The goal of the project is to evolve the Internet into the Interspace, in particular by bringing professional search and display of structured documents to the Net. To accomplish this, we are constructing a large-scale digital library testbed of SGML journal articles while concurrently performing the underlying research to provide effective interaction across networks with such a library. That is, we are critically concerned with both the functionality needed to interact with structured documents and the infrastructure required to scale such functionality up to a global community. Because of our goal of directly affecting the Internet, we are heavily involved with industrial partners in the publishing and software industries as a major part of their future strategy. The production digital library we are prototyping at the University of Illinois will thus be quickly transferred into commercial production, whose functionality will increase over time as more of our research results are incorporated into the testbed.
The system model we are pushing has distributed repositories run by the publishers containing their collections. The "library" provides the integration functions of searching and displaying across the multiple heterogeneous sources (this is the traditional reference function of a library, often referred to as "A& I" for abstracting and indexing.) Finally, the user can interact with multiple displays corresponding to multiple styles of searching the information in the collections for different purposes. Superficially, this looks like a standard client-gateway-server model, but the "distributed library model" must deal directly with the problems of providing transparent interaction across the differing semantics of heterogeneous information sources. (Here, differing SGML semantics of tag definitions (DTD), style definitions (DSSSL), and query languages.)
This quarter we held our first Publisher Workshop and it is now clear that our partners are strong enough and there is enough "real" SGML for us to move into initial production of the testbed over the summer as planned. The evaluation efforts studied the prospective usage in focus groups and are well organized to begin usability testing next quarter. The technology efforts produced a major result with the first interactive prototype of a search interface with A& I from both a thesaurus generated by human indexers (INSPEC hierarchy) and a thesaurus generated by a statistical program (co- occurrence graph). Details on all of these accomplishments are given below. We are trying very hard to have strong efforts in all of these aspects of testbed construction, testbed evaluation, and technology research.
Finally, this quarter we also hosted the first DLI-project-wide workshop held at a project site. Investigators from the projects, officers from the agencies, and invited guests (e.g. NCSA Industrial Partners) all came to the University of Illinois. There were overview presentations of each project, technical talks by each project, and a variety of discussion groups to share information. The format seems to have been successful at enabling everyone to lay out their positions and we all can now begin the hard work of trying to work together. Subsequent workshops are accordingly likely to be more informal.
The major accomplishment for the testbed this quarter was making the first small SGML repository then holding a Publisher Workshop to discuss the resulting process of building a digital library with our partners. Fifteen of the major publishers in engineering and science sent their managers of electronic publication plus other technical personnel to our workshop. The attached attendee list represents a very influential group of publishers. They were extremely supportive of our project, and at the end insisted on coming back in six months at their own expense to have a hands- on workshop. (We have budgeted in our grant a yearly workshop for our publishing partners.)
Despite the status of being an international standard, the actual state of SGML in real publishers is not very far along. A few of the partners are now producing some of their journals in SGML and it is these we will be using in our initial production period. The DTDs (tags definitions) are still evolving and the publishers know little about styles (typically the SGML is automatically generated from the electronic texts before being transformed into the format for the print version; thus the publishers have never displayed the SGML before and we are the first ones to actually look at it!).
At the workshop, it became clear that our strategy will have to be to do everything ourselves locally to generate structured document repositories, except actually generating the electronic text: parse the DTDs, define the styles, add additional tags for indexing purposes, perform the indexing, build the database, etc. At the hands-on workshops, we will be gradually training our partners to set up their own repositories. Thus, except for internal simulations, we will not be able to test a distributed repository for some time (after the publishers are further along in their technology). Remember that our overall goal is to make distributed repositories a viable technology in the Internet, rather than, for example, putting up as many documents as possible immediately.
