchers always 
> defined a reference length; e.g. jaw length as a fraction of body 
> length, etc, and then the units are pure and it makes sense to take a 
> logarithm.
>
> I suppose that's the role the standard unit plays when researchers 
> measure a bunch of values in some unit (e.g. cm) and then take the 
> logarithm, we can think of this as a dimensionless value: 
> length(measured)/length(cm).  But this begs the question of how we 
> denote the reference unit (denominator).  I can add it to the column 
> metadata (attributeDiscription) in plain text, but it seems a better 
> solution would provide a more machine-readable way to, e.g. recover 
> the initial lengths.
>
> Thanks for the feedback and suggestions.
>
>
> On Tue, Nov 26, 2013 at 10:08 AM, Matt Jones <jones at nceas.ucsb.edu 
> <mailto:jones at nceas.ucsb.edu>> wrote:
>
>     I think that both the input to and output of the log function are
>     dimensionless numbers.  So I would use dimensionless in my EML
>     unit. See these discussions:
>
>     http://physics.stackexchange.com/questions/13060/what-is-the-logarithm-of-a-kilometer-is-it-a-dimensionless-number
>     http://www.sciforums.com/showthread.php?64346-what-s-the-UNIT-after-you-take-the-natural-log
>
>     Matt
>
>
>     On Tue, Nov 26, 2013 at 6:34 AM, Carl Boettiger
>     <cboettig at gmail.com <mailto:cboettig at gmail.com>> wrote:
>
>         Hi eml-dev,
>
>         Any advice on the best way to write the EML attribute metadata
>         for, e.g. a column whose units are log-transformed lengths
>         measured in centimeters?  (Logarithms being 'pure',
>         formally).  (I see definitions for powers but not logs among
>         the standardUnits
>         <http://knb.ecoinformatics.org/software/eml/eml-2.1.1/eml-unitTypeDefinitions.html>)
>
>
>         Would this require a custom unit definition? Seems like this
>         might be a reasonably common use-case, so curious how others
>         are handling it.
>
>         Thanks and sorry for the bother,
>
>         - Carl
>
>         (referred from reml/issues/57
>         <https://github.com/ropensci/reml/issues/57>)
>
>         -- 
>         Carl Boettiger
>         UC Santa Cruz
>         http://carlboettiger.info/
>
>         _______________________________________________
>         Eml-dev mailing list
>         Eml-dev at ecoinformatics.org <mailto:Eml-dev at ecoinformatics.org>
>         http://lists.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/eml-dev
>
>
>
>
>
> -- 
> Carl Boettiger
> UC Santa Cruz
> http://carlboettiger.info/
>
>
> _______________________________________________
> Eml-dev mailing list
> Eml-dev at ecoinformatics.org
> http://lists.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/eml-dev