May 21-22, 2003
State and University Library Göttingen
Recorder: Kizer Walker
Wednesday, May 21
Prof. Elmar Mittler of SUB Göttingen opened the meeting and welcomed participants.
The group discussed the implementation and maintenance of registries of digitized mathematics materials as an important basis for coordinating digitization efforts. SUB Göttingen (Hans Becker) will maintain a registry for digitized journals. Cellule MathDoc's LiNuM (Livres Numérisés Mathématiques) can function as an interim books registry; it brings together bibliographic information from several mathematics book digitization efforts:
Rolf Jeltsch posed the question of whether the name "DML" describes the current planning project and its participants or whether it refers to the proposed digital library and a future organization that will sustain it. How open is participation in the future project?
Participants agreed that primary purpose of the present meeting is to lay the groundwork for a project that comes after the final report to NSF. Participation in the future project must be open and inclusive.
Keith Dennis stressed that the product of the current meeting should be a steering committee for the ongoing DML project and that this steering committee should consist of representatives of digitization projects. The responsibility of this group would be to facilitate communication among stakeholders. It is not to be a closed or permanent committee.
The issue of DML structure arose again in this context. Discussants agreed that the goal is not a DML organization, but that some flexible structure would be needed.
Discussion of Steering Committee report by working group topic
Based on discussion of the DML working group interim reports at the Steering Committee and IMU Liaison Committee Meeting in March, the Cornell group prepared a draft summary of the interim reports for discussion at the Planning Meeting in Göttingen. This document will be revised for submission to NSF on the basis of discussion at the Göttingen meeting.
The working group chairs (Keith Dennis, Bernd Wegner) summarized their interim report.
Regarding the content parameters proposed in the draft report, John Ewing noted that determining the "top-to-bottom" scope (i.e., the range of mathematics in terms of levels of complexity) will be the most complicated. There is interest in parts of the math community in digitizing elementary textbooks. Since individual DML digitization projects would seek their own funding, these are likely to be included, even though some would consider this a dilution of the collection.
The content section of the draft report proposes a global registry listing materials that have been digitized and prioritizing items targeted for digitization. Pierre Bérard pointed out that this really entails two types of registry: one for coordinating workflow for digitization projects and one showing availability of digitized material. Can MathSciNet and Zentralblatt fulfill the latter function?
Ewing said that both services are committed to adding links to material that is already in their databases, but noted that adding records is expensive. Bernd Wegner proposed that the two services share the effort of adding the older material and Ewing agreed that the two should cooperate in appropriate ways.
The group discussed the problem of persistent URLs and the relative merits of DOI and URN.
2. Rights and licenses
Pierre Bérard summarized the interim report for the Rights and Licenses working group.
Drawing on the NUMDAM projectís experience negotiating with Elsevier France, Bérard recommended the following model: Author transfers copyright to the institution that owns the journal; the institution then transfers partial rights to the publisher, for a limited time.
Ewing suggested that there are two aspects of the copyright question that concern the DML: how to deal with the existing copyright situation, and how to effect change to copyright. With regard to the latter, the aim should be to have the author assure that the publication falls into the public domain in reasonable time. This would not require the author to retain copyright or transfer it to an institution, etc. The original term of 14 years, renewable to 28, should be reinstated as a limit.
CEIC members said that ways to address copyright term limitation are under discussion in the committee. A statement is expected to come out of the May 24-25 CEIC meeting in Berlin.
Ulf Rehmann raised the question of whether creating a new version of a work by digitizing it gives rise to new copyright. Participants agreed that this might be the case, depending on the copyright status of the original, and discussed how to insure that new barriers to use are avoided. David Tranah noted that no new rights should arise for material that is currently under copyright: the digitizer would be acting under contract with the original copyright holder, who would own rights to both versions.
Conclusions: The DML report should recommend that the "original owner" (not necessarily the "publisher") retain copyright to the digitized version.
3. Economic model
David Tranah summarized the interim report for the Economic Model working group.
Tranah proposed that the following assumptions underlie the DML economic model:
The prospect that the DML effort might result in journal price increases was highly contentious among participants.
Bérard contended that just as governments have traditionally incurred costs for maintaining library buildings, etc., they should be tapped for maintenance of the digital library.
Group discussed the distinction between maintenance and archiving and the proper roles for public institutions, libraries, publishers in these processes.
Jean Poland recommended that costs associated with DML not be tied to journal prices; costs could instead be divided among DML members as a membership fee.
Conclusions: No decisions were reached with respect to covering costs of digitization and maintenance. Final report will leave the question open.
Kizer Walker summarized the interim report for the Archiving working group.
