Practical Policy Recommendations (Endorsed)

 

Practical policy Working Group

Recommendation Title: Machine Actionable Policy Templates

Impact: Can be used to enforce management, automate administrative tasks, validate assessment criteria, and automate scientific analyses

Recommendation package DOI: http://dx.doi.org/10.15497/83E1B3F9-7E17-484A-A466-B3E5775121CC

Co-Chairs:

Reagan Moore, RENCI

Rainer Stotzka, Karlsruhe Institute of Technology

 

Computer actionable policies are used to enforce management, automate administrative tasks, validate assessment criteria, and automate scientific analyses.  The benefits of using policies include minimization of the amount of labor needed to manage a collection, the ability to publish to the users the rules that are being used, and the ability to automate process management.

Currently all sites and scientific communities use their own set of policies, if any. A generic set of policies that can be revised and adapted by user communities and site managers who need to build up their own data collection in a trusted environment does not exist.

What are the Goals?

The goals of the working group therefore are

  • To bring together practitioners in policy making and policy implementation
  • To identify typical application scenarios for policies such as replication, preservation etc.
  • To collect and to register practical policies
  • To enable sharing, revising, adapting, and re-using of computer actionable policies

Please download the full Practical Policy (PP) WG recommendation package below, as individual sections. 

Please use the comment function below for questions and suggestions. Please note that you need to login in order to comment. 

 

 

Group content visibility: 
Use group defaults
  • Elizabeth Griffin's picture

    Author: Elizabeth Griffin

    Date: 26 Mar, 2015

    What you say might work in a simple, strightforward, undeveloped environment, but the actual state of affairs is probably a tad more complicated than that.

    The stated goals are
    (a) To bring together practitioners in policy making and policy implementation
    (b) To identify typical application scenarios for policies such as replication, preservation etc.
    (c) To collect and to register practical policies
    (d) To enable sharing, revising, adapting, and re-using of computer actionable policies.

    Now data producers (observers, research practitioners, lab workers, etc.) will, from sheer necessity of research organization, already have their own forms of data management. If there is no identifiable domain repository and handling system, then it will simply be something that works for them; and if they are not part of a grand collaboration with outside partners, then what they have will be adequate for them, so they will not cheerfully embrace a request to change 'for the common good', or certainly not unless someone comes along and does it for them. Researchers usually have objectives and date targets to meet, and their funding agencies do not expect them to use their time formatting data just so that others can use them.

    If I take Astronomy as an example, an international scheme for data deposit (largely signalled by space missions but by multi-national ground-based ones too) is well established because the number of astronomers in research is relatively manageable, and because people have been working on these very problems, for the sake of sharing data, ever since astronomical data became born-digital, so the basic concepts were built in from the get-go. That is why it has been possible to build such a thing as a Virtual Observatory, whereby data sets can be made fully interoperable, and even retrofitted to that standard. But take almost any other domain, where data-survival has been more common than data management, and where the numbers of people involved are larger. Individualism will be rife, and trying to supplant it with a unified all-singing all-dancing system is going to take several revolutions and
    coups. Who is going to do all that? The researcher's remit will not have changed, and there will be no fresh funding to pay for the necessary reformatting, etc. Each lab experiment will probably need a tame data-science expert to assist, and I am not sure that I can see what the justification for that extra funding will look like.

    Elizabeth Griffin

  • Reagan Moore's picture

    Author: Reagan Moore

    Date: 27 Apr, 2015

    The comment discusses the concept that data producers already have internal data management systems and have no incentive to share data.  The implication is that the objective of the Research Data Alliance to promote international research collaborations is not feasible.

    The response is that multiple communities are forming national and
    international collaborations that require support for federating
    existing data repositories.  Existing collaborations include:
    - EUDAT
    - International Neuroinformatics Coordinating Facility
    - the iPlant Collaborative
    - DataNet Federation Consortium
    - French National Institute for Nuclear Physics and Particle Physics
    Computer Center
     
    The incentives to change to policy-based systems come from requirements for scalability (the current system is unable to manage the load), requirements for distributed data management (the data sources and collaborators are distributed), and requirements for interoperability (the need to be able to migrate to more cost effective data management technologies).
     
    Policy based system provide a way to federate existing data repositories to support research collaborations, automate enforcement of management decisions, automate administrative tasks, and automate validation of assessment criteria.  The examples presented in the Practical Policy documents come from three production environments, and represent the state-of-the-art in distributed data management.
     
    Reagan Moore

  • Jill Kowalchuk's picture

    Author: Jill Kowalchuk

    Date: 29 Apr, 2015

    My comments are not related to the actual work done, but instead related to the motivation and the listed benefits or goals. I reviewed the documentation through the lens of what organizations/indivdiuals do I know who are not currently members of RDA or engaged in RDA efforts who might be able to benefit from this work/output.

    My suggestion would be some additional wording be added to the summary (above) to highlight the benefits or more specifically "why this might be useful to groups" - it might also be useful to take it one step further to suggest some types of organizations which might find this work useful.

    I actually found much of what I was looking for in terms of benefits in Reagan's comments above so that could be reused/tailed quite easily/effectively. I would agree with Elizabeth's comments in that using policies such as what is noted above from the perpsective of an individual researcher or research group could be quite challenging. However, within my jurisdiction I am seeing more and more of these researchers relying on data service providers to provide the necessary infrastructure and services to support data management. The more of these service providers that implement comment policies such as what is stated in this outcome could help drive scalability, interoperability (and the other benefits suggested by Reagan).

     

    Cheers,

    Jill

  • Jill Kowalchuk's picture

    Author: Jill Kowalchuk

    Date: 29 Apr, 2015

    My apologies as my previous comment might not be 100% relevant (sorry I am new to navigating the RDA website/wiki). I noticed when I navigated back to the working group page some of what I was looking for (benefits) was included on that introductory page. I came to the recommendations page through the email link Herman sent in February. It could still be useful to add a few sentences to this page related to benefits to make it easier to identify. 

    Cheers,
    Jill

submit a comment