"Use of HTTP URIs (URLs) for the identification of library elements, vocabularies and bibliographic data..."
Perhaps the term "data" should be exchanged as it is especially important to use HTTP URIs for bibliographic _resources_ and not only for the data about them.
This paragraph contains a more differentiated view on open data than the previous paragraphs which touch the topic. As said in other comments, a general explanation of 'Open Data' (especially compared to 'Linked Data') is missing in the report.
There are also issues of concern around attribution licenses and linked data.
Attribution only works and thus has practical value at a dataset level. Given the composite nature of RDF, any single triple could be referenced or reused by another record or service. Attribution does not work practically in this context.
The report should substantiate its assertions regarding the value of linked data more explicitly. It would be instructive include examples of the benefits derived by other communities.
The British Library welcomes the work of the work of the W3C incubator group on library linked data. The British Library has been experimenting with the practicalities of expression the British National Bibliographic as linked data and our comments draw on this experience.
BL experience illustrates that issues such as identification of the real object distinct from the concept of an object are still very much alive. It seems prudent while such fundamental debates remain unresolved to err on the side of caution and identify both separately. Real work is needed on use cases to illustrate that identification of the real object is sufficient for all needs, not just library requirements.
BL will publish experience of converting BNB from MARC 21 to LOD. In principle, BL is also open to publishing information on the tools employed and where possible, the tools themselves.
The tools are only part of the equation; the expertise necessary for their effective deployment should not be underestimated. BL’s experience certainly confirms the expectation that this is an iterative process.
http://www.bl.uk/bibliographic/datafree.html
Statistical analyses of redundancy in current metadata processes may identify a lot of waste, but comparing well established standards with emerging standards is complex. Measuring linked open data against current processes will be difficult. A further complexity is that there is not agreement within the LOD community on significant issues: such as types of persistent identifier to be preferred; application of RDF model: there are differences of opinion concerning class/property; use of literals; use of blank nodes. This makes engagement with LOD complex, confusing and costly
Agree value of authority files as LOD.
Release of national bibliographies as LOD also has potential to generate a lot of data without excessive duplication and could provide the hooks for holdings of individual libraries.
Not sure where to put general comments because it deals most with the outcome of the paper i leave it in this section:
When I first read the paper i was disappointed. I think I expected a little bit more concrte recommentations. Later I learned that the task for the group is to evaluate is a library linked data group is needed under the roof of W3C. or how it is written in the success criteria ", leading to a clear and agreed view regarding what further standards and guidelines should be developed, and what organization should be set up in order to develop them." in the deliverables this point is described as " propose a way forward for these communities to participate productively in further W3C standardization actions." I try to focus on this point during my comments.
A main question is to less covered in this recommendations, this is in my person view the question of complexity. If the W3C should be active in the field of library data which complexity they should take about, must it really the level of internal library formats like RDA and MARC or should be more general that it understandable and useable by a broader audience. If we decide for the second point (and this is the only option on a W3C level else it could be better done by Library of Congress or IFLA) earns it automatically that not the full complexity of library data can be handled by a W3C standard. A second question is also for which types of library data we need to develop standards. Three very easy upcoming topics are bibliographic data, vocabularies and classifications, but this list don't have to be complete.
Already in the benefits the conflict of complexity is not covered very well instead the goal reads more to find a systems which is suitable for all needs even for internal library use. In my person view it looks like the discussion focused to much on the point how the work of this group ca replace existing systems in libraries but this can be much better discussed in a library only community and don't need the W3C. Certainly the discussion the library community should be connected to the W3C discussion, so that in future the transition costs to an easy to reuse by a broader community format will be easily realized. If this focus would be changed a lot of the barriers would also fall away because the goal won't be to change the library system in the basics what means that also not all libraries have to adapt such a new system this will always take a large amount of time like the transition process from MAB (the german internal library format) to MARC21 is showing. Also the aspects of the costs won't be so high because there will be no force to change.
The future work of a library data group as part of the W3C should focus on the last two problems first which rights protection should data have, so that is can be easily reused outside of the library world and how this data can be more easily shared outside of the library world. Under this conditions a graph paradigm is still useful but not a must. The vision of global unique identifiers sounds nice but in a bottom up approach like the semantic web not realizable, specially because there are already different existing identifiers for library objects. The build of effective mapping and resolver systems like VIAF sounds more promising.
The resulting recommendations are nice for library management but they are not answering the main question:
A)Which standards must be developed
B) How should an organization look like which care about the standards?
With the outcome of the group i would answer this question in the following way:
A) Existing library standards are certifying they need only translated in a best practice way in a linked data format.
B) There is no need for a further work of a W3C group, this development can be done by existing library standarisation comities like the Library of Congress Future of MARC group or different groups as part of IFLA
But in my person view this are the wrong answers because this will lead again to standards which are suitable and understandable by libraries but not enough by external players like citation management software or other existing tools.
Indeed the answers should be
A) We need the development of a lightly W3C standards e.g. for the representation of bibliographic data which are able to deal which the complexity that is mainly useful for non-library systems.
B) A group should created where the voting weights from library community is max 50% of the total votes else it will happen like in the incubator group that to many librarians will be presented and create an library only standard.
The work of a W3C group is only needed if the result will be broadly excepted in the web community not in the library world.
I honestly think this is the best way to get the library culture to buy in to linked data implementations and the semantic web. If librarians become part owners of the process, claiming their role as information professionals, staking their interests in assisting controlled vocabulary mapping and development and improving access to patrons... then we'll make progress. You may need to frame this in terms of the Ranganathan 5 laws, modernized:
1) Information is for use.
2) Every user his/her information.
3) Every bit of information, its user.
4) Save the time of the user.
5) Information access is a growing organism.
Assisting in getting the right information in the right hands at the point of need: isn't that what librarianship is all about?
Returning to the basic focus of our profession may help to sell the change.
This is a wiki draft, and the editorial work to fix things like case will happen when it is turned into a "real" document. It would be good to know, for things like "web", what people find clearest in terms of case.
Comments by Section