NSF Program: CISE (IRIS), NSF 03-141
Award Number: IRI94-11330
PI Name: Terence R. Smith
Period Covered By This Report: 02/01/97-04/31/97
PI Institution: UCSB & Date: May 1st, 1997
PI Address: Department of Computer Science, UCSB, Santa Barbara, CA 93106
1.0 Summary of Overall Progress - The Alexandria Digital Library research teams made significant progress during this reporting period. The library team continued research on the problem of defining level of geographic detail for geospatial data s ets. Research in User Interface evaluation methods have focussed on developing statistically robust models of user behavior patterns for the current Web interface. The Image Processing team has made much progress in developing wavelet based algori thms for progressive delivery of wavelet coefficients. The Parallel Processing team has been evaluating the applications of adaptive client-server partitioning and scheduling techniques in wavelet image browsing and postscript file document browsing. The Information Systems team made advances in the following areas: development of new disk allocation policies for two-dimensional data; the impact of media exchanges in robotic storage libraries; multidimensional indexing; materialized views selection al gorithm; and the scalable architecture for locating heterogeneous information sources (Pharos). Each team has published the results of their research (see section 8 of this report).
The ADL advisory board met and made two major recommendations to the project: 1) to focus more on collection development; and 2) to hire a project manager. The ADL testbed system activities included formation of a new User Interface Development team. ADL also conducted a major Design Review Workshop in February in which approximately 32 participants evaluated the ADL user interface design and requirements.
2.0 Research Progress and plans
2.1 Library Team - Research on the problem of defining level of geographic detail for geospatial data sets has concluded with the publication of a paper in Geographic and Environmental Modeling and a book on Scale in Remote Sensing and GIS from CRC Press. Research continues on the definition of fuzzy geographic footprints for both information objects and queries. A paper is in preparation on ADL’s work in this area. Efforts are under way to define the concept of a "geolibrary", defi ned as a distributed store of georeferenced information objects. A paper is in press, and it is likely that the Mapping Science Committee of the National Research Council will hold a workshop on this topic later this year. Goodchild is currently chair of the MSC.
Problems associated with on-the-fly registration of geospatial data sets, required for successful integration of data in ADL, are being studied, in part under a new grant from NIMA, and in part through funding from Caltrans and the US Department of Tra nsportation. These methods are being studied and implemented in a variety of clients, including clients mounted in moving vehicles, where retrieval of information from distributed sources is essential to effective navigation. A new initiative of the Nati onal Center for Geographic Information and Analysis, titled "Interoperating GISs", will bring many of the issues faced by ADL to the fore in a combined international conference and workshop planned in Santa Barbara in December 1997. This effort is being c onducted jointly with the Open GIS Consortium, and Terry Smith and Mike Goodchild are both participating in the organizing committees.
2.1.1 Research Plans of Team for Next Six Months - The Library Team plans to continue the activities described above, which are consistent with the plans developed for the team in March. Increasing focus will be given to collections as the p roject moves closer to public access and operationalization.
2.2 User Interface and Evaluation Team - Research in User Interface evaluation methods have focused on developing statistically robust models of user behavior patterns for the current Web interface. User behavior patterns on the Web interface h ave been modeled using spatial interaction models, commonly applied in geographical analysis of transportation routing, migration flows, and journey-to-work studies. We work from a metaphor of users jumping from one Web page to another as they work withi n ADL. We employ spatial interaction models to estimate a matrix of distances between ADL Web pages. Interpretations of matrix cells indicate use patterns of (for example) accessing on-line help files, returning from a tutorial to continue a search. W e can recover and compare the use patterns of searches via the catalog or via the gazetteer.
As reported in previous quarterly reports, we continue to run focus groups on the current and planned interfaces. Our new focus group mechanism is to have one of our staff run a search using a laptop and overhead projector, with cached Netscape sessio ns. We ran a focus group in March for 20 University of Colorado library staff and administrators in this fashion, also showing them screen shots of the ADL user interface design, and thirdly showing them screen shots of the USGS National Atlas interface design.