The initial journals encoded into our repository will center around computer science (from the IEEE Computer Society and the IEEE as a whole) and physics (from the American Physical Society and the American Institute of Physics). We will be running usability tests of the prototype interface (see below). In addition, we will be setting up public terminals in the Grainger Engineering Library, in the Computer Science Department, and in the Beckman Institute (for the computational physicists at NCSA). Note that at the University of Illinois, both computer science and physics are in the engineering college and thus served by its library. We have identified enthusiastic early users to debug the system and evolve the interface. The next quarter will see the initial deployment.
We are also experimenting with other remote repositories. One of our partners is the American Astronomical Society, who have a grant from NSF to place a SGML repository of one of their journals on the Net with a Z39.50 interface. We plan to interface to this as it becomes available as a way of experimenting with the fully-fledged distributed repository environment that we will support when the bulk of the publishers begin to build their own repositories. Some journal articles in the astronomy repository also mention names of astronomical objects, which will enable us to build live links to the another digital library, the BIMA database of radio telescope images supported at Illinois by an NSF Grand Challenge grant.
Building the testbed is centered in the Grainger Engineering Library. The primary efforts are in document encoding and database search. The encoding involves gathering the SGML from the publishers, enhancing the tags for indexing purposes, and specifying the styles necessary for reasonable display. Currently, there is quite a bit of manual labor involved for each article, but we are quickly understanding how to automate the encoding process. The search uses the structure of the SGML to provide fine-grained retrieval in a variety of ways. Currently, we are using a small project-built search engine based on SQL. However, we have also extensively investigated existing commercial full-text engines that can handle SGML and scale up to the required usage. This quarter, we hosted full-day collaboration visits from the major vendors of full-text SGML, including BRS, EBT, OpenText, and OCLC. We will be continuing to obtain copies of each of these systems for evaluation, with the goal of choosing one soon to be our large-scale reliable search engine. A streamlined user interface for the database search with multiple interaction paths exists in prototype and is a major emphasis in the usability experiments. The DLI lead on the testbed and the database is Bill Mischo while the lead on the encoding effort is Tim Cole.
In particular, the testbed activities focused on: 1) setting up streamlined document transfer procedures from the APS, AIP, IEEE Computer Society, and ASCE to the Project; 2) continuing to develop manual and software procedures for more efficient SGML processing (in particular, to capture images, handle entities, and add necessary metadata to each document); 3) building software to translate APS, ASCE, and IEEE CS DTDs into a standard canonical form for effective indexing and retrieval; 4) writing software to fully index SGML documents (utilizing the ability of SGML to indicate content and structure of a document) and put the data into an SQL database structure; 5) testing the efficacy of the SQL database structure; 6) designing and
testing user interface software modules for full-text retrieval; 7) working with SoftQuad on procedures to effectively display the SGML documents in the interface environment. This included algorithms and procedures for dealing with the many configuration issues that arise when using Panorama for a heterogeneous collection of science-oriented documents, including particularly special characters and graphics.
In addition to this testbed prototype being constructed in the Grainger, the Mosaic project in NCSA is working on software infrastructure to support digital libraries. On the client end, they have built the CCI (Common Client Interface) package, which will enable the separate SGML viewer being used to display the documents (currently Panorama from SoftQuad) to be part of the Web (that is, links may be followed to other documents directly from an SGML document displayed in Panorama). On the server end, they are developing stateful gateways (CGI -- Common Gateway Interface) for SQL. These gateways will not only link across the network to the search engines, but also record the history of the search session so that the results of previous searches can be refined. For SQL, the state is recorded by saving intermediate join tables. We will subsequently be developing a Z39.50 gateway as well. Jason Ng is building the gateways.
The stateful gateway is our immediate approach to the problem of providing translation to/from multiple heterogeneous distributed repositories. As part of our DLI project, NCSA is also designing new server architectures to handle the demands for next generation Web services, specifically intended for distributed repositories. These encompass both search engines and intermediate gateways, and include all the required functions: depositing from external sources, supporting metadata for search, supporting indexes for search, storing the documents, performing the search, handling external links, interfacing to the network, and so on. Conceptual diagrams of these functions exist already and we are beginning to develop the details of how to actually make repositories work in practice. The DLI lead on this effort is Joseph Hardin with conceptual details being worked by Dan LaLiberte. Implementation of these diagrams will eventually include solutions to the problems being discovered in the testbed construction, such as mapping fields across repositories (e.g. every publisher has a different definition of "author").
