[San Diego agenda]
[presentation summaries]
[list of San Diego attendees]

Dear Colleagues,

Appended below are short summaries of most of the presentations made at the Informal Meeting on Mathematics Digitization, which took place at the Joint Mathematics Meetings in San Diego on Jan. 7, 2002. In addition, there is a note from Martin Grötschel who was unable to attend the meeting, and an update from Rolf Jeltsch.

This meeting brought together a number of mathematicians, librarians (Cornell, Courant, Göttingen), publishers & journals (CUP, de Gruyter, Duke U. Press, IP, Pacific J. Math., Springer), representatives of various related projects (arXiv, Euclid, GDZ, NUMDAM, MathDL), institutions (AIM, Clay), organizations (NSF) and Societies (AMS, EMS, IMU, MAA, SIAM) as well as others with an interest in the digitization of mathematics. In addition we were fortunate to have the executive editors of Mathematical Reviews and Zentralblatt für Mathematik as well.

I'd like to thank all of the participants for their interest and their help in making the meeting a success. In particular, I'd like to thank Philippe Tondeur for all of his efforts in starting and publicizing this important international project, and John Ewing for writing "Twenty Centuries of Mathematics: Digitizing and disseminating the past mathematical literature" (http://www.ams.org/ewing/Twenty_centuries.pdf). I'd also like to thank those who made presentations at the meeting and in particular those who have provided the summaries that appear below.

It appears that many discussions of the proposed International Project to digitize the mathematical literature are now or soon will be underway as is evident from the following:

Math Digitization was discussed at the annual workshop (http://www.cecm.sfu.ca/ceic/) of the CEIC (http://www.ceic.math.ca/). Since January, Philippe Tondeur has had a number of additional discussions with members of the international community on this topic. The EMS will have a "brainstorming" meeting in April where Math Digitization will be one of the special topics (see Rolf Jeltsch's report below). There will also be a satellite meeting to the ICM: "A workshop on Electronic Information and Communication in Mathematics", Beijing, August 29-31, 2002 (http://icms.math.tsinghua.edu.cn/). The NSF has agreed to fund two small meetings during the coming year to start the planning for an International Cooperative Project to construct a Digital Mathematics Library.

I hope that these notes from the meeting will be of some use to you. Should you have any related information or updates, I would appreciate hearing about them.

Best regards,

Keith Dennis

January 7, 2002 San Diego, California: Informal Meeting on Mathematics Digitization

Original Meeting Agenda:

January 7, 2002 San Diego, California: Informal Meeting on Mathematics Digitization
Joint Mathematics Meetings

Summaries of the Presentations:

Digital Mathematical Library (Philippe Tondeur)

The Vision

We would like to propose a project which can be realized by the international cooperation of research organizations around the world. It is the creation of a digital (virtual) mathematics library (DML) of the scholarly literature in the mathematical sciences as part of the fundamental world infrastructure for the mathematical sciences and related fields. The realization of this project could have a significant impact on the way mathematics is done and used in the 21st Century. If we are successful with this project, its future impact on mathematics, the sciences , and education could be the most significant event since the invention of scholarly journals as a replacement for the private correspondence of individual scientists. It could also serve as a prototype for a new model of scientific cooperation, a new paradigm for the future of science in our electronically connected world.

This is an updated version of the dream of the Alexandrine Library. The proposed project would be a distributed system, sustained by the existing University Library System of the world, with access to users following established current practices.

Realization
We see three phases to the realization of this project:
a design phase, an implementation phase, followed by a sustained operation phase.

Design Phase.
The task is to set technical standards for making the scholarly mathematical literature accessible online, and to negotiate a protocol for making future mathematical literature similarly available. The design team should be an international consortium from the profession and interested parties - to be supported by planning grants from various research organizations. A representative of the U.S. National Science Foundation (Philippe Tondeur, Director of the Division of Mathematical Sciences) expresses NSF's willingness to participate. DMS liaison is Program Director Henry Warchall.

Issues to be addressed include content, format, copyright, and archiving. This Design Phase is beginning Spring 2002 in a meeting to be organized by Cornell University, under the leadership of Sarah Thomas and Keith Dennis. It is funded by a planning grant from the Digital Library Activity at the U.S. National Science Foundation (NSF Program Director Lee Zia).

Implementation Phase.
The implementation phase or "construction phase" is a massive digitization project of the existing scholarly mathematical sciences literature. It is to be conceived as a distributed system mirrored on many servers around the world. This is a significant capital project of the order of U.S. $100M, to be carried out as an international collective effort of Science Research Foundations around the world. The U.S. National Science Foundation leadership (Rita Colwell, Director of NSF) is keenly interested in the participation of NSF in such an international effort.

Sustained Operation in the Long Term.
This is an essential aspect of this project, to make it part of the world s scientific infrastructure. This is a function expected to be assumed by the system of University Libraries around the world. Critical ingredients are web connections and site licenses for access.

Essential aspects for THE success of this project:
Competition and cooperation need to be effectively balanced.
Internationality
Sustainability

REFERENCES:
Twenty Centuries of Mathematics:
Digitizing and Disseminating the Past Mathematical Literature.
A white paper on the subject (version, by John Ewing, Executive Director, American Mathematical Society (ewing@ams.org):
See http://www.ams.org/ewing/

Remarks from John Ewing

I wrote the paper to outline the reasons it should be done . . . as well as the reasons it would be almost impossible to do. Enumerating the obstacles is as important as enumerating the advantages if one hopes to actually carry out the project. The four major categories of obstacles --- selecting the content, choosing the format, dealing with copyright, and archiving the material --- are formidable and will not be easy to overcome. This is a huge task, with huge benefits.

Carrying out the initial task of digitizing the material is only the first step, however. Disseminating and updating that material will be an ongoing project for years to come, and I have suggested that the best way to accomplish this is to give "ownership" of the material to many different groups in the mathematics community. Those groups can add value to the raw material in many different ways, and then offer it to mathematicians community in different formats (sometimes on a subscription basis). This has the advantage of giving the entire community of mathematicians an interest in carrying the project forward into the indefinite future.

Most importantly, the project should *not* be viewed as the property of a single group. From the beginning, this must be international, with many players and broad interests represented. Even those who cannot contribute monetarily should have a role in the project. Making the project international in this way makes it harder to accomplish; it also makes it far more likely to succeed ultimately.

EMS and Digitalization. Comments by Rolf Jeltsch, President, European Mathematical Society (EMS)

We have acchieved a high level of recognintion with the European Union (EU) Commission in Brussels, especially with the Department General (DG) for Research. Currently we try to influence the definition of the policy of the 6th Framework Program (FP6) in such a way that it can be used for mathematics too.

Digitalisation EMS wants to use the FP6 which has a new program included. It aims in general at strengthening European Science & Technology infrastructure. There is a separate 'sub-progamme' for this purpose. To create an electronic (virtual) library for mathematical literature seems to fit well to this programme. The project are called 'integrated programs'. It calls for an integrated action. EU can support an effort to improve the infrastructure by several individual member states. Due to this construction it is EMS who should act together with some of its corporate soocieties.

Currently France an Germany are already in the process of digitalizing part of their literature. Finland has made strong political statements in direction that it finds the digitalization an important issue.

The current 'short' time goals of EMS are:

• define the project
• identify countries in Europe which want to participate
• make an application with EU for funding

Schedule:

25 February 2002
Meeting of Ph. Busquin, (EU commissionar of DG Research) with Ph. Tondeur, (Division Director Division of Mathematical Sciences), Luc Lemaire, (vice-president EMS and liason officer EMS to Brussels) and R. Jeltsch (president EMS).
This meeting had to be postponed on short notice to 26 March 2002.

19-21 April
Brainstorming weekend with about 40 representatives of corporate societies, parties involved and EMS officials in Berlingen, Switzerland.
The following four topics will be discussed:

• digitalizsation of math. literature
• Zentralblatt Math
• Structure of meetings
• Publishing of EMS

1-2 June
Presentation of the EoI at the council meeting of EMS in Oslo

7 June
deadline for submitting EoI to EU.

Sept-Nov
Presentataion/workshop in Brussels

NUMDAM
NUMérisation de Documents Anciens Mathématiques
Digitization of ancient mathematical documents
www-mathdoc.ujf-grenoble.fr/NUMDAM/

Pierre Bérard (Pierre.Berard@ujf-grenoble.fr)

France produces twenty or so high level mathematical journals, some of which figure among the best internationally, and a few of which are over 100 years old. It was therefore desirable to set up an important digitization operation, the NUMDAM project, for these journals. It is planned to make links between the mathematical databases (Jahrbuch, Zentralblatt MATH and Mathematical Reviews) and the digitized collections on the one hand, and between the digitized collections and other mathematical collections on the other hand, via cooperation agreements.

