[GO-ESSP] Re: Units and NERC DataGrid

[John Caron]

Robert Miner wrote:

>Hi.
>
>You wrote:
>
>  
>
>>I also feel that what we need is more along the lines of a semantic web
>>resource rather than a downloadable executable approach: in other words
>>a units ontology that provides a units vocabulary plus the knowledge
>>about which units may be interconverted and how.  Such an ontology could
>>be interlinked with parameter vocabularies indicating the subset of
>>units that may sensibly be associated with particular parameters.
>>    
>>
>
>For what it's worth, your ontology idea sounds more like what the W3C
>Math group hears requested from its user community.  
>
>As I follow the discussions here, it seems to me that there are
>several kinds of units "standards" people need.  One is some sort of
>markup for defining and describing units and their interrelations.
>The GML schema, for example, seems to live mostly in this space.
>
>However, another group of users isn't so interested in the units
>lanugage per se.  Instead, they want an authoritative central registry
>of units of some sort, so that units definitions can be registered,
>referenced, retrieved and converted by various user communities.  One
>would expect standard physical units would be pretty much static, and
>widely referenced.  By the same token, more specialized or
>idiosyncrantic units would be more volatile, and primarily shared
>within small communities of practice.
>
>In the case of MathML particularly, I don't think people really want
>to inline large blocks of units markup.  Rather, they just want to
>indicate that this "mu" is the unit defined at XXX.  
>
>--Robert
>
>------------------------------------------------------------------
>Dr. Robert Miner                                RobertM at dessci.com
>W3C Math Interest Group Co-Chair                      651-223-2883
>Design Science, Inc.   "How Science Communicates"   www.dessci.com
>------------------------------------------------------------------
>
>
>  
>

This sounds pretty much right to me. I'll just add an elaboration, based 
on what I think is needed for scientific datasets.

Data writers choose their units based on their own "communities of 
practice". They are sometimes willing to follow standards if those 
standards really fit their domain, but sometimes even then they wont, or 
dont know about them, Also theres plenty of archives that predate 
whatever standards now exist. What middleware like THREDDS wants to do 
is to map the dataset's units to a standard, by annotating the dataset 
in some way.

There are at least 2 ways you might want to do that. 1) for each place a 
unit is used, indicate what its YYY unit is in standard XXX. 2) for the 
entire dataset, point to a document that is a mapping of their 
vocabulary to a standard, so that every instance of the unit string yyy 
gets mapped to YYY at XXX. With that indirection, you can hopefully add 
a single annotation to a dataset or collection of datasets and transform 
the units to standard units. Also, you can give people lists of unit 
strings to use, and not burden then with XML arcana.

There is also likely a need for mappings between standards. Its 
important to let communities grow their standards with loose couplings 
between other efforts.  (Just to clarify- these would be different 
standards that use the same language to specify). This is probably a 
different mapping language then the first, but perhaps there are some 
common pieces, or perhaps they can be usages of the same language.

(Note:  Im: not quite sure how this discussion got off the 
scimarkuplang-list . ive put it back on. I wonder if interested 
individuals should get on that and continue there?)