A major obstacle to the further development of digital libraries, and the national information infrastructure as a whole, is the lack of adequate means of providing digital objects on any basis other than free, unrestricted access. Publishers wish to specify terms of use and ensure those terms are enforced (optionally collecting payment), before providing valuable materials on the net. While the latter topics are the subject of much commercial activity, the former seems to have been largely neglected. We propose a workshop to begin defining a common machine-readable language for expressing the terms and conditions of use of objects.
We will invite leading members of the library, publishing, economic, legal, policy, and technical communities to define basic requirements and to propose concrete pilot projects. Such projects can serve as the basis of joint funding and research, and can provide common focus for collaborative prototyping.
Discussion
A workshop on the technology of terms and conditions is timely, given the state of electronic publishing, commerce, and information interchange. Lack of means to share materials in a controlled manner is impeding research in digital libraries. The commercial sector is quickly bringing technology to the market with no guarantee that it will interoperate or provide the required controls.
Examples of the kinds of terms and conditions we have in mind include expiration date, number of times that an item can be used, means and location of use (e.g. home use only, not for export), user class (e.g., students, or over 21), price, etc.. Conditions can be stated by the creator of content or by the publisher. Conditions of use might well be included in a search (e.g. find movies under $5), or perhaps even subject to negotiation.
As one example of the difficulties causes by the lack of such a language, consider the Networked Computer Science Technical Report Library (NCSTRL), an ARPA-sponsored project now including works from about fifty universities. University technical reports have a relatively low financial value, compared to, say, a scholarly journal (to say nothing of Terminator 2), and are typically written by authors eager for an audience, yet still some sites have been reluctant to contribute because of lack of means for expressing terms of use. Terms and conditions in NCSTRL are currently expressed in English as "plain text" statements on Web pages. Humans may read these, but they do not provide for machine interpretation, and hence they can not serve to guide automatic replication, to remove reports from circulation when required, or to collect payment. Several Digital Library Initiative sites work with materials that constitute a publisher's most valued works. Since the means to control sharing is lacking, it is difficult even to demonstrate research, much less cooperate on shared research.
The commercial sector is quickly bringing forth technologies for access control (e.g. IBM's Cryptolopes, EPR's DigiBox) and secure payment (CyberCash, ECash, etc) but is not usefully addressing the need for a language of terms and conditions. Current practices in technologies of terms and conditions do not fit the needs of all parties involved: i.e. authors, publishers, users (readers), and do not extend to international requirements. However, integrating points of view and requirements from different communities has proven to be difficult (Klavans 1996).
Luckily, we need not not wait for the technology to make progress, because cryptographic security is not required in all applications. In some cases, one can simply trust participants to act in accordance with agreements. The obstacle is not lack of desire to cooperate, but rather lack of means of expressing desire. In other cases, one may use watermarking to detect illicit copying after the fact, and then use the legal system to prosecute. But even for those cases where strong protection is required (as in the paid viewing of Terminator), the terms and conditions language is logically prior to the security and payment features. Consumers need to know what rights they are buying and what price they are paying before making the purchase.
Furthermore, even when strong protection is required, it is useful that there be a common language for expressing the actual terms, regardless of the means of enforcement, to allow for interoperation among systems. While we do not plan to address cryptographic interoperation in this workshop, it remains a motivating factor.
Our focus will be on establishing a basic set of criteria, including a full list of facts, methods for representation, and technology problems for different aspects of terms and conditions. This will involve:
Terms and conditions are best thought of as one kind of meta-data, along with cataloging information, reviews and ratings, and ownership information. We hope the results of this workshop will fit into the framework begun in the 1995 Metadata Workshop at Dublin, Ohio, and continued in 1996 at Warwick, England[7].
We will adopt the strategy of the Metadata workshops of limiting the scope of the workshop in the hope of making small but useful steps forward. We will not attempt to devise a language capable of expressing any possible set of terms and conditions, however detailed - this would be likely be "AI complete". Rather, we will attempt to determine whether a subset exists that is small enough to be formalized, and large enough to support many useful applications. Furthermore, we will rule several topics out of scope, including
We will if necessary further limit the scope, believing it more
valuable to carefully define one portion of the full space than
to obtain only generalities about a larger area.
Participants
The proposed list of invited participants is weighed towards senior
people in their fields, with some junior and mid-level participants as
well. Since the workshop is small, and since the topics cover several
subject areas, it is desirable to focus on a more senior-level group.
Potential participants will be drawn from several areas including
library, technical, policy, legal, publishing, and economics.
Dates, Location, Steering Committee
The proposed dates are September 24-26, 1996. The conference can be
held at the Columbia University Conference Center at Arden House,
which accommodates up to 30 attendees. The organizers are
Judith L. Klavans, Director of the Center
for Research on Information Access at Columbia University
and James R. Davis (Xerox PARC). The
organizing Steering Committee consists of William Arms, CNRI; Carl
Lagoze, Cornell University; and David Millman, Columbia University.
Workshop Structure
The goal is to have a balance of talks since one of the issues we plan to address is interaction across the four topical areas: policy, legal, economic, technical. The proposed schedule will be flexible to accommodate to the final set of attendees. The workshop will be structured for smaller discussion groups in break-out sessions covering topics to be defined. The purpose is to encourage further discussion and interaction over critical issues rather than discussions divided by community. Reports from the breakout groups will form part of the final report. Since the group is small, the proposed structure also enables group discussion of such issues. Indeed, even the topics for the break-out sessions will be determined by the people present at the time, although the leaders will provide some suggested areas for discussion.
| Evening 9/24/96 | |
|---|---|
| 6:30-8:00 pm | Dinner |
| 8:30-9:30 pm |
Plenary Presentation (30 minutes) Burning issues-overview
Discussion (30 minutes) Interdisciplinary viewpoints |
Day One - 9/25/96 |
| 9:00-12:00 am | Presentations (20 min. talk, 10 min. discuss)
5 presentations - plenary legal, legal (international) policy, technology economic |
| 2:00-5:30 pm | Breakout discussion and Breakout's
Plenary Discussion on breakout topics Breakout Sessions |
| 6:30 pm | Dinner | Day Two - 9/26/96 |
| 9:00-12:00 am | Reports on Breakout Groups -plenary |
| 12:00-2:00 pm | Lunch with Breakout Groups respond to general session issues |
| 2:00-5:00 pm | Planning Session - plenary Identification of
What do we agree on? Future collaborations & Where do we go from here? |
Before the workshop begins, we will create a Web site with an overview and outline of the workshop, including an annotation system [2] allowing commentary on the documents there. This site will be open to the public, allowing participation even by those unable to attend in person. Our current plans call for only brief "position papers", but we will post such as we receive on this site.
During the workshop, a designated rapporteur will be assigned to each session. The rapporteur will be responsible for taking detailed minutes for several reports. First, there will be a brief verbal report at the end of the workshop (one or two slides) on each breakout, as well as some slides summarizing the presentations themselves. This will occur during the workshop itself. On Day Two, just after the workshop, there will be a technical memo produced which summarizes in more detail the workshop itself, including abstracts of all papers and discussion.
After the workshop, we will upload summaries to the Web site, and will write an article for D-Lib magazine covering the workshop and its conclusions. In addition, the organizers will seek a special issue of a publications such as CACM, with a selected group of invited articles from the workshop.
Desired properties of projects are that they be technically practical and legally demonstrative but they need not solve all aspects of information commerce. Collaborative projects, those with strong evaluation components and those able to facilitate international interoperation will be especially encouraged. We will address issues such as scalability and interoperation, and we will identify and recommend transmission and security architectures using open standards.
For example, a project might develop an RFC for a machine-readable notation to encode contractual information such as pricing method (subscription vs. one-time), price, units of purchase, "return" policy and permissions for resale, re-purpose, backup and download. Or a project might develop a security infrastructure, implemented using on-the-fly parsing of such machine-readable contracts. These are but two examples of the type of project we envision but we intend to leave this open-ended to enable participants to freely generate new ideas aimed at incorporating future vision into current needs.