The Cellule MathDoc is in charge of setting up and managing the NUMDAM project on behalf of CNRS.

General principles
The first phase, funded by CNRS, will enable digitization of five general-interest mathematical journals (about 220 000 pages or 8 000 articles approx.). The digitization operations will begin in January 2002.

• The NUMDAM program is, above all, a scientific operation aiming at giving wide access to the digitized collections for researchers, science historians and librarians. The programme seeks maximum quality digitization, enabling OCR processing, thence indexation, of the texts and excellent quality for screen visualisation and printing.

• The operation was conceived with the idea of national (in particular with Gallica - BnF) and international cooperations.

• Access to the digitized collections will be as widely open as possible.

• Access to the database of digitized items will be free.

Choices
The choices made up are as follows:

• 600 dpi black and white digitization for typical pages, greyscale or colour for the rare plates and photos.
• Good quality cataloguing of articles (XML with a "home-made" DTD), OCR to allow fulltext searches: (except mathematical formulae, which could be processed at a later stage, technology allowing), cataloguing of bibliographies (with the possibility of refining searches to cited authors or titles).
• Delivery of articles to the users in multipage image form, with an ergonomical format (PDF, DjVu).
• Preservation of the economical balance of living journals by setting up ad hoc moving walls.

Project Euclid
Zsuzsa Koltay (Cornell University)

MR's role in the Digitization of Mathematics
Jane Kister (jek@ams.org)

MR itself has a long history of digitization going back 20 years; the retrodigitization of MR back to its founding in 1940 was completed in 1999. With the introduction of MathSciNet in 1996 the full potential of the highly structured and fully digitized MRDB could begin to be fully realized. Since then the number of internal and external links in MathSciNet has continued to grow; in particular, there are now over 180,000 items (10% of the MRDB) with direct links to the original article.

Last year, a new digitization-cum-linking project was begun: the addition to the MRDB of reference lists from original articles, with links to the associated MRDB entries. This has not only facilitated the navigability of the literature but has also enabled the start of citation indexing. There are now over 8,000 recent reference lists from 65 journals. This year, reference lists from 1997-99 will be digitized and added to the MRDB, with links. Publishers will also be encouraged to submit reference lists (in prescribed format) from other journals so that they too may be added to the MRDB.

Another ongoing project will be the addition of bibliographic data from pre-1940 items that have been digitized, with links from the MRDB to the original.

We believe that the MRDB will continue to be an indispensable research tool for mathematicians in the fully digitized era just as it has been in the past.

The Electronic Mathematics Archives Ininiative (EMANI)
Gertraud Griepke (Springer-Verlag, Heidelberg), Bernd Wegner (TU Berlin)

In the first half of 2001 several projects arose to try to tackle this question and to develop models for the archiving of electronic content. A project aimed specifically at mathematics was created - the Electronic Mathematics Archives Network Initiative (EMANI). Since a distributed architecture is deemed better suitable to the task and would reduce the load on the partners of such a project, a network was proposed. This may lead to a more open approach for extending the project from an initial restricted solution to a more comprehensive enterprise.

Though some of the subsequent arguments may apply to all sciences, they are of particular importance for mathematics: Mathematicians and professionals applying mathematics need quick, reliable and integrated access to mathematical publications. Long-term availability of publications is a particular need in mathematics. Electronic storage and offers of publications in mathematics are confronting content producers and libraries with challenging problems. How do they maintain and provide these services in the future? Digitizing of print-only publications and the adjustment of these offers to current facilities provided for electronic publications leads to a new series of problems to be solved. Electronic publishing offers a variety of additional information in mathematics, which has to be integrated into the access of traditional types of publications.

EMANI will need to address all the above aspects, leading to ambitious aims and goals for the project.

