Clarifying the usage of ds/cdataset

[jypeter]

I have been spending some time with dataset and explore to check the content of the IPSL mirror, reading the documentation of cdataset : define a dataset object and some of the source code many times

I think there is something missing or ambiguous about this, but I can't completely explain it. Or maybe I have just been too tired lately :-(

I think I have trouble putting my mind around the dataset object, and remembering the documentation, because the dataset is at the same:

• used to create a query of what we are looking for
• used to get the result of the query by using its explore method

And this is not really what the description says (btw, there must be a word missing in the first sentence): A CLIMAF dataset is a description of what the data (rather than the data itself or a file). It is basically a set of pairs attribute-value. The list of attributes actually used to describe a dataset is defined by the project it refers to.

I wonder if it would not be easier to document this in terms of a search/query with facets. Using this would also match what people do when they select facets in the search interface of an ESGF node

Another problem I have is that I'm not sure when the actual globbing search is performed. When looking at the response time of the commands (in the interpreter), I have a feeling that creating the ds object only initializes it (e.g. we create the query), and that the actual query is executed when we use explore. But that's not always the case, so maybe it depends on how many wildcards there are in the query?

Hmmm, I hope what I wrote makes sense

[senesis]

Thanks Jean-Yves for your tireless contributions to improving CliMAF, and its doc. You are quite right here. We will work on that subject, and welcome further suggestions.