Multiple doc types

[kpsherva]

We discussed “unmerging” records and linking them via related_identifiers field would be more proper from the point of view of cataloguing rules and how DataCite is handling the records.
From the technical aspect of handling future record updates this solution is also preferable.

Our alternative was to leave the records merged and select the leading resource type. This solution was proposed as a mitigation to users searching for the most prominent resource type – journal article.
We discussed both together and later internally and we might have appropriate solutions for this problem in case we go with unmerged records.
From the point of view of DataCite this solution is not viable and it brings a lot of complications while updating record versions and files.

We agreed on exploring the needed components for implementation of umerging approach, namely:

1. SIS Team to provide list of “source” fields to let us distinguish which value belongs to which resource type while unmerging
2. SIS Team to provide: list of possible updates operations
3. SIS Team to answer: is there a way of distinguishing which file belongs to which resource type, if yes, how?
4. SIS Team to consider in which cases of updates operations we need to consider creating a new version of the record.

Clarification from CDS: a new version of record is created when file of this record was updated to a new version (something has changed, f.e. typo in the PDF, or PDF received a review, author applied feedback, and this new version of file is now uploaded to the record). The resource type should not change between record versions – if resource type is changed, it implies that new separate record should be created.

5. Inspire team to expose “hidden” arXiv files to CDS via REST API – what would be your rough estimation of delivering this change on Inspire side?
6. Inspire team to expose accelerator and experiment detailed data on the record directly, similarly as done for authors affiliations – same question as above

Then we have a list of questions to answer regarding the files (that we didn’t cover during the meeting):

Which fields in the record schema can contain file references, either internally hosted or external (e.g., documents, arxiv_eprints, others)?

Is there a way to determine all available files for a record, regardless of whether they’re hosted internally or externally?

Is there a recommended way to detect if a record has no internally hosted files but does reference an external file?

Do you provide license or access information (e.g., open access, embargoed) for externally referenced files?
Are there cases where the documents field references files that are not downloadable (e.g., due to embargoes or expired links)?
Since our use case includes long-term archiving, is it acceptable to download referenced files, or are there any copyright/licensing concerns we should be aware of (when coming from INSPIRE)?

[kpsherva]

Meeting notes:

• TODO (Karolina) check how many files we could distinguish preprint
• could we have multiple journals and conferences ?

[salomerohr]

Sara will send the Excel file listing all the fields and possible rules and issues today.