Views
Comments by two participants sum up the desire of most, if not all, participants to determine conclusively what to do or not do with the OKI OSIDs? and not merely iterate past attempts to use them or evaluate their usefulness.
1. Chris Mackie expressed Mellon’s need to know definitively whether to recommend or require, or merely advocate that grantees for Mellon projects use the OSIDs? to increase interoperability.
2. Brad Wheeler described the “seven years of pain” he has experienced from several attempts to use the OSIDs? and expressed a desire to avoid yet another cycle that fails to yield a definite conclusion.
From the discussion in the meeting, it appeared that participants agree the Repository OSID has significant adoption and implementation and that this OSID is an effective approach to achieving many to many interoperability involving repository access. Requiring projects to use the Repository OSID, where it applies, appears to be a good strategy for Mellon. “Official” approval by Mellon might contribute to voluntary adoption and use of this solution, as well as facilitate the consolidation of support that the OKI Consortium hopes to achieve. Further work in the general Repository area could yield additional benefits. However, it should be noted that Mellon already has seeded this area.
Recommendation: The Repository OSID has succeeded in providing interoperability. Mellon’s investment in products and approaches in this area has produced initial successes. Direct action by Mellon is not likely to materially influence the market forces that are driving further adoption of the Repository OSID and is, therefore, not advisable. The value of indirect actions is noted above.
From the discussion in the meeting, it appeared that participants agree the Grade Reporting area constitutes a relatively well-understood and bounded interoperability problem for which a focused, near term effort has the potential for resulting in a standard that enables interoperability. While it is somewhat mundane technically, this area looks like a good test for an OSID and any other approach. Collaborative work in this area would exercise the governance and process elements of the nascent O.K.I. Consortium and test its ability to organize and facilitate community effort involving participants from Mellon projects and the community at large.
Recommendation: A strategic push through Mellon funding and advocacy might yield convergence on a general model and interoperable implementations of grade reporting services (i.e., a complete solution in the Digital Marketplace environment using an OSID and Sakai in combination). That outcome would dramatically increase interoperability in practice and would be widely adopted in the field.
The discussion of identity management seemed more inchoate on several dimensions than that for repositories or grading. Perhaps this indicates it is a place for Mellon to provide incremental investment to test the utility of the AuthN? and AuthZ? OSIDs?. On the other hand, the influence of market forces and activity outside the Mellon community, and the work that already will be undertaken in the DM and KS projects, are likely to dominate improvements in interoperability in this area. Recommendation: Unless a specific target problem that will not be addressed or overwhelmed by other efforts can be defined, Mellon policy or activity in this area will not have much effect on either identity management or interoperability.
The discussion among participants indicated that reusing workflow engines and providing substitutability are valid interoperability concerns. However, this area appears to be different in kind from repositories, grading, and identity management in that it is not tied to specific functionality. Multiple existing approaches (models and engines) exist. But a single, high level model does not. The OSID in this area is untested.
Recommendation: The risk is substantial that activity by Mellon will degenerate towards open loop refinement of either an approach (e.g., the workflow OSID) or an implementation (e.g., the KEWE engine), rather than resulting in improved interoperability. Unless the activity can be tied to concrete outcomes and specific metrics, it should not be undertaken.
For understandable reasons, participants in the meeting made position statements or logically possible, tentative, or untestable assertions. These are safer than evidence-based comparisons or assertions that can be validated. But they have little, if any, value for decision-making. There appeared to be no opposition to using the kind of criteria and metrics that were identified in the pre-meeting report and used to guide discussion during the meeting. However, other than characterizing outcomes as “seven years of pain”, or asserting that an OSID is “widely adopted” or “not in production”, the participants did not offer metrics by which the effect of an approach to interoperability or the benefit or cost of achieving interoperability can be assessed.
Recommendation: Mellon’s goal is interoperability, rather than the success of a particular means for accomplishing interoperability. Investments tied to consolidating criteria and metrics for interoperability, using them to measure interim progress and project outcomes, and awards or incremental funding contingent on performance will influence interoperability more than top down mandates or direct funding of a particular approach or a particular project.
