Page History

Test ARK proposal #1 (May 11 brought to NAAN WG)

N2T-based test ARKs for non-EZID users

This is a much narrower version of the proposal brought in last month, which was to expand the role of NAAN registry to record shoulders under the four non-institutional NAANs. We need agreement on how INCIPIT might proceed with N2T-based test ARKs that do not conflict with EZID test ARKs. A major benefit of the proposal would be to protect exclusive "test" shoulder access rights of both EZID and INCIPIT.

Proposed: With its oversight role for the NAAN namespace, this group agrees to reserve, under the 99999 NAAN,

• 1 shoulder for exclusive use of the INCIPIT service
• 1-3 "historical" shoulders for exclusive use of the EZID service (uses for testing longer than 2 weeks)

Some discussion:

1. A 99999 NAAN in ARKs found in the wild signifies a status of "test".
2. No organization owns or was meant to own the 99999 NAAN.
3. Historically, EZID has used 2 shoulders under the 99999 NAAN, taking advantage of these portions of the public/common space as if it had exclusive rights to them. This proposal would formally recognize those rights in the future.
4. This proposal would also grant similar rights to a test shoulder for INCIPIT.

The current proposal has no impact on the NAAN registry. The test shoulders reserved in this way would be recorded for now in a separate file. 

Test ARK proposal #2 (this might co-exist with or go forward instead of proposal #1)

In the May 11 NAAN WG discussion, Bertrand reminded us of Roxanna's idea (2020.04.20) of a convention that automatically and implicitly reserves a test shoulder under 99999 whenever a new NAAN, eg 12345, is created.

Summary proposal: If ark:/12345/foo is registered to redirect to X/foo, then ark:/99999/912345_bar will redirect to X/912345_bar. In general, ark:/99999/9<NAAN>_... will redirect to X/9<NAAN>_....

Why surround the NAAN with a 9 on the left and '_' on the right? Because (a) the 9 preserves the first digit convention for the shoulder (with a digit that is fixed) and the '_' marks the end of the NAAN in case NAANs are assigned that have more than 5 characters. Question: will this interfere with check digit calculation?