proposal for GLIMS analysis package file names

[Siri Jodha Singh Khalsa]

As described in the GLIMS data transfer format specification (see the 
"Proposal for RC --> NSIDC data transfer methodology: Rev. 1.11" email from 
Bruce Raup dated Thu, 27 Sep 2001 10:22:02),  the results of an analysis 
will be contained in multiple files. Herein we propose a naming convention 
for these files.

Although we do not constrain file names to the DOS 8.3 format, we do retain 
3-letter extensions to designate the file types:

ASCII files:
.pvl for Parameter-Value Language files
.tsv for Tab-Separated Value format files
.end for Endnote format files

ESRI Shapefiles:
.shp for the shape file
.shx for the index file accompanying a shape file
.dbf for the dBASE table

Multiple glaciers in a region may be analyzed in one "analysis session," 
and the results conveyed in a single "analysis package." NSIDC must be able 
to associate the segments in the shapefile with the analysis session that 
produced them. Each segment in a shapefile will be identified by its 
associated glacierID, but because there may be multiple analyses of any one 
glacier over time, we ask that there be only one analysis of any given 
glacier in an "analysis package."

In an analysis package, the basename (everything in the filename before the 
period that sets off the extension) should be common for the .pvl, .tsv, 
.shp, .shx, and .dbf files.  A suggested convention for the basename is 
GLIMS_Region_number as NNN, an abbreviated (8 character or less) geographic 
placename for the specific region analyzed, and the date the analysis was 
completed in YYYYMMDD format. This creates a basename of at most 19 characters.

For example, an analysis of glaciers in the Chugach Mountains of Alaska 
(GLIMS region 4) done on August 18th, 2002 would have a basename of 
004CHUGACH20020818.  If during the transfer the basename is truncated to 8 
characters by some protocol, the most important identifying information 
should still be visible (e.g. 004CHU~1.pvl)

All the files for one analysis session should be packaged into a gzipped 
tar file (unix) or a zip file (DOS/Windows), conforming to the basename 
convention, e.g. 004CHUGACH20020818.tar.gz or 004CHUGACH20020818.zip.

For a file containing literature references (*.end), the basename may not 
be appropriate, because the references could refer to multiple regions or 
be from multiple RCs.  Links to the glaciers referenced in the documents 
will be made through the information in the file, so the file naming 
convention is not critical. However, we would like to avoid receiving files 
with duplicate names, so we suggest a convention of first author surname 
plus an 8-character (maximum) geographic placename plus a publication year, 
appended by a, b, c, etc., if necessary. For example, BISHOPHIMILAYA1997a.end.

Comments are welcome.


Siri Jodha Singh Khalsa, Ph.D. %%% Emergent IT, Inc.
ECS Science Coordinator for the National Snow and Ice Data Center
Campus Box 449 / 1540 30th St.
Boulder, CO 80309
Phone: (303) 492-1445 FAX: (303) 492-2468
http://spot.colorado.edu/~khalsa
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://nsidc.org/pipermail/glims/attachments/20011123/cf0eaab7/attachment.html>