Thomas Fischer noted that the DML must ultimately archive "born-digital" material as well as page images. He urged that the DML discussion differentiate between preserving bytestream (a problem Fischer characterized as more or less "solved") and preserving presentation format.
Ewing raised content permanence as an archiving issue. To what extent are there consistent practices among publishers regarding modification of electronic documents after publication? Participants discussed the uses of errata in the electronic environment as well as versioning (with links to original).
Gertraud Griepke stressed that the archive must include successive versions where these exist and she volunteered Springer material for testing.
In connection with the content permanence discussion, Wegner noted that for material that isnít static (e.g., multimedia), it will be necessary to address which version to archive. Would DML need authorís permission to archive a "frozen" version? What kind of document can DML guarantee to archive?
The Archiving WG interim report describes an OAIS (Open Archival Information System) repository. Fischer contended that for a distributed archive such as the DML will create, the OAIS model would need to be extended to allow for side-to-side movement. Marcy Rosenkrantz noted that some of this work has been done for inter-institutional joint projects involving interoperable searching.
Since the co-chairs of the Metadata Working Group could not be present at the meeting, this discussion was deferred.
6. Technical standards
Thierry Bouche and Ulf Rehmann submitted a final version (May 18, 2003) of the Technical Standards WG report (in paper form).
Discussion focused principally on the extent to which DML should provide and promote DjVu as a delivery format. DjVu was recommended along with PDF in the Technical Standards interim report, but was given secondary status in the draft of the DML final report, based on discussions in the Steering Committee. Rehmann and Bouche argued for stronger promotion of the format. Borwein pointed out that if DML recommends DjVu, it should also recommend that content providers deliver user support for the format.
Conclusions: Final report should take a more neutral position regarding DjVu.
Reports on production processes at the Göttingen Digitization Centre (GDZ) and NUMDAM
The group heard PowerPoint presentations from members of the GDZ team:
Thursday, May 22
Promotion of the DML
Keith Dennis introduced the issue of how to raise awareness and support for the DML among interested parties internationally.
Dennis described a widespread lack of awareness among mathematicians of digital materials that are already available. In order to reach the mathematics community, literature digitized as part of the DML must be disseminated and linked through the networks upon which mathematicians already rely. DML digitization projects must offer records in standard formats to MathSciNet and Zentralblatt and to libraries for their online catalogs. Zentralblatt and MathSciNet should add these records to their systems and add links to the digitized versions of materials they have already indexed. Libraries should add the records to their catalogs.
In addition, the DML must maintain a registry of digitized mathematics and a listing with up-to-date information on the digitization projects. The DML needs to maintain a network of volunteers (on the model of Project Gutenberg), and the DML should collect and disseminate information between digitization projects (on costs, statistics, tools, etc.).
The group discussed various other approaches to information dissemination and promotion. It was suggested that CEIC act as coordinator of information on standardizing a DML interface, and related issues.
Relationship of IMU and DML
John Ball affirmed that the IMU is "solidly behind" the DML effort and asked what coordinating and/or promotional role the IMU could play that would have the confidence of the individual projects.
IMU has already shown substantial support for DML in appointing David Mumford and Rolf Jeltsch as liaisons to the effort and in backing CEIC participation. Six of eight CEIC members were present in Göttingen. Jonathan Borwein said that CEIC has been cautious not to infringe on the autonomy of the DML effort, but noted that CEIC is ready to step in to support the DML if this is what the member projects want.
The group showed general support for some coordinating or supporting role for IMU/CEIC; discussion centered on how to insure the autonomy of the local projects.
The group discussed how DML can begin to provide a shared "front end" for access to digitized materials. It was agreed that technical standards and standards for bibliographic data need to be worked out in discussion among MR/Zentralblatt, the CEIC, and other DML participants. Peter Michor suggested that CEIC invite the digitization partners to a meeting to discuss these matters.
Conclusions: The group agreed to a CEIC-sponsored meeting in early 2004 to discuss technical and bibliographic standards for the DML.
Next steps/ summary
A new draft of the DML final report to NSF will be circulated to participants for comment. A response is required from everyone before the report is submitted.
The Steering Committee that led the planning phase of the DML is to dissolve at the end of the Göttingen meeting. Keith Dennis proposed that a new Steering Committee be selected to lead the DML effort after the close of the NSF-supported planning phase. He urged that the project move forward initially with interim leadership from within the existing group, while remaining open to expansion.
David Morrison suggested the following immediate goals for a new Steering Committee:
The group agreed that the new Steering Committee would include:
The immediate goals of new Steering Committee are: