Feature/opends 0.4.0

[samleeflang]

PR with the new version of openDS based on the feedback we got during the Public Review.

• Heavier use of 'ods:hasAgents
• Remove ods:ID for an existing term (schema or dcterms)
• More explicit and extensive descriptions
• Split object into separate files
• Change naming convention of object (single nested object hasXXX multiple nested objects hasXXXs
• renamed ods:MaterialEntity -> ods:SpecimenPart
• A lot of other small changes.

Please focus on the schemas, the examples haven't all been validated yet as the schema's I need the schemas online for that.

[wouteraddink]

one more thing: with ISO 639-2 it would be better to use the terminological codes, so ISO 639-2/T and not ISO 639-2/B (bibliographic) variant in cases where there is a variant, which is mainly there for legacy bibliographic catalog purposes. So the code for Dutch should be "nld" not "dut" for example. The first one is ISO 639-2/T and the second one is ISO 639-2/B.

[samleeflang]

one more thing: with ISO 639-2 it would be better to use the terminological codes, so ISO 639-2/T and not ISO 639-2/B (bibliographic) variant in cases where there is a variant, which is mainly there for legacy bibliographic catalog purposes. So the code for Dutch should be "nld" not "dut" for example. The first one is ISO 639-2/T and the second one is ISO 639-2/B.

I tried to keep in line with Audoban Core with this, which just accepts any ISO 639-2 standard and points to a controlled vocab which puts Dutch -> dut. We could deviate, but I am unsure if we can then still use ac:metadataLanguage and ac:metadataLanguageLiteral. We could also scrap the metadataLanguages as we already has a discussion about it usefullness. We would then keep only dcterms:language which we could limit to ISO 639-2/T.