We welcome content providers who can make metadata available to DigitalNZ using OAI-PMH, the Open Archives Initiative Protocol for Metadata Harvesting.

The way OAI harvesting works is like a duet between two partners:

• data providers, who provide metadata as an XML document following an agreed specification, and
• service providers (such as DigitalNZ) who harvest metadata from data providers and use it to provide useful services.

If your data repository supports OAI-PMH you can use the protocol to contribute metadata to DigitalNZ. We also have other harvesting options.

Contributing to DigitalNZ will be simple if you already provide metadata to an OAI service provider. All you need to do is let us know the base URL for your OAI feed.

Take a tutorial to find out more about OAI-PMH

Becoming an OAI-PMH data provider

To set up OAI as a data provider you need:

• Metadata on content, stored in a database (or ‘repository’). Each content item must have a unique identifier.
• Repository software with OAI-PMH enabled (or the resource to develop it).
• To configure the repository to supply the metadata that DigitalNZ needs.

Some examples of software with OAI-PMH built in are:

• Kete
• DSpace
• Eprints
• Fedora

If you don’t already have OAI-PMH you will need to discuss the feasibility of becoming an OAI-PMH data provider with your software vendor or technicians.

Configuring the OAI-PMH repository

Configuring your OAI-PMH repository can be difficult, though most software has sensible defaults. Talk to your vendor or technicians if you are unsure.

DigitalNZ currently harvests OAI-PMH using the required oai_dc format, which allows you to transfer metadata using the 15 original unqualified or Simple Dublin Core elements.

Read descriptions of Simple Dublin Core elements

Dublin Core conventions for DigitalNZ

Simple (unqualified) Dublin Core is enough to get you started in DigitalNZ. We ask that you also follow a few conventions:

• Every record must include a URL that users can follow to get to the resource from a search result page. This should be stored in dc:identifier.
• Every record must include a title in the dc:title field.
• Every record should have dc:rights metadata, and we strongly encourage you to encode this as a URL.
• To really work well in DigitalNZ search, records should have dc:description, dc:date, dc:creator, dc:coverage and dc:subject metadata if at all possible.

Find out more about using Dublin Core

Make it Digital guide to describing content

Links to thumbnails and digital resources

There are a few more conventions that we ask you to follow in order to make your metadata really useful in DigitalNZ:

• Every record that describes a digital resource should include the URL of a high-resolution version of the digital resource in the dc:source field.
• Every record should include a link to a thumbnail, and every image must include a link to a thumbnail.

Unfortunately there’s no obvious place to store thumbnail metadata in Simple Dublin Core. Our harvester is flexible and can handle several different conventions. For example, other content providers have:

• put a thumbnail URL in dc:relation, or
• put a thumbnail URL in other fields but ending with “thumb.jpg”, or
• provided a pattern that can be used to generate a thumbnail URL from a URL in dc:identifier or dc:source.

Need help?

Don’t forget that advice is available to you from DigitalNZ if you need help configuring your OAI-PMH feed. If it is necessary our OAI-PMH harvester can help tidy up your metadata by transforming values, adding and removing fields, and even filtering out records that you don’t want us to expose.

Just get in touch if you have any questions.

What happens next?

When we know the base URL for your metadata feed, we will go in and do a test harvest. After ironing out any issues with that first harvest, we’ll continue to check for changes each night at 2 AM (unless you let us know otherwise). You don’t need to do anything!

Remember this is just one of the available options. If you're not sure whether this is right for you, drop us a line and we can talk about it.