2.2.1 Research Plans of Team for Next Six Months - To follow on the user pattern modeling, we will disaggregate the transaction logs for specific user target groups in the beta tester pool (earth/space scientists, K-12 teachers, and librarian s) as well as disaggregating ADR panelists and system designers at UCSB. Reitsma, Rock, Tsou, and Buttenfield will work on this effort. Constructing spatial interaction models for specific groups and comparing the parameterized interaction estimates wi ll help to identify differences in use patterns and use error patterns.
To follow on the georeferencing project with the Denver Public Library, we will begin GPS positioning of individual photographs for a pilot group of about one hundred photos of mining and mill sites in the Front Range area in June, when the snow clears from the Continental Divide regions. Smith and Buttenfield will work on this effort. We will continue negotiations with the GIPSY Team to utilize their software for probabilistic georeferencing of the metadata content. Andrew Smith will begin to write up this work for his Master's Thesis project in the fall, and has submitted an abstract to GIS/LIS to present this work at the national conference in the fall.
2.3 Information Systems Team - The team worked with interoperability issues in Java and CORBA and in particular process with migration issues related to digital libraries.
Development of new disk allocation policies for two-dimensional data: It has been shown that strictly optimal solutions to the problem of declustering two-dimensionally tiled data onto multiple I/O devices (possibly disks) in order to achieve m aximum parallelism for rectangular queries, do not exist, except for some special cases. The team has developed new allocation techniques that achieve very good performance in a general setting. These techniques are superior to the existing approaches es pecially for the cases where no strictly optimal solution exists.
Impact of Media Exchanges in Robotic Storage Libraries: In an earlier paper (Scheduling Tertiary I/O in Database Applications) the team investigated the problem of scheduling I/O requests for tertiary libraries. The development was based upon t he assumption that since the cost of exchanging media (tapes or disks) in robotic libraries is so high, it would be beneficial to eliminate exchanges as far as possible. In this study we put this assumption (that is made by most researchers studying I/O in tertiary storage) to test. We showed that though the assumption is not valid all the time, it is so most of the time. Also the gains obtained by not making the assumption are small whereas the resulting computation is much more intense. This paper w as submitted to IOPADS 97.
Multidimensional Indexing: The team addressed dynamic range searching in multiple dimensions. A new structure called the O-tree was proposed. For n-dimensional point data, this structure achieves n(d-1)/d range query time complexity and x\log n amortized time complexity for insertions and deletions. We proved that this query time complexity is optimal when data is not replicated. We then examined how the query time complexity can be improved by replication of data. Based on Bentley's multi-leve l range structures, we proposed a dynamic query structure.
More experiments of materialized views: A set of more systematic experiments of materialized views have been conducted in both Sybase and Illustra DBMSs (the initial experiments were only done in Sybase). In addition to the confirmation that ma terialized views give a significant query performance gain in Illustra, the experiments focused on the scalability and applicability issues of the materialized view approach. In particular, queries with different properties were tested. These properties i nclude join ratios, table numbers and sizes, answer sizes etc. The results show a large range of speedups (from 1.5 times faster to nearly 3000 times) due to those different properties.
Materialized views selection algorithm: The optimal conditions for our proposed materialized views selection algorithm were studied. A near optimal condition was found which guarantees that the selection is within 63% of the ``optimal''.
Scalable Architecture for Locating Heterogeneous Information Sources (Pharos): The team is currently in the process of building a small prototype of the proposed Pharos architecture. In order to accomplish this we have installed the informatio n retrieval system, LSI (Latent Semantic Indexing Package) from Bellcore. This software will be used for automatic classification of information content of sources under Pharos. In addition, we are currently in the process of analyzing UCSB's library cat alog data (approximately 4 million MARC records). These records will be used as sample data for Pharos development.
2.3.1 Research Plans of Team for Next Six Months - Tertiary Storage : Evaluate the performance of declustering schemes with real disks.