2. Aims and goals of the initiative
For the core of the network, a co-operational system of reference libraries and content providers (like publishers and editors) will be set up. Ideally, the final version should serve a long list of purposes:

The initial action will be to store the digital content in mathematics provided by the content providers at the reference libraries. This will be complemented by retrodigitizing all printed publications in mathematics (provided by content providers) at the reference libraries. This would eventually cover a major part of the publications in electronic versions. On this basis first measures can be undertaken to address the long-term preservation of this content in readable form.

However, just to have the content stored somewhere will not be sufficient. Retrospective digitization may lead only to scanned images, which hopefully could be accessed in some repository. A necessary enhancement will be to improve the usability of the retro digitized publications via advanced linking and searching facilities and also to provide convenient and affordable access of the stored content to mathematicians and professionals worldwide.

The reference libraries may even serve as a reference system for other libraries that want to store and provide part of the content or refresh their existing offers by updated material. Bearing in mind the long time scale of the publications provided through the network, (articles from the 19th century to current publications), a system of distribution agents will be needed. This may be a good reason to develop new business models for the distribution of mathematical publications in a combined enterprise between reference libraries and content providers.

3. The starting point
Owing to the complex nature of project the enterprise should initially start on a small-scale. Once the architecture and the action plan have been more precisely defined, an extension to the scope could be considered.

Current library partners to implement the initiative:

• The Cornell University Library, Ithaca, N.Y. Cornell has a good tradition in retrospective digitization projects and is also involved in the archiving discussion for other sciences.

• The State and University Library Goettingen - Also involved with important retrospective digitization projects like ERAM and DIEPER. Additionally, the SUB Goettingen is obliged to collect all publications in mathematics.

• The Tsinghua University Library, Beijing - Tsinghua has experience with the digitization of Chinese publications. They are a Chinese centre of excellence for installing and offering electronic publications.

The participation of the Library in Orsay, Paris is the subject of current negotiations.

Current content providers involved:

• Springer-Verlag.
• Birkhaeuser Verlag.
• Teubner Verlag.
• Vieweg Verlag
• The electronic library ELibM offered through EMIS.
• The European Mathematical Information Service.

The four publishers have a long tradition in publishing mathematics. Their collections include several of the best journals in mathematics. In contrast to this the ElibM is a co-operation of several journals and editors on a voluntary basis bundling electronic offers in a worldwide system of WWW-servers.

It has been agreed that not all partners can address all the aims of the initiative from the start and that even in the long run the reference libraries may serve only basic purposes. One reason for this is that the reference libraries depend on different economic systems. Furthermore, their participation in other projects must be taken into account.

Nevertheless, it is a common understanding that the storage of the content in a repository will have priority in the near future and that in general copies of the content stored in the system should be deposited at all reference libraries as a matter of safety. Updated versions of the content should be exchanged periodically. The partners of the initiative have agreed to provide their own resources to support the aims of the initiative and this will become important in a later phase of the project.

The first phase of the initiative is the transfer of available electronic content from the content providers to the reference libraries where it will be checked to ensure that the files can be used for the archiving. Adjustments will be made to files not suitable for archiving and recommendations on improving subsequent delivery will be developed.

ERAM

The aim of the Jahrbuch Project is to provide a digital library of important documents in mathematics in the 19th and 20th century. The library consists of two parts:

One part is a digital archive at the Staats und Universitätsbibliothek Goettingen (SUB) of complete articles storing selected original publications from the period indicated. Presently in Göttingen nearly all articles of the journals Mathematische Annalen, Abhandlungen der Königlichen Gesellschaft der Wissenschaften in Göttingen, Mathematische Zeitschrift and more are scanned.

The other part is a database capturing and enhancing the complete content of the Jahrbuch über die Fortschritte der Mathematik (JFM).

The JFM was the first comprehensive reviewing journal in mathematics. It contains information on almost all publications in mathematics and its applications in the years 1868 - 1942; more than 200.000 documents were reviewed, among others by famous mathematicians such a Felix Klein, Sophus Lie, or Richard Courant. Keyboarding of all these data is in progress. But the content of the JFM lacks some important features of modern databases such as classification codes, keywords, and translated titles. Therefore mathematicians from all over the world volunteer on providing the additional information needed. Moreover they add comments when appropriate (stressing the importance of further developments) and select important publications for storage in the digital archive. Presently the experts added data up to volume 31 (Zahlentheorie), volume 18 (Funktionentheorie), volume 14 (Reihen), volume 13 (Wahrscheinlichkeitstheorie) and more.