To make distributed repositories work in practice, it is necessary to move beyond the syntactic solutions interfacing query protocols between search engines to the semantic solutions of performing a canonical mapping between the repository contents.
User-centered research related to testbed development continues to be conducted by faculty and students in the Graduate School of Library and Information Science, Sociology, and Economics. The primary focus in this quarter has been on testbed usability studies, observation of users of a variety of electronic media, and the development of system instrumentation mechanisms.
As a warmup to the usability of the testbed prototype itself, a series of usability tests of the thesaurus component of the testbed (see below under Technology) were undertaken and summary results were presented to the thesaurus designer. The tests were conducted in the Grainger usability lab. Subjects (including a graduate assistant who works at the reference desk in the Grainger library, as well as with a graduate student, undergraduate, and faculty member in computer science and engineering) were given a brief demonstration of the system and were then asked to perform two thesaurus search tasks. The search sessions were captured online as well as videotaped. The researchers also wrote down the users questions and problems during the sessions and concluded each session with a series of questions about the users impressions of, and recommendations regarding, the prototype thesaurus.
The usability tests revealed a general appreciation of the utility of the prototype thesaurus, but also identified problems that users had with the design of the thesaurus interface, including general difficulties with manipulating windows, insufficient labeling and instructions, and lack of clarity in the purpose of certain thesaurus functions. During this quarter the evaluation team also helped to draw up a preliminary plan for an experiment designed to compare the effectiveness of both the INSPEC and automatic thesaurae being developed for potential inclusion in the DLI testbed.
The evaluation team is currently planning for the first stage of deployment of the prototype DLI testbed. Several public sites for deployment at the University of Illinois have been identified. The sites have been selected to allow the study of system use within both library and workteam environments, and across several engineering disciplines, by various segments of the academic engineering community, such as faculty members, graduate and undergraduate students, and librarians. The next quarter will see both continued usability testing of the prototype along with the collection of data about system usage in these actual work and learning situations through observations, interviews, and the solicitation of user feedback.
Members of the evaluation team also began conducting a series of observations of the use of various electronic information resources and tools by students at nearby University High School. Some interviews with students and teachers are also being conducted. The purpose of these observations is to study the manner in which students obtain information through the use of a variety of computer applications. Both the computer labs and the library are serving as observation sites. Learning about the needs and habits of current high school students will give us insights into their future needs as college students; it will also suggest the potential utility and usability of DLI testbed for younger segments of the population. In addition, these observations provide an opportunity to investigate the use of a range of electronic resources (e.g., multimedia fulltext sources) relevant to development of functions and features of the DLI testbed.
Another important activity of the evaluation team during this quarter was the development of a preliminary version of a DLI testbed user registration form and procedures for implementing it. User registration procedures are necessary to collect basic demographic data on testbed users and protect publishers against unauthorized use of their testbed materials. Some of the basic issues that need to be dealt with are the technical feasibility of automating the registration process, determining whether all testbed usage will require registration or whether unregistered usage will be permitted at certain public sites, and determining the manner in which registration forms will be processed and the data they contain stored and analyzed. The next steps will be to obtain consensus from system designers and publishers on the registration process, implement the registration form in the appropriate online application, and test the processes by which users are authenticated and data about them are collected.
