Summary of Comments Received on MSC Principles for Evaluating Metadata Standards
The ALCTS/LITA Metadata Standards Committee posted a second draft of the Draft Principles for Evaluating Metadata Standards on October 27, 2015, on the metaware.buzz site. This second draft incorporated feedback received on the first draft. The committee held a second open comment period from the time this draft was posted through the ALA Midwinter 2016 meeting in Boston. We received comments directly on the post itself, on a web form set up for this purpose, from the audience following a presentation on the Principles at an ALA Midwinter 2016 session sponsored by the ALCTS Metadata Interest Group, and at the committee’s in person meeting at ALA Midwinter 2016.
This document summarizes the feedback received on the second draft in those venues.
We received a number of general comments on the document and its approach. Several commenters requested more detailed information on why the Principles exist and what use cases they are intended to fulfill. We will write an introduction to the Principles document that addresses these and related issues. This introduction will define the scope of the Principles to be the standards and how standards construction impacts metadata being created, rather than best practices for creation of metadata itself, recognize that the information in the Principles is a distillation for easy consumption of lessons learned by many metadata communities over time; and acknowledge that some of the ideas in the Principles document are difficult to achieve, or that indeed methods for achieving them are still being defined by various metadata communities. The Principles as such are in some sense aspirational.
We received requests for clearer definitions for some terminology such as “add value” and “network”, which we will achieve through linking to an established, externally-maintained glossary, and for including examples in the document, which we will endeavor to do to illustrate application of individual principles wherever possible. We will also ensure that the document adequately covers content standards, controlled vocabularies, and system- or machine-created metadata.
Comments regarding “metadata standards should be part of the network,” centered around the concept of “by reference” models. We made an embarrassing error in language that caused this principle to originally say the opposite of what we intended, and have corrected this. We will enhance the text of this principle to focus on standards providing context and relationships to other vocabularies as a core expression of network-connected metadata, and cite the “by reference” model as exemplifying this principle. One commenter noted that occasionally implementations might choose to store both “strings” and “things”, which we will address through language in the introduction covering pragmatic considerations taking over when circumstances warrant.
We received a number of comments on the concept of openness in this principle, “Metadata and metadata standards should be open and reusable.” We will clarify that openness here refers to the metadata and not the resource, that explicit licenses for metadata promote reusability, and that openness is tied to governance and maintenance community actions. As we are clarifying that metadata creation practices are out of scope for this document, we will further edit this principle to reflect that, and remove the recommendation that all metadata itself should be open. While committee members believe strongly in the utility of open metadata, we do agree with a commenter that the focus of this document is elsewhere.
We received one comment on the principle “Metadata creation should benefit user communities,” requesting clarity of its meaning. We will enhance this text to indicate that depth of metadata defined by a standard should directly fulfill a real world use case, and ensure that this principle and its heading focus on the role of the standard and not the practice of metadata creation.
Commenters on the principle “Metadata standards should support new research methods” requested greater precision and explanation of how this would occur, and to clarify that metadata can be reused for other purposes as well. We will expand this principle to focus on flexibility of metadata and the ability to do interesting things with it, with new research methods as one example. In addition, as suggested by a commenter, we will mention formal expression of metadata models through languages such as RDF Schema, RelaxNG, or XML schema as an example of making metadata more machine-processable.
One commenter on “Metadata standards should have an active maintenance and governance community” pointed out that it’s a stretch to claim that metadata standards in LAM institutions have been stable for 40 years. We agree, and will adjust the text to focus more tightly on the increasing pace of change.
Feedback on the principle “Standards should be extensible, embeddable, and interoperable” pointed out that modularity of metadata standards is affected by differing underlying intellectual models between standards. We agree, and will update the language here to reflect that.
Comments on “Metadata standards should follow the rules of ‘graceful degradation’ and ‘responsive design’” focused on two areas. First, a commenter noted that graceful degradation is balanced by the concept of progressive enhancement. We will consider including both here, perhaps referencing Tim Berners-Lee’s progressive levels of implementation for Linked Data. Second, a commenter requested greater clarity of meaning for the phrase “more sophisticated uses/ontological inferencing.” We will endeavor to achieve this in a revision.
A few new ideas emerged from comments that we will incorporate into the Principles document. First, we will cover the role of documentation for metadata standards, that they should provide sufficient information about the standard, be useful and usable, have examples, and be friendly to multilingualism. Second, we will introduce the concept of inherent bias within any human-created construct, recognizing that good metadata embraces diverse viewpoints and that metadata structures themselves likely encode some (typically unconscious) biases of their creators. While we cannot eliminate this bias, we can raise awareness of it. To assist with effective presentation of these issues, we will consult with ALA’s various ethnic caucuses to review draft language. Finally, privacy issues surrounding metadata implementations are significant enough to deserve mention in this document. We will treat de-identifying metadata as a use case for the graceful degradation principle (#7).
We thank all who took the time to comment on this second draft of the Principles for Evaluating Metadata Standards. The comments are extremely useful and will result in a stronger final product. The Metadata Standards Committee is currently working on making the changes described here and in doing a final review of the document. We anticipate releasing a new version of the Principles in spring 2016.