eml 2 conference call

[Peter McCartney]

Ill remember to thank Bill for creating RCN so i had an excuse to miss it.



regarding date values, i dont think it would be serving anyone's purpose to
punt and declare that dates are ordinal scales with no concept of the
relative difference between 1/1/2000, 1/2/2000, and 1/1/2002. I agree that
software packages seem to differ in how then do the  math to assess the
differences, but they do do it. So i think you're going to find a big hole
in your foot if you choose the ordinal path for dates.


-----Original Message-----
From: Matt Jones [mailto:jones at nceas.ucsb.edu]
Sent: Tuesday, October 29, 2002 3:40 PM
To: eml-dev at ecoinformatics.org
Subject: eml 2 conference call


Today we had a conference call to make some decisions about the release 
of EML 2.0.0.  Present were: Matt Jones, Chad Berkley, Scott Chapal, 
David Blankman, and Tim Bergsma.  Here's an exceedingly brief summary 
for those not present:

1) Discussed changes to measurementScale, attributeDomain, and unit that 
are in CVS. Agreed these were ok, but fairly substantial (thus, probably 
need a rc3 so peter and others can see the changes).  Everybody liked 
that unit was no longer required (or even possible) for nominal attributes.

2) Discussed how date value work, and agreed we need a consistent way to 
flag that a value is a date.  Agreed that they are ordinal-scale values 
(ordered, but intervals between units are not a constant number of 
seconds).  Decided to modify the ordinal type to specifically deal with 
date values because these are so prevalent and cause such extreme 
problems for software.  Matt will do this.

3) Discussed units, specifically how dimensionless works.  Discussed pH, 
Reynolds numbers, and other dimensionless values.  Decided there was a 
clear need for dimensionless in the unit dictionary.  This brought up a 
bit of a conversation about "dimensionlesss" units where the units of 
the quantity in the general sense cancel, but not in the particular 
sense.  For example, grams Carbon per grams soil.  The g/g cancel, but 
that doesn't tell you that it is wrong to add grams Carbon with grams 
Phosphorus.  Got into discussion of semantics.  Agreed that we don't 
fully understand these issues, and would need months to work them out 
fully.  Agreed to go with the more mechanical approach where units don't 
have a semantic component for now.  So, someone describing an attribute 
with units "grams Carbon" would just use "grams" for the unit. 
Dimensionless units can either be labeled dimensionless, or g/g if the 
user feels it is important to preserve the info about the constituent 
parts.  We came up with a couple of examples where this would be an 
issue.  Chad agreed to make units changes.  Agreed we need another 
formal, systematic review of the unit dictionary -- higgins will do it, 
and David will ask Barbara Benson if she could do it.

4) Decided to write some of this rationale into the spec.  Tim will 
write up after chad creates a section.

5) Agreed we need a Release Candidate 3, because the changes were fairly 
substantial.  Matt and Chad will prepare and release eml2.0.0rc3 on 
Wednesday Oct 30.  It will be available for 1 week, and we will release 
EML 2.0.0 on Wed., Nov 4.  We will ask for formal approval of all EML 
Project memebers for this final release.

That's about all I remember :-)  We talked for about 2.5 hours, so you 
know how that goes.

Matt

-- 
*******************************************************************
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
*******************************************************************

_______________________________________________
eml-dev mailing list
eml-dev at ecoinformatics.org
http://www.ecoinformatics.org/mailman/listinfo/eml-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mercury.nceas.ucsb.edu/ecoinformatics/pipermail/eml-dev/attachments/20021030/6f70b5c9/attachment.htm