Instrumentation related to capturing data on testbed use also progressed during this quarter. The Grainger team has instrumented the current testbed prototype to record actions taken by users as they perform searches. The Mosaic team has produced a pilot instrumented version of that application as well, in consultation with the evaluation team. Work on designing an instrumented version of Panorama is just beginning; this instrumentation would be used to record the manner in which people display and read documents in the testbed collection. The next steps in the instrumentation effort will be to test the robustness and utility of the instrumented software. Careful consideration needs to be given to the nature and handling of what will be a huge store of usage data. In addition, the manner in which data from the different applications that make up the complete testbed will be integrated, over time, must be investigated.
During this quarter, Bishop conducted an empirical investigation of the phenomenon of the virtual art gallery by conducting interviews with artists and museum administrators, implementing an online survey of visitors to several Internet art exhibits sponsored by the University of Illinois, and analyzing server statistics on the use of those exhibits (Bishop and Squier, 1995, in press). The study was not undertaken as a primary activity of the DLI evaluation team, but it is relevant in that it provided some exposure to online data collection methods. It also illuminates general user behavior of networked information services and stimulates thinking about new models for the presentation of information in networked multimedia environments. Bishop also conducted a comparative evaluation of seven different electronic journals to gain insights into current mechanisms for (and problems in) structuring, navigating, reading, and archiving scholarly material in electronic formats (Bishop, 1995, in press).
The activities of the evaluation team during this quarter have led to a growing realization of the need for a theoretical framework to guide research on digital libraries and the nature of individual and group processes for creating and using knowledge. The realization stems from their own internal studies as well as from their interaction with other DLI evaluation researchers and their review of the current landscape for networked knowledge production. Thus, the evaluation team will concentrate on developing its own theoretical framework in the next quarter and will continue its efforts to share ideas and results with the broader research community. One important avenue for sharing results continues to be the dissemination of informal working papers via the project homepage.
Co-PI Bishop continues as the evaluation team leader and Star continues as the technical lead for ethnography and research. In addition to their responsibility for evaluation research related to the University of Illinois project, they have played an important role in promoting evaluation efforts across the other DLI projects. Bishop serves as chairperson of the DLI- project-wide working group on user-centered evaluation and they have developed and received funding for a research symposium on user-centered digital library design and evaluation that will be held in October 1995. Star has participated in several national meetings related to furthering humanities and social science efforts for digital libraries, and Bishop will participate in a panel on the DLI initiative at the ACM's Computer Human Interaction (CHI- 95) conference in May as well as host a user research session at the Digital Library 95 conference in Austin this June. Greg Newby (Assistant Professor) and Bob Sandusky (doctoral student) in the Graduate School of Library and Information Science joined the evaluation team toward the end of this quarter. They will take the lead in the technical aspects of developing and implementing evaluation procedures which necessitate extensive online data collection and analysis, such as system instrumentation and the processing of user registration data.
The Technology Research on our DLI project is doing longer-term research into increasing the interaction with information across the network, with the goals of eventually moving this new technology into the testbed.
The major accomplishment of the technology research in this quarter was in the area of Semantic Retrieval. We have built the first prototype of what search intermediaries in the Net might really be like for digital libraries. (This was also the topic of our talk at the DLI workshop.) Our research is based upon indexes with semantics beyond full-text search. It might be regarded as the computer equivalent of the traditional library "A& I", Abstracting and Indexing Services.
Since text search is still word matching at the base, scalable approaches to deeper semantics involve interactive programs for suggesting alternative terms to search for. Interactive term suggesters use a classification scheme containing words within the subject domain, often referred to as a "domain thesaurus". Traditional A& I has this thesaurus generated by human indexers (professional librarians), who assign several thesaurus terms to each document in a collection. We are examining having the user interact directly with a thesaurus, both the manual classification and some new sophisticated automatic classification schemes.
Our prototype enables a user to search a collection of documents using a variety of indexes, each with its own interface and with the ability to drag- and-drop terms found in one index to another in the hopes of finding a suitable set of terms for a desired full-text search. Since our initial testbed collection concentrates on computer science and physics, we are using the INSPEC thesaurus produced by the IEE, which broadly covers the journals in this subject area. The prototype displays the INSPEC thesaurus as a tree, showing broader and narrower terms, and as a graph, for related terms. In addition to this human-generated manual index, there is also a computer- generated automatic index.
