Like all the recommendations discussed here, this is in the context of metadata for published datasets, not for datasets undergoing processing within an organisation.
Encourage data producers to consider the metadata at the same time that they are designing the data product. They should consider:
• what term might users search for that could form part of the dataset Title? This in turn may influence how a larger set of data could be partitioned thematically so that users more easily find the data they are looking for;
• in what way will they partition the data: a series with datasets, or a dataset with tiles, etc.; and
• that they need to avoid mixing in the same dataset data that is only available under one set of (licensing) conditions with data that is available under another set of conditions.
• Under what conditions does the status of the data change? What will trigger a new release of the data (with an updated Dataset reference date)? What should then happen to the old “version” of the data, and does this mean a new metadata record – for the new version, or the old one?
• What is the relationship for this dataset between Temporal extent, Dataset reference date and Maintenance information?
Some of these decisions might only affect the data and metadata managed internally within the data publisher (or aggregator). Other organisations may make all intermediate stages available to the public (“publish them”).
See https://www.agi.org.uk/agi-groups/standards-committee/uk-gemini/40-gemini/1051-uk-gemini-v2-2-specification-for-discovery-metadata-for-geospatial-resources#4.1 where this information could perhaps reside.
Also, where the licence under which the resource is available changes, it may sometimes be best to issue a new metadata record. This is a decision for the publishing organisation. That implies that a change of licence which is known in advance, for example the end of an embargo period, requires clearer guidance – effectively a discussion of the lifecycle of the metadata record.
Like all the recommendations discussed here, this is in the context of metadata for published datasets, not for datasets undergoing processing within an organisation. Encourage data producers to consider the metadata at the same time that they are designing the data product. They should consider: • what term might users search for that could form part of the dataset Title? This in turn may influence how a larger set of data could be partitioned thematically so that users more easily find the data they are looking for; • in what way will they partition the data: a series with datasets, or a dataset with tiles, etc.; and • that they need to avoid mixing in the same dataset data that is only available under one set of (licensing) conditions with data that is available under another set of conditions. • Under what conditions does the status of the data change? What will trigger a new release of the data (with an updated Dataset reference date)? What should then happen to the old “version” of the data, and does this mean a new metadata record – for the new version, or the old one? • What is the relationship for this dataset between Temporal extent, Dataset reference date and Maintenance information? Some of these decisions might only affect the data and metadata managed internally within the data publisher (or aggregator). Other organisations may make all intermediate stages available to the public (“publish them”). See https://www.agi.org.uk/agi-groups/standards-committee/uk-gemini/40-gemini/1051-uk-gemini-v2-2-specification-for-discovery-metadata-for-geospatial-resources#4.1 where this information could perhaps reside. Also, where the licence under which the resource is available changes, it may sometimes be best to issue a new metadata record. This is a decision for the publishing organisation. That implies that a change of licence which is known in advance, for example the end of an embargo period, requires clearer guidance – effectively a discussion of the lifecycle of the metadata record.