Tony Hunt's recent message on the convergence of
     experimental data and data for modeling agrees with my
     experiences at CIAT and CIMMYT.  We are fast moving to a
     situation where researchers will want to have all of their
     data stored in a database and have the option to examine
     the results with models.  One might add that in
     international centers, where data are publicaly owned,
     administrators may start requiring such databases.  As you
     finish a project, data will be deposited in a data
     wharehouse as a condition for terminating the project.
 
     Concerning the present file system in DSSAT3, I encourage
     people to take Tony up on his request for feedback.  The
     better we make the system, the more widely it will be used
     and supported.
 
     My thoughts are as follows:
 
     1)  Site location should include latitude, longitude,
     altitude.  Ten years from now a farm or institute may have
     changed names or owners, so a FILEX should always be tied
     to an exact location.  GPS-level accuracy can be assumed
     (e.g. decimal fractions of seconds).
 
     2)  Associated publications should be cited in FILEX.  For the
     moment, I use the "notes" section for references, but I
     would prefer to see a separate section.  My experience
     with other databases is that "note" sections create
     a lot of problems for searches.
 
     3)  Data distribution limitations should be noted.  People
     may want to submit data to databases with restrictions on
     distribution (particularly "in-house" databases).  Options
     might be:
 
         Public domain - no limitation on use
         Acknowledgment requested - may be used, but contingent
     upon acknowledgment in publications
         Institutionally held - cannot be released externally
     without permission
 
 
     4)  Number of treatments should be increased.  At the
     moment FILEX accepts up to 999 treatments, but that is done
     using up the blank in column 3.  I would add two columns to
     permit up 9999 treatments.  This is mostly of interest for GIS
     applications, but I have run replicated photoperiod
     screening trials with over 1000 entries.
 
     Best regards,
     Jeff White
 
 
 
 
----- End forwarded message
 
 
 
-----------------------------------------------------------------
CENTRO INTERNACIONAL DE MEJORAMIENTO DE MAIZ Y TRIGO (CIMMYT,INT)
        International Maize and Wheat Improvement Center
 
            Note INTERNET address:   [log in to unmask]
 
   Lisboa 27  Apdo. Postal 6-641  CP 06600  Mexico, D.F. MEXICO
Telephone: +52(5)726-9091                     FAX: +52(5)726-7559
-----------------------------------------------------------------