Multidimensional Indexing: The team is currently exploring new techniques to support efficient querying of images by their content. Traditional index structures such as R*-trees, X-trees, SS-trees do not scale with the dimensionality of data. T ypical figures show that as the dimensionality increases from 2 to 48, the query performance degrades by a factor of 50 in these index structures. The solution to indexing such high-dimensional data is reduction of the dimensionality of data. The QBIC sy stem at IBM adopts this approach. In this approach, the data is first transformed using SVD or KL-transforms. These transforms yield the most distinguishing dimensions for the dataset. These dimensions are then used in indexing the data. This approach ass umes the database is static, i.e., insertions and deletions of images are not supported. An expanding digital library has to handle such on-line insertions and deletions. Consequently, we are examining how dimensionality reduction can be applied in the co ntext of dynamic databases. Preliminary results with our techniques indicate an order of factor improvement in query performance over traditional index structures.
Database performance tuning: The possibility of using outer-joins in our materialized view approach is being studied. With the natural join, a view can not be used to answer its subqueries because the natural join doesn't contain those dangling tuples. But the outer-join contains both the natural join part and the dangling tuples part. A materialized view of the outer join thus can be used to answer more queries on the additional cost of only a linear space increasing. Both experiments and theo retical work are being conducted on this. Although our experiments have confirmed that using materialized views in answering queries is a gain, it has a large range of speedups (from 1.5 times faster to nearly 3000 times). Several different properties of the queries such as join ratios, table sizes etc., contribute to this large range. The possible rules are being studied and tried to be incorporated into our database performance tuning approach.
Pharos: In the short term (3+ months), the team plans to build a prototype that would enable searching news groups (which will be seen as collections by Pharos). We will make the system available on the Web for general use and testing. The lon g term plan (6+ months) is to conduct analysis of other types of collections (sizes, document distribution patterns, etc.) and enhance our current simulation model to compare subject-based queries to keyword-based queries.
Interoperability: The team will continue to work on interoperability issues, and to develop semantic data model for project. Will also develop user interface for demo.
2.4 The Image Processing Team - During the past few months much progress has been made in developing wavelet based algorithms for progressive delivery of wavelet coefficients. A new balanced rounding integer wavelet transform for color/multispec tral image coding has been proposed. The team continues to make progress in error detection and masking for progressive image delivery (ICIP97 paper). In image segmentation, a robust scheme has been developed for color image segmentation and tested on o ver few thousand images with visually acceptable results (CVPR'97 paper). We are investigating different approaches for similarity matching in the feature space.
2.4.1 Research Plans of Team for Next Six Months - To continue work on progressive multiresolution image delivery including subregion retrieval. Generalize work on error detection/correction. Extend visual thesaurus to integrate multispect ral, texture, and shape attributes as well as spatial relationships.
2.5 The Parallel Processing Team - The team has been evaluating the applications of adaptive client-server partitioning and scheduling techniques in wavelet image browsing and postscript file document browsing. We are also developing a Java based client-side resource monitor to dynamically report client load and client server bandwidth. A piggyback strategy is used to evaluate the client-server bandwidth without creating extra network traffic, and we are studying and improving the accuracy of its measurement.
The team has been evaluating the applications of adaptive client-server partitioning and scheduling techniques in wavelet image browsing and postscript file document browsing. Taking advantage of multi-processor server support with client resourc es can lead to significantly improved response times. The main research challenge is the effective management and utilization of resources from multi-processor WWW servers and client-site machines. Blindly transferring load onto clients may not be advi sable, since the bytecode performance of Java is usually much slower than a client machine's potential. A careful design of the scheduling strategy is needed to avoid imposing too much burden on client machines.
Adaptive scheduling techniques have been developed that optimize the use of a multiprocessor server with client resources by predicting demands of requests on I/O, CPU and network capabilities. We also provide a performance analysis under simplified assumptions for understanding the impact of system loads and network bandwidth when using our scheduling strategy. The experimental results show that substantial performance improvement is obtained by shifting computation to client-site machines and f or some applications, it reduces network bandwidth requirements. Our scheme is adaptive to the variation of client resource availability.
