Actually, using my catalog as an example, I was thinking that the parent
dataset "metadata" (html) page like this first one:
http://dodsdev.gso.uri.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://dodsdev.gso.uri.edu:8080/thredds/catalog.xml&dataset=SSM/I%20Pathfinder%20Wind%20Speed%20-%20Level%202
would have a link in it to its CatalogRef:
http://dodsdev.gso.uri.edu:8080/thredds/catalogServices?catalog=http://dodsdev.gso.uri.edu:8080/thredds/cataloggen/catalogs/JPL054.xml
and also in reverse, a dataset node like this:
http://dodsdev.gso.uri.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://dodsdev.gso.uri.edu:8080/thredds/cataloggen/catalogs/JPL054.xml&dataset=1987-217
would have a link back to its parent metadata page (first page given
above):
http://dodsdev.gso.uri.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://dodsdev.gso.uri.edu:8080/thredds/catalog.xml&dataset=SSM/I%20Pathfinder%20Wind%20Speed%20-%20Level%202
In your example it would be the same as having a link from your data node:
http://motherlode.ucar.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://motherlode.ucar.edu:8080/thredds/idd/model/DGEX/Alaska_12km/catalog.xml&dataset=NCEP/DGEX/Alaska_12km/DGEX_Alaska_12km_20051221_0000.grib2
to this html page for the parent dataset:
http://motherlode.ucar.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://motherlode.ucar.edu:8080/thredds/idd/model/DGEX/Alaska_12km/catalog.xml&dataset=NCEP/DGEX/Alaska_12km
What do you think, is at least one of these links pretty doable? Thanks
again.
Carl
John Caron wrote:
address@hidden wrote:
Late in the day, probably didn't make myself very clear. We felt
that online users who accessed the catalog might find themselves
navigating within these html pages. And as a convenience, if they
were on the parent dataset page (for a particular dataset) where the
complete metadata is presented, we wanted to make it easy to directly
go to the CatalogRef listing for that dataset without going back to
the main thredds catalog page at http://dodsdev.gso.uri.edu/thredds/
and linking through there.
so a catalog like
http://motherlode.ucar.edu:8080/thredds/catalogServices?catalog=http://motherlode.ucar.edu:8080/thredds/idd/model/DGEX/Alaska_12km/catalog.xml
should link to its parent catalog:
http://motherlode.ucar.edu:8080/thredds/catalogServices?catalog=http://motherlode.ucar.edu:8080/thredds/idd/dgex_model.xml
??
And the other potential link I referred to would be down at the
granule level. On our present catalog, if you go down to metadata at
the granule level, which is accessed by the links in the CatalogRef
element for each
dataset, we were suggesting a
link down at this granule level back up to the parent dataset
(metadata) page that I mentioned in the previous paragraph.
so a dataset "node" like
http://motherlode.ucar.edu:8080/thredds/catalog.html?cmd=subset&catalog=http://motherlode.ucar.edu:8080/thredds/idd/model/DGEX/Alaska_12km/catalog.xml&dataset=NCEP/DGEX/Alaska_12km/DGEX_Alaska_12km_20051221_0000.grib2
should link to its containing catalog, like:
http://motherlode.ucar.edu:8080/thredds/catalogServices?catalog=http://motherlode.ucar.edu:8080/thredds/idd/model/DGEX/Alaska_12km/catalog.xml
??
If thats what you mean, that seems reasonable, the second one is easy
to implement, but Im not sure if the first one is; catalogs dont know
who their parents are, if they exist, and the parent doesnt have to be
unique.