Erik, The condition you are describing would definitely be best served by a (low,high] bounds condition (exclusive lower, inclusive upper). I think we need to maintain some form of consistency regarding bounds for the sake of users and automated software, and I feel that double-inclusive bounds are a bad plan, but I also think you have posed a use case that points out a weakness in the convention about bounds. I'm now going to sit around and think too much about the implications of doing (low,high] bounds without any metadata to indicate which way the bounds are to be interpreted, then get back to you with more thoughts.
Grace and peace, Jim [image: CICS-NC] <http://www.cicsnc.org/>Visit us on Facebook <http://www.facebook.com/cicsnc> *Jim Biard* *Research Scholar* Cooperative Institute for Climate and Satellites NC <http://cicsnc.org/> North Carolina State University <http://ncsu.edu/> NOAA National Centers for Environmental Information <http://ncdc.noaa.gov/> *formerly NOAA’s National Climatic Data Center* 151 Patton Ave, Asheville, NC 28801 e: [email protected] o: +1 828 271 4900 *Connect with us on Facebook for climate <http://www.facebook.com/NOAANCEIclimate> and ocean and geophysics <http://www.facebook.com/NOAANCEIoceangeo> information, and follow us on Twitter at @NOAANCEIclimate <http://www.twitter.com/NOAANCEIclimate>and @NOAANCEIocngeo <http://www.twitter.com/NOAANCEIocngeo>.* On Mon, Mar 26, 2018 at 5:21 PM, Erik Quaeghebeur < [email protected]> wrote: > Dear Karl, > > > Thanks for your reaction. > > I think one could argue that a "sample" taken *on* the second is most >> representative of an interval extending from half a second prior to the >> sample time and half a second following the sample time, […] >> >> On the other hand, if there is some finite response response time of your >> instrument (say, of order 1 sec), you could argue that a sample taken at >> time t really represents an average over some preceding interval, […] >> > > I hadn't considered taking the sampling times as representatives of > lower-level intervals. That will often be the case. But I think that to > avoid having to know all details of the underlying data generation process, > we may assume for purposes of the bounds variable for the derived > statistics that the samples are instantaneous. (Feel free to correct me.) > > > > Best, > > Erik > > -- > https://ac.erikquaeghebeur.name > _______________________________________________ > CF-metadata mailing list > [email protected] > http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata >
_______________________________________________ CF-metadata mailing list [email protected] http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