The automatic thesaurus index is based on term co-occurrence, i.e. a matrix is generated of how frequently each term occurs in the same "sentence" as each other term. Running this on a large collection generates a graph, where terms are connected by their frequency of occurring together. This graph, sometimes called a "concept space", can be used to suggest alternative terms from a given term. Note that the suggestions are words that occur together but not necessarily synonyms. Thus the automatic classification is based on "context, whereas the manual classification is based on "meaning".
A concept space corresponding to the INSPEC thesaurus was generated, based on 400,000 abstracts from the last 3 years of INSPEC coverage. This computation took 23 hours on the NCSA Power Challenge and was the principal user of the supercomputer for the 3 weeks preceding the DLI workshop! The prototype interface enabled the user to search both the manual and the automatic index and see the advantages of each. For example, searching for "information systems" in the manual thesaurus yields terms of more specific information systems, such as geographic information systems or medical data processing. On the other hand, the automatic thesaurus instead lists applications of information systems, computers used to implement information systems, professions dealing with information systems, and so on. Depending on the need, either index might produce valuable alternative terms for the user.
User experiments are proceeding to evaluate the utility of this multiple index approach, but our philosophy is definitely that multiple interfaces and multiple indexes are necessary for effective search across multiple repositories. The DLI leads on this work are Hsinchun Chen at the University of Arizona for the automatic classification, Eric Johnson (Grainger Library) for the interface software, and Pauline Cochrane (GSLIS) an internationally renowned expert on A& I.
We are also investigating methods by which this semantic retrieval and beyond might be scaled up to large distributed repositories (the Net or the NII). The Computing and Communications Group at NCSA is doing performance measurements of the NCSA Web Servers and these results will be integrated with the forthcoming instrumentation for the DLI to try to predict the usage patterns of a large-scale digital library.
This group, notably Nancy Yeager and Bob McGrath, are also doing research, supported by the DLI and by the Illinois NASA CAN grant, on how to effectively implement scalable versions of unique invariant document names (URNs rather than URLs). They produced a document comparing the NCSA requirements to several existing implementations (Handles, DNS, LIFNS, Whois++). The Handle System from CNRI has been chosen as the initial system to experiment with.
The Software Development Group at NCSA is addressing the question of document classification in the Web. Jointly with OCLC, they organized the Dublin Metadata Workshop, held at OCLC in April, to attempt to generate a standard list of metadata objects for documents. In the A& I context discussed above, these might be thought as of enabling the author of a document to add appropriate indexing for the document (keywords, subjects, title, etc). If this effort proves successful and publishing documents directly from editors becomes common, this will be another important type of indexing for documents in the Net. See the report Metadata Workshop: The Essential Elements of Network Object Description and the references below for the people involved. http://www.oclc.org:5046/conferences/metadata/metadata.html
Thus, the eventual strategy would be to have authors classify their own documents, following by classification of collections (groups of documents) by indexers, either manual for important static collections or automatic for "unimportant" dynamic ones.
The final interaction research topic is development of New Architectures, which will eventually produce the system properly referred to as the Interspace. This will be based on the computing model where the operating system is distributed objects and every machine does some combination of every step in the publishing cycle, from authoring to indexing to searching. It is thus a peer-peer rather than a client-server model. This architecture makes the assumption that interoperability is solved, then concentrates on analysis, how to cross-correlate objects from multiple sources to solve problems. A preliminary design of this architecture is continuing, with the first implementation efforts shortly approaching. The lead here is Bruce Schatz.
