Data search by parameter name (CF metadata convention vocabulary)

[rico-hengst]

A user wants to search for datasets by parameter name. The user expects standardisized naming convention (CF metadata convetion vocabulary) of the parameters to avoid duplicate search events.

[bjoernbroetz]

There could be a search on cf-names, so you type "ozone" and you get the 5 cf names with "ozone" in the name.

That's maybe the place to talk about ontologies and controlled vocabularies. The user might want to type "O3" and would still like to find the ozone related datasets.

[joerg-halo]

Notes from Zoom-session on 22 February 2021

• We recommend a rather strict application of the CF-conventions for variable naming
• Ontologies might appear to be a nice solution for providing more comfort to data providers. However, there are bad experiences with the upkeep of the ontologies from the MSS-tool. Furthermore, also the use of ontologies will not render the use of every variable naming possible. There have to be reglementation anyway. Thus, we do not want to use ontologies for datasets which are uploaded nowadays.
• The HALO database has already the capability to recognize if uploaded data comply with required data format or naming conventions. However, right now, the output of this check is only a message if the uploaded data do not comply. Data are accepted anyway.
• In the past, there was not a too strict enforcement of the CF-convention in order to achieve a higher acceptance of the HALO database. However, we think that nowadays the hurdle of enforced CF-conventions would be acceptable to the data providers. First of all, times have changed and the need to comply with naming conventions is more commonly accepted. Furthermore, nowadays the HALO database is the accepted data repository for HALO data. This increases the acceptance of rules enforced by the HALO database.
• How to deal with data which was uploaded to the database in the past, without compliance to CF-conventions? A script might be deployed to revise data which are already uploaded to the database but do not comply with CF-conventions. However, the preferred solution is to use ontologies for this scenario.
• There might be cases, where the CF-conventions do not fulfill all requirements of users of the HALO database. We should be open for other requests in this regard. The global change master directory might also be helpful in this regard. Futhermore, it is wothwhile to establish a direct contact to the team behind the CF-conventions.