data access for virtual collections

[ferrighi]

Some mmd records can express a virtual collection, i.e. a collection of files or directories on a thredds server. The virtual collection does not have a data access represented by the vocabulary currently in use (http, opendap, etc). Currently such records use the related_information "Dataset landing page" to connect to the data. In vision of disconnecting landing pages from thredds, I would suggest we add a data_access type as for example "thredds catalogue" or more generically "virtual collection".

As for today the HTTP definition is "Direct access to the full data file. May require authentication, but should point directly to the data file or a catalogue containing the data." So in principle we could use HTTP for data access, but it creates ambiguity between a direct download and a resolving url.

[ferrighi]

This is covered now by the design of the landing pages which include a link to the children. I still think it would be useful to have a url pointing to the folder of the whole collection on thredds, but I close this for now. We can reopen if necessary.

[ferrighi]

I would like to push for this now, so we can proceed with harmonisation of information in our systems. My proposition is:

• Addition of term "THREDDS" into the Access Type list
• Proposed mapping:

MMD	DIF9	DIF10	OSGEO
THREDDS	GET DATA, THREDDS DATA	USE SERVICE API, THREDDS DATA	UNIDATA:THREDDS

[ferrighi]

@mortenwh do you have any comment on this? I'm ready with a PR.

[mortenwh]

Not really - have you created the PR?

[ferrighi]

After our discussion to move this field from data_access to related_information (where the type Dataset landing page already exists), a new proposition would be the following:

related_information/type = "Data server landing page"

The url will point to any application server landing page, e.g. thredds, hyrax and erddap.

Possible translation to dif/iso will be a bit more tricky. But for DIF:

USE SERVICE API, THREDDS DATA: "The URL for access to the Thematic Real-time Environmental Distributed Data Services (THREDDS). The THREDDS Data Server (TDS) is a web server that provides metadata and data access for scientific datasets, using OPeNDAP, OGC WMS and WCS, HTTP, and other remote data access protocols."

I've checked some examples in DIF10, and it seems this field refers to the .html pages of thredds, so in case there is a URL from thredds we could map it to this field.

For ISO/OSGEO:

UNIDATA:THREDDS : Link is a THREDDS Catalog URL that provides the XML for traversing programatically.

so it is not the .html, but .xml. In this case I would skip this translation.

[ferrighi]

@steingod would this be more consistent?