One last research topic was proposed this quarter in a supplement to the main DLI grant. The lead is Tom Huang in the Beckman Institute and the topic is image processing. In particular, how can images be encoded to increase network throughout and how can features be extracted to improve search? This issues are being examined in the context of an application in the arts and humanities dealing with a collection of digitized paintings provided by the Getty Foundation. A secondary application is the astronomy image database discussed above. The hope is that this research will expand the range of media and the range of subjects that our digital library technology is applicable to.
Beside the publishers, we had a very active quarter with respect to partners and gained several new significant ones.
We received a major equipment grant from Hewlett-Packard in excess of $1M. Roughly half of this is allocated to the Computer Science Department to set up a teaching lab in digital video. The other half is directly allocated to the DLI project in several parts. The Grainger Library will be receiving a large file server and several large workstations as the central machine servers for the testbed repository. The Graduate School of Library and Information Science will be receiving a workstation cluster to set up a teaching lab to train professionals in information systems. NCSA will be receiving workstations for work on the digital library project, both to the C& C (Hardware) and the SDG (Software) group.
We are dealing with several partners and prospective partners with respect to full-text search engines: BRS, OCLC, EBT, Opentext. We met this quarter with high-ranking officials from each and a substantial arrangement with each is in progress.
We met several times with Microsoft to discuss our digital library technology vis-a-vis their new network operating systems and on-line information services. We have a pre-alpha arrangement with them and a closer relationship is under negotiation.
As discussed above, we are serving as an initial test site of the Handle software from CNRI.
We have reached an arrangement with the NetBill project at Carnegie- Mellon University (M. Sirbu et al) which will provide us with a way of experimenting with the charging schemes that our publisher partners are so interested in.
Bishop, A. P. (1995, in press). Scholarly journals on the net: A reader’s assessment. Library Trends, 43 (4), 544-
Bishop, A. P. and Squier, J. (1995, in press). Artists on the Internet. In INET ‘95 Proceedings.
Chen, H. "Machine Learning for Information Retrieval: Neural Networks, Symbolic Learning, and Genetic Algorithms, Journal of the American Society for Information Science, 46(3): 194-216, 1995.
Chen, H., Martinez, J., Ng, T., and Schatz, B. "A Concept Space Approach to Addressing the Vocabulary Problem in Scientific Information Retrieval: An Experiment on the Worm Community System", Journal of the American Society for Information Science, 46(3): 175-193, 1995.
Cole, T. "Digital Library Projects", LITA Newsletter, 16(2): 25-27, Spring 1995.
Johnson, E. H. and Cochrane, P. A. "A Hypertextual Interface for a Searcher's Thesaurus". Digital Libraries '95 (Austin, June 11-13, 1995). In press.
Schatz, B. "Building the Interspace: The Illinois Digital Library Project." Communications of the ACM, 38(4): 62-63, April 1995.
Yeager, N. "Critique of Existing URN Implementations, when assessed against NCSA URN Requirements", http://www.ncsa.uiuc.edu/InformationServers/Horizon/URN/critique.html
Public Presentations
Bishop, A. and Mischo, W. Project briefing by Mischo and Bishop at the Coalition for Networked Information’s spring meeting on April 10, 1995.
Crawford, H. DLI Sociology Presentation, Illinois Library Association Conferece, Peoria, Illinois, May 5.
Hardin, J. WEB METADATA Workshop "The Essential Elements of Network Description", Dublin, OH. Moderated a large group session, co- organizer of the workshop, Feb. 28 - March 3.
Hardin, J. World Wide Web 3 Conference, Darmstadt, Germany, International World Wide Web Conference Committee, April 10-15.
Hardin, J. ARPA/NCSA/MITRE meeting to discuss enabling searching over the Net, Vienna, Virginia, April 16-18.
Hardin, J. Presentation on NCSA and the World Wide Web at the Symposium on Computers in Health Care Education, Philadelphia, PA, April 26-28
Johnson, E. DLI Publisher's Workshop, Grainger Engineering Library Information Center, University of Illinois at Urbana-Champaign, March 23- 24, 1995. Demonstrated thesaurus navagation software, loaded with the INSPEC Thesaurus, to DLI publishing partners.