Furthermore the database contains internal links and external links. About 2700 external links lead to the digital archive in Göttingen. But intending a comprehensive digital library there is also cooperation with other archives. At present there are about 360 links to the Cornell Historic Math Book Collection, about 550 links to scanned articles of the Monatshefte für Mathematik und Physik of the project DIEPER (University Linz/Graz), about 8500 links to scanned articles of the Journal de mathematiques pures et appliquées , Comptes Rendus and Annalen der Physik of the project GALLICA (Bibliothéque nationale de France), few links to Berlin-Brandenburgische Akademie der Wissenschaften, Classical Works, Selecta, and Opera Omnia and Dokumentenserver der Humboldt-Universität zu Berlin.

The JFM database is accessible at the server of the European Mathematical Society. It is free of charge. Search and display features are the same as those for Zentralblatt MATH. The combined installation of both databases at the EMIS server present comprehensive information on mathematical literature from 1868 until now.

The Jahrbuch Project is sponsored by the Deutsche Forschungsgemeinschaft (DFG). The editors of these activities are Keith Dennis (Cornell University), Bernd Wegner (Technische Universität Berlin) and Elmar Mittler (Staats- und Universitätsbibliothek Göttingen).

For more information, contact:

Dr. Silke Göbel
c/o Fachinformationszentrum Karlsruhe
Abteilung Mathematik und Informatik
Franklinstr. 11
10587 Berlin
Email: jfm@zblmath.fiz-karlsruhe.de

URLs for references listed above:
http://www.emis.de/projects/JFM/
http://gdz.sub.uni-goettingen.de/
http://library5.library.cornell.edu/cdl-math-browse.html
http://dieper.aib.uni-linz.ac.at/index.html
http://www.bbaw.de/bibliothek/digital/index.html
http://www.emis.de/classics/index.html
http://dochost.rz.hu-berlin.de/listen/dissertationen/historische-liste.html
http://www.emis.de/
http://www.emis.de/ZMATH

Libraries: Activities at Cornell

Jean Poland

CUL is working with Duke University Press on Project Euclid, a partnership of independent publishers of mathematics and statistics journals that is developing digital access to their publications (http://projecteuclid.org). CUL, The University of Michigan Library, and the State and University Library of Goettingen (Germany) are cooperating in a mathematic digitization project jointly funded by NSF and the Deutsche Forschungsgemeinschaft. CUL is also working with Goettingen, Tsinghua University Library (China), and Springer-Verlag publishers to archive mathematical literature.

Libraries are natural keepers of the scholarly record. We have traditionally archived the research literature by shelving, caring for, and making accessible print materials. Cornell University Library is extending that tradition of preservation to the electronic environment. We have been trusted repositories of the printed word and are becoming the repositories of choice for digital materials. We are aware of the issues surrounding access and use; we are working with subject colleagues on developing standards for subject-based repositories and appropriate metadata standards.

CUL has submitted a grant proposal to the NSF for the planning phase of an interantional Digital Mathematics Library. If the grant is successful we anticipate several meetings of stakeholders - mathematicians, technologists, publishers, and librarians - that will result in a proposal to NSF for implementation. I will keep the participants of this meeting informed of the progress of our proposal.

Jean Poland
Associate University Librarian for Engineering, Mathematics, and Physical Sciences
Cornell University
jp126@cornell.edu
http://www.library.cornell.edu/EMPSL/

Comments of Martin Grötschel:

The idea that John Ewing put forward in his outline "Twenty centuries of mathematics" of creating a Digital Mathematical Library (DML) that "contains" all scholarly mathematical literature worldwide is a challenging and enterprising vision. It would be a milestone towards best possible access to the mathematical literature and to scientific publications in general.

It would be fantastic if almost everything ever written in mathematics was accessible to each mathematician and anyone else interested everywhere on the globe. This would create the best conditions for fair competition and for fruitful cooperation.

