Re: [thredds] Collection Metadata from Granules

Simon.Pigot@xxxxxxxx wrote:

Re: granularity

I don't have an answer on what level of granularity is appropriate - only to 
mention that the harvester will have simple check boxes that allow you to 
harvest services, collections and atomic dataset metadata - its a device to 
which other options could be added. The whole concept of searching/identifying 
and displaying hierarchical relations between dataset metadata records as well 
as relationships between services records and the datasets they deliver is 
still at a very early stage of development (at least in GeoNetwork :-)).
We use the harvest="true" attribute to identify those THREDDS dataset elements for which we want to create these kind of metadata records. This is a mechanism to make sure you dont have to open 1000's of identical files.