More GS2 OAI testing stuff

[ak19]

​http://www.openarchives.org/Register/ValidateSite is the usual (official) OAI validation site that we use.

1. Mariana Pichinini found another useful URL to validate against:

​http://oval.base-search.net/oval/validate (or ​http://oval.base-search.net/oval/oval)

After setting up GS2 and collections for OAI, plug in the URL: ​http://HOST.RESEARCH.scms.waikato.ac.nz:PORT/greenstone/cgi-bin/oaiserver.cgi

2. While there are no errors in validating 2.84.5, there is future work in this area, to make our OAI server perfect.

Oval defines WARNINGS and RECOMMENDATIONS as follows

• WARNING: Protocol violation or malformed content data
• RECOMMENDATION: Deprecated protocol feature or content-related issue

Of the warnings and recommendations resulting for a test of 2.84.5, the following are pertinent as they can't be fixed by merely tweaking a collection:

• RECOMMENDATION: Deleting strategy is "no" – recommended is persistent or transient.

• RECOMMENDATION: resumptionToken should contain expirationDate information.

• RECOMMENDATION: resumptionToken should contain completeListSize information.

• WARNING: Requests seem to be redirected to: " ​http://akeake.resnet.scms.waikato.ac.nz:8283/greenstone/cgi-bin/oaiserver.cgi"

Not sure what we can do about the following warning, since it's just the way GS3 works:

WARNING: Requests seem to be redirected to: "​http://HOST.RESEARCH.scms.waikato.ac.nz:PORT/greenstone/cgi-bin/oaiserver.cgi"

Recommendations and warnings that merely require tweaking GS2 config settings or those of a collection, or just require correct metadata to be present in the OAI-enabled collections (and which are therefore things that GS2 users can control and configure):

• RECOMMENDATION: ListRecords batch size too small (5), should be at least 100.

• RECOMMENDATION: dc:language should conform to ISO 639, found "English"

• WARNING: Records should at least contain the DC elements: date, creator, identifier, type, title. Found a record (oai:greenstone:demo:HASH9717c1cfafd71954cc6e79) missing the following DC element(s): date, type, creator.