Johnson, E. DLI-Wide Workshop, Beckman Institute, University of Illinois at Urbana-Champaign, April 24-25, 1995. Demonstrated interoperability of thesaurus navigation software and H. Chen's concept space in a generalized retrieval environment.
LaLiberte, D. WEB METADATA Workshop "The Essential Elements of Network Description", Dublin, OH. Moderated two small group sessions, February 28 - March 3.
LaLiberte, D. World Wide Web 3 Conference, Darmstadt, Germany, paper presentation on "A Protocal for Scalable Group and Public Annotation Services", and presentation of NSCA's WEB architecture developments, specifically the repository concept, April 10 - 15.
Magliery, T. SGML Documentation '95 Conf. and SGML Open Tech. Committee Meeting, Long Beach, CA, March 8-12.
Magliery, T. and Sanderson, B. World Wide Web 3 Conference, Darmstadt, Germany, Tutorial “CCI Applications: Local and Remote”, April 10 - 15.
Mischo, W. DLI presentation at the Illinois Library Association Meeting, Peoria, IL, May 5.
Ng, J. NOAA, Federal Consortium Development, Needs and Progress, Washington, DC, March 23-26.
Schatz, B. "Building the Interspace: Digital Library Infrastructure for a University Engineering Community", AAAS Meeting, Atlanta, February 19, invited talk in Digital Libraries session.
Schatz, B and Mischo, W. "The Illinois Digital Library Project", NCSA Industrial Partners Workshop, April 25.
Shapiro, M. Presentation of the URN scheme to the URI working group, Internet Engineering Task Force Conference in Danvers, Massachusetts, April 1-6.
Access (Chris Adasiewicz), "Shaping Tomorrow's Internet", Spring 1995.
Kodak-Follet: Demonstration of the DLI testbed, 2/1
Dept. of English, UIUC/Michael Mullin, and visitors from Folger Shakespeare Library: Discussions on research & development issues involved in implementation of interactive multimedia technology into the Internet and demonstration of DLI testbed, 2/17
Microsoft: Negotiations and technical collaborations with Microsoft on DLI and Mosaic Issues, 2/27
INSPEC: Visit by Jim Ashling to discuss incorporation of INSPEC thesaurus into DLI testbed, 3/6
EBT (Electronic Book Technologies): Meeting with DLI testbed team to discuss use of EBT software, 3/15
Unicef: Demonstration of DLI testbed, 3/15
American Theological Association: Demonstration of DLI testbed, 3/16
University of Arizona, Hsinchun Chen, Discuss automated program to build relationship graphs of words for any Net document based on co-occurrence within sentences with server, 3/16
Marvin Sirbu, Tom Wagner of CMU-Philadelphia and Conrad Tselepsis, representative Mellon-Bank: All Day Meeting "Overview of the NetBill Project" covering Pre-Commerical Trial of NetBill from a Business Perspective-Providing the Billing Mechanisms, "NetBill: The Client Side Software and Porting it to the Mac, "Coordinated Efforts from a Management Perspective:, "Security Plug-In Module Discussions and , "Modification of NCSA httpd Server to Support the Extensions Necessary for Security Framework" , 3/20
University of Wisconsin, Madison, Demonstration of DLI testbed, 3/21
CNRI: Visit from Bill (William) Arms, 3/21
Schlemburger: Demonstration of DLI testbed, 4/6
BRS/Dataware: Meeting with DLI testbed team to discuss use of BRS software, 4/27
CIC (Committee on Institutional Cooperation - Consortium of Big 10, University of Chicago and UI at Chicago), 5/2
OpenText, Meeting with DLI testbed team to discuss use of OpenText software, 5/3
University of Illinois at Chicago, Demonstration of the DLI testbed, 5/5
Tribune Co. Board of Directors, Demonstration of DLI testbed, 5/9
Go back to the DLI progress reports page