But everything has its price. This means, for the time being, that such a project needs a thorough and very careful preparation, the involvement of the best forces available, and sufficient funds (in a case where fund raising may be very delicate). Only persistent work will guarantee success in the long run. The project is a great opportunity for mathematics and the mathematical sciences. All mathematical organizations worldwide should contribute to this project (at least morally) and should in some way join these efforts.

I feel that for the German part, the DMV could be apt to take over coordinating activities as far as a German contribution to the Digital Mathematical Library is concerned. Of course, detailed enquiries are necessary to find out whether, when, and where, for instance, competent staff and other resources are available for this highly demanding project.

Several German mathematical institutions have already gathered experiences in carrying out projects of this kind, however, of much smaller scale.

It would be interesting to know which organizations or countries are able and willing to support this very promising project financially so that the respective experts assigned to the project can meet in order to discuss all relevant issues. Many questions to be addressed in this project are still open. And when the different phases outlined in the project, i.e., the design, implementation, and long-term operation are under way, ever new questions and problems will undoubtedly arise.

In such a multi-national endeavour, every issue will require great attention, and already now, I have some problems in mind that I briefly mark out in the following. I agrree with Philippe Tondeur completely that we first need a conceptual planning phase for the DLM project.

Best regards

Martin

Here is a list of problems that, in my opinion, need to be addressed and in many cases must be solved before the project can be launched:

• Organization of the DLM:
  • Who are the partners?
  • Who is responsible for what?
    Agreements between the participating institutions are needed
  • Role of IMU: could the IMU act as the host for the DLM activity?
  • Is IMU able to effectively take care of upcoming issues?
  • Organization during the project phase and afterwards!

• Sustainability
  • It cannot be assumed that there "will be a large number of organizations" taking care of the library by "some magic". The project at least has to come up with an operational scheme and some real world test, whether the chosen scheme has some perspective to develop and survive in the future.

• Selecting the content of the DLM
  • Who should decide what will be part of the DLM (possibly mathematical experts)?
  • How can this be organized (workflow)?

• Content of DLM: Raw material and added value services
  
  • The relation between raw material and added value services has to be clarified (terms and conditions for the use of raw material by service providers)
  • What is the role of service providers?
  • Who protects server logs from misuse?

• Embedding of present and future publications
  • DLM should be more than an archive for the past mathematical literature
  • New mathematical literature has to be inserted into the DLM
  • Agreements with publishers have to be made.
  • Regarding journal articles and other material, conversion tools are needed.
  • Who decides on the workflow?

• Model for the handling and management of the copyright of the publications
  • Who owns the copyright of the publications in the DLM?
  • Who owns the copyright on the DLM as a collection?
  • Agreements with copyright holders need to be worked out.

• Content analysis
  • Model and standards, metadata, tools for generating metadata

• Access to the material in the DLM:
  • Terms and conditions, more concretely: Will it be allowed to download a substantial part of DLM for persons/organizations?
  • Stability, entry points and user-interface need to worked out.
  • Concept for mirroring: Does a mirror have local copies of the content - or does it just have a copy of the index?

• Retrieval in the DLM:
  • User-friendly and effective methods for the search and navigation in the DLM have to be designed.
  • Who takes care of classification and ordering, reference linking?

• Identification of publications:
  • Identification via bibliographic metadata versus byte identity (byte identity may be problematic for scans, bibliographis data requires a common standard),
  • Alternatively, only weak identification such as a hint: "might be the same"

• Standards for the digitization:
  • This should be solved in concordance with the international activities in the field of digital libraries.

• Standards and interfaces for deeply formatted documents, e.g., XML (MathML), TeX
  • This cannot be solved by mathematics independently from the standardization process of the Web.

• Digitization of the non-digital literature:
  • Who will do the work?
  • How will the past literature be distributed to various institutions (disjunctive or overlapping)?

• Long-term archiving
  • Strategy, organization, who is responsible for what, tools?

• Choice of formats
  • Proprietary versus standard formats and conversions (methods and tools)

• Compatibility with other scientific disciplines
  • The DLM should be interoperable with the digital libraries in other disciplines
• ...

This list should not be read as a threat, it just shows that we are addressing a complex issue!

January 7, 2002 San Diego, California: Informal Meeting on Mathematics Digitization

Invited or in attendence