1. Chris Mackie expressed Mellon’s need to know definitively whether to recommend or require, or merely advocate that grantees for Mellon projects use the OSIDs? to increase interoperability.
2. Brad Wheeler described the “seven years of pain” he has experienced from several attempts to use the OSIDs? and expressed a desire to avoid yet another cycle that fails to yield a definite conclusion.
From the discussion in the meeting, it appeared that participants agree the Repository OSID has significant adoption and implementation and that this OSID is an effective approach to achieving many to many interoperability involving repository access. Requiring projects to use the Repository OSID, where it applies, appears to be a good strategy for Mellon. “Official” approval by Mellon might contribute to voluntary adoption and use of this solution, as well as facilitate the consolidation of support that the OKI Consortium hopes to achieve. Further work in the general Repository area could yield additional benefits. However, it should be noted that Mellon already has seeded this area.
Recommendation: The Repository OSID has succeeded in providing interoperability. Mellon’s investment in products and approaches in this area has produced initial successes. Direct action by Mellon is not likely to materially influence the market forces that are driving further adoption of the Repository OSID and is, therefore, not advisable. The value of indirect actions is noted above.
From the discussion in the meeting, it appeared that participants agree the Grade Reporting area constitutes a relatively well-understood and bounded interoperability problem for which a focused, near term effort has the potential for resulting in a standard that enables interoperability. While it is somewhat mundane technically, this area looks like a good test for an OSID and any other approach. Collaborative work in this area would exercise the governance and process elements of the nascent O.K.I. Consortium and test its ability to organize and facilitate community effort involving participants from Mellon projects and the community at large.
Recommendation: A strategic push through Mellon funding and advocacy might yield convergence on a general model and interoperable implementations of grade reporting services (i.e., a complete solution in the Digital Marketplace environment using an OSID and Sakai in combination). That outcome would dramatically increase interoperability in practice and would be widely adopted in the field.
The discussion of identity management seemed more inchoate on several dimensions than that for repositories or grading. Perhaps this indicates it is a place for Mellon to provide incremental investment to test the utility of the AuthN? and AuthZ? OSIDs?. On the other hand, the influence of market forces and activity outside the Mellon community, and the work that already will be undertaken in the DM and KS projects, are likely to dominate improvements in interoperability in this area. Recommendation: Unless a specific target problem that will not be addressed or overwhelmed by other efforts can be defined, Mellon policy or activity in this area will not have much effect on either identity management or interoperability.
The discussion among participants indicated that reusing workflow engines and providing substitutability are valid interoperability concerns. However, this area appears to be different in kind from repositories, grading, and identity management in that it is not tied to specific functionality. Multiple existing approaches (models and engines) exist. But a single, high level model does not. The OSID in this area is untested.
Recommendation: The risk is substantial that activity by Mellon will degenerate towards open loop refinement of either an approach (e.g., the workflow OSID) or an implementation (e.g., the KEWE engine), rather than resulting in improved interoperability. Unless the activity can be tied to concrete outcomes and specific metrics, it should not be undertaken.
For understandable reasons, participants in the meeting made position statements or logically possible, tentative, or untestable assertions. These are safer than evidence-based comparisons or assertions that can be validated. But they have little, if any, value for decision-making. There appeared to be no opposition to using the kind of criteria and metrics that were identified in the pre-meeting report and used to guide discussion during the meeting. However, other than characterizing outcomes as “seven years of pain”, or asserting that an OSID is “widely adopted” or “not in production”, the participants did not offer metrics by which the effect of an approach to interoperability or the benefit or cost of achieving interoperability can be assessed.
Recommendation: Mellon’s goal is interoperability, rather than the success of a particular means for accomplishing interoperability. Investments tied to consolidating criteria and metrics for interoperability, using them to measure interim progress and project outcomes, and awards or incremental funding contingent on performance will influence interoperability more than top down mandates or direct funding of a particular approach or a particular project.