The number of records are not the same.
So my question is the following:
Should the oer set of the OAI-PMH automatically be the same as the frontend category or are the set in phaidra instances individually configured?
Thanks.
Tobias
PS:
If any Phaidra is missing and we got no ticket for indexing it into OERSI yet feel free to contact me.
first of all, welcome to our community and thank you for reaching out to us!
The number of records could indeed differ. I check the two cases and the reasons are different though:
in case of fhstp, it seems that there is a bug in OAI update script which prevents successful update. I’ll fix it and will let you know.
in case of univie, there are 2 objects (o:2057819 and o:2129210) which have been deleted. Such objects do not appear in the search anymore, but are still in OAI provider to pass the information about deletion to harvesters. You can see that in <header status="deleted">, but you probably know that.
I’m not 100% sure if you currently have tickets for all instances which have OER set set up, but we can check. @Narloth do you have a list of PHAIDRA instances integrated with OERHub?
One additional thing:
It seems that in the repositories you make the distinction between learning object/educational resource and open educational resource. Since I thought that Phaidra is an open access repository kind of thing I thought that all learning objects/educational resources would be oer in Phaidra.
Could you specify the difference between them in Phaidra context?
PS: Nice Discourse you have here. hbz also started one for metadata related topics for the DACH-region: https://metadaten.community/
PHAIDRA is best suited for open access material but, apart from the fact that we also accept material with restricted access, some people see material as open access simply if you can access it, others only if it has any of the licenses they find open. That is also the case in OER.
The “learning object” was an object type in PHAIDRA long before OER was a thing around here, so once the OER people came they found the OER objects should be marked specifically as OER because not all our learning objects, regardless of access restrictions, are OER by their(?) definition. I think that makes sense, although it might seem a bit confusing in the facets…
I forgot to mention that in KPH, the numbers can also differ but for a different reason: in PHAIDRA one can create and OER by using the OER submit form which then only allows open licenses and is additionaly using the hcrt vocabulary for object types which then makes the objects easily compatible with Austria’s OERHub. However, one can also simply go to a normal submit form, mark an object as learning object and use an open license. In univie, such object will still be part of oer set in oai-phm, but in KPH not, because they only wanted object submitted through their (slightly modified) OER submit to be part of oer set.
Also @rasta you wrote that you are providing OERHub with hcrt resource types these seem not be available via oai_openaire is there a way to fetch them via oai pmh?
I am checking out the LOM metadata via OAI-PMH it seems that LOM is missing information that is provided via oai-open-aire.g.:
publication dates (<datacite:date dateType="Issued"> is not the same as <lom:lifecycle><lom:datetime> compare: oai:phaidra.fh-kaernten.at:o:15)
In the LOM Datal it seems that there is no possibility to differentiate between an Organization and a Person with regard to contributors in the VCard 3.0 that is only possible with VCard 4.0. Compare: oai:phaidra.fh-kaernten.at:o:45 (openair allow this in <datacite:creatorName nameType="Organizational" >Carinthia Nature Tech</datacite:creatorName>
File links are not given but could be constructed from the given data. openair give this in: <file mimeType="application/pdf" objectType="fulltext" >https://phaidra.fh-kaernten.at/api/objects/o:64/download</file>
Relations between records are not given in LOM.Compare: oai:phaidra.fh-kaernten.at:o:45 openair allows this with <datacite:relatedIdentifiers> <datacite:relatedIdentifier relatedIdentifierType="URL" relationType="IsPartOf" >https://phaidra.fh-kaernten.at/o:25</datacite:relatedIdentifier> </datacite:relatedIdentifiers>
Currently we use the OAI_Openair for the metadata since it has more info and I additionally harvest LOM but only for the HCRT and locally enrich the openair data with it.
I am not sure where to add this info into open_air but it would be great if we would not have to harvest the oai-pmh twice each time. But currently our set up seems to work,