2021-11-16 - Meeting DDI Scientific Board URN TWG

Participants

  • Carsten Thiel

  • Darren Bell

  • Hilde Orten

  • Jared Lyle

  • Joachim Wackerow

  • Wendy Thomas

Agenda

  1. Wording discussion (see CESSDA Vocabulary Service Example below)

    1. Resolution – get from a URN to a DDI instance

    2. Registration – create a URN for a DDI instance and store the URN and the connection (i.e. in the registry and its database) so that it can be resolved

  2. Use cases:

    1. Resolve a DDI URN

      1. Human found a DDI URN, needs a website to type it in to resolve

      2. Software wants to resolve a URN automatically

    2. Register a DDI URN 

      1. Build a system for publishing instances 

      2. Manually register URN and publish instance on a website

  3. Recap of open questions

    1. One URN ←→ one URL or multiple URL

      1. Is this a central decision for the entire namespace or can each agency decide?

    2. Specification of possible answers to a resolution query

      1. Could be a list of services

      2. One option: URN to URL

    3. Central or decentral

      1. Registry

      2. Resolver

  4. Planned outputs of the TWG

    1. Policy/recommendation as basis for what kind of technical services to be built

    2. Recommendation for what kind of services we provide

    3. Options for using the system built by Colectica

    4. Options we provide to agencies

      1. What do those mean for them

    5. Cost for alliance and/or agencies

  5. Next Steps

    1. Understand interest in decentralised or centralised solution

      1. Need to provide pros and cons and implications for either option

    2. Stakeholders in each case:

      1. End users

      2. Agencies

      3. DDI Alliance as managing organisation

  6. DoNM

    1. 7. Dec same time

 

Notes Nov 16

  • CESSDA Example covers the general scenarios

  • In most cases the metadata is open, the data is protected

  • In some cases even metadata can be protected, but that is not a problem of the resolution system

 

Work for this temporary working group: 

  1. Flesh out the example and summarize all implications for each option

    1. Explain which types of use might be possible or impossible in each scenario

  2. Then present to stakeholders to gauge interest

  3. Then make recommendation to Alliance which (combination) to implement

    1. Short-term vs long-term based on implementation possibilities/plan and needs

What we need to define (independent of the internal setup) first:

  • Intro: Why is identification and resolution of metadata necessary and public? [Darren+Hilde]

  • Rules

    • Technical details from the internet draft [Achim+Carsten]

    • Must also be understandable on a service level [Wendy+Jared]

    • Should include all optional rules independent of what we later recommend

  • Documentation of Rules

  • Documentation of Steps around that

  • Open Source implementation for the most simple case that can be reused by anyone

Descriptions needed in final document

  1. User point of view

  2. Admin point of view of the entire infrastructure

  3. Registration of the URNs at the basis

 

 

Example: CESSDA Vocabulary Service (CVS)

CESSDA has registered the agency int.cessda.cv within the ddi namespace.

For CESSDA the CMV uses URNs of the form urn:ddi:int.cessda.cv:TopicClassification:4.1

Distributed registry with distributed resolution:

  • This URN is only known to the CVS.

  • The DDI Agency Resolver points to CVS for resolution of the URN

Distributed registry with centralised resolution:

  • This URN is only known to the CVS

  • The DDI Agency Resolver points to the DDI URN Resolver for resolution of the URN which returns the answer after querying the CVS for the information in the background

Centralised Registry and resolution

  • The URN is stored in the DDI URN Resolver by the CVS on creation/update

  • The DDI Agency Resolver points to the DDI URN Resolver for resolution of the URN which returns the answer

It is foreseen that other entities will use the CVS. These could have and manage their URNs themselves in an arbitrary namespace or have a CESSDA-provided agency within the DDI namespace, say urn:ddi:int.cessda.eosc:Glossary:1.1, where the same as above holds.

Special case: DDI CVs in the CVS use URNs of the form urn:ddi:int.ddi.cv:DateType:1.1, these could be either managed by CVS or by the DDI Alliance, just as above.