In supporting the implementation of our partitioning and scheduling scheme, we need to develop a client-side resource monitor to dynamically report client load and client-server bandwidth. We have used a piggyback strategy in our Java implementation to evaluate the client-server bandwidth without creating extra network traffic, and we are studying and improving the accuracy of its measurement.
The team has also been developing a JAVA implementation of the wavelet color image browsing scheme. It uses multi-threading to improve the client computation and communication performance and the code is universally executable in any standard WWW brows er.
2.5.1 Research Plans of Team for Next Six Months - We plan to study and evaluate the ADL test bed performance on our parallel machine and workstation clusters when there are a large amount of concurrent requests from Internet. We are curren tly working with the implementation team (Qi and Terry) to port the ADL system on our machine. We plan to incorporate SWEB with the ADL test bed and the main research and development difficulty is to efficiently migrate and maintain states when balancing the processor load by dynamic scheduling. We are also developing a caching mechanism to improve the WWW/ADL server performance. Two of our Ph.D students are in the end stage of research and they will spend time to complete their theses.
3.0 Testbed Development Progress and Plans
3.1 Implementation Team - The implementation team began weekly iterations of screen designs posted to threaded WWW discussion board for subteam feedback.
URL for screen design iterations:
http://www.arts.ucsb.edu/~nideffer/alexandria/index.html
URL for message board: http://www.arts.ucsb.edu/~nideffer/alexandria/wwwboard/index.html
The team formed a workgroup specifically oriented toward interface design and implementation (Robert Nideffer, Nathan Freitas, Kevin Lovette); and proposed and began designing a cross-platform UID and evaluation lab. The first version of the ADL int erface was created as a multi-threaded object-based java applet. Specific accomplishment in this implementation include the following: successfully implemented first version of query object, which handles queries to the actual ADL database; designed a u nified search screen to the user; designed "look-n-feel" object to allow quick and easy modification of applet font, color, and other settings; integrated gazetteer, catalog and map-based searching; designed basic concept of intra-applet communication bet ween four areas: map, search, result, and user workspace; simplified set of search types for metadata attributes; allowed users to search non-contiguous spatial areas; allowed users to select search areas on map (rather than use whole map map window as se arch area).
3.2 Development Plans for Next Six Months - The team will implement a second version redesign of the ADL user interface: vector-based map browser, platform independent look-n-feel, heirarchical sorting in results/workspace, scalable windowing, u ser definable color preferences, move to Java 1.1. We also began redesign of external Web-pages to maintain consistency in the look and feel of ADL's on-line presence
3.3 Testbed Developments Involving Other Teams - Texture search of images developed by the Image Processing Team is being integrated with the ADL testbed.
The Information Systems Team implemented the proposed materialized views in the ADL web prototype system. A translation module has been built to analyze and translate user queries to the equivalent ones with materialized views. The module was demonstr ated during the sit visit and the original unaccepted queries due to the unreasonable response time can now be issued. The answers are returned within a few seconds or a few minutes at the most. This supports the possibility of making a fully populated AD L operational.
In addition, the Parallel Processing Team has been developing a JAVA implementation of the wavelet color image browsing scheme. It uses multi-threading to improve the client computation and communication performance and the code is universally executab le in any standard WWW browser.
3.4 Alexandria Design Review - The second Alexandria Design Review Workshop was held 19-21 February in Santa Barbara in order to review user requirements for the Alexandria Digital Library and to discuss plans for testbed development and researc h progress in the coming eighteen months.
Thirty-two representatives from the public sector, private sector, and academia came from U.S. (one from Canada) for two-and-a-half days to iterate between plenary and small group discussion, software demonstrations, and informal discourse. The intenti on is that panel members will continue to be proactive in the coming year, through on-line activities (working with the Library testbed), through personal contact with ADL project staff, and by face-to-face meetings at national conferences in the coming y ear.
In preparation for the meeting, the Steering Committee agreed upon a set of objectives to guide the content of the agenda. A three-part focus was established to deal in turn with issues of search, browse, and retrieval; content and processing; and inte rface and navigation. These three areas encompass a number of the more difficult problems to be resolved immediately. Their selection was based primarily upon feedback gained in focus groups, survey questionnaires, and videotaped or recorded interactions with actual and potential digital library patrons over the first two years of the Alexandria Project. Many specific topics were discussed including ways for users to view the contents of the library (an overview of the holdings ), giving the user a "road map" of the services of the library, visualization (presentation) of search results, iterative search processes, and customization of the user interface.
From the wide range of comments and recommendations contributed by the Panelists that are detailed in this report, some strong themes can be identified. First, the Panel is supportive of the efforts of the research teams working at Santa Barbara and at Colorado. Panelists and funding agency representatives agreed that keeping a careful balance between developing an operational digital library and contributing to theoretical and applied research is necessary. The importance of the content of the library was strongly supported, as were the enabling technologies for distributed ingesting of data with attention to the representation of data provenance and verifica tion. Collection development criteria were discussed at length, with emphasis on focusing on particular geographic areas and collecting all forms of georeferenced information about that place. Panelists recommended that ADL develop enabling technologies a nd specifications for connecting local collections to ADL, with distributed searching capability.
The overall thread that runs through the recommendations is that ADL needs to focus better on what can be accomplished in the short time left in the current phase of the project. Most panelists reported that the format of the Workshop worked well, but they would like for it to have been more focused and they expect to see specific responses from ADL on the recommendations made.
4.0 ADL Operations
4.1 Data Selection and Loading - The first meeting of the ADL data-loading committee met. The committee is comprised of representatives from the San Diego Super Computer Center, the Southern California Earthquake Center's Crustal Studies group at UCSB, the Map and Image Laboratory, the NSF Ecology Center at UCSB, the NCGIA, and ADL staff. It was agreed that a data loading plan should be constructed representative of broad general interest datasets. It was also agreed that other data for ms e.g., text, cross-sections, seismic, should be included in the plan so those exploring ADL will better understand the scope of materials that may be accessed. It was determined that since much large-scale aerial-photography was available for the South ern California Area, the area ADL will initially focus on, that sample historic based flights should be scanned as part of the phase one plan.
The loading plan for phase 1 consists of selections of digital, analog and metadata content. These collections transcend media types and include cartographic materials, imagery, text and metadata corpus. When loading is complete, about 850,000 metada ta records and 400GB's of actual data will be available to ADL users. The target date for U.C. user testing is Oct. 1997. The scanning plan, now in draft 2, is a collection of procedures for scanning aerial photography and determining the geographical l ocation of each image. Using a flatbed scanner, each analog image is scanned at 600dpi, and metadata generated. Footprint information is determined using ArcInfo. Two California counties are completed and two more scheduled for scanning.
4.2 Operational-Library Committee - Tying ADL to the Library On-line Public Access Catalog (OPAC): The UCSB library has initiated a committee to coordinate ADL with its on-line catalog. The charge of the committee is to define how the system s functionality will interact, determine how users will be passed-off between the two systems and what services will be required to support an operational digital library operation. One of the first objectives of the committee has been to develop user sc enarios to better understand interface linking requirements. Base level interaction has started between the OPAC vendor and ADL systems personnel.
4.3 Other Collection Development Activities - The Colorado Team began a pilot study for the Denver Public Library, to explore georeferencing a series of historical landscape photography of the Front Range region. The series fits the ADL crite ria (already scanned, unique, large -- half a million photos) and beneficially can be stored on-line in the Denver Public Library). Their metadata is in USMARC format. However the photographs lack georeferencing. This spring we designed a pilot project to georeference some of the photos, and to explore ways to automate the georeferencing process directly from metadata content. This work will continue through the summer. A paper has been accepted (Smith and Buttenfield, 1997) for presentation at GIS/L IS '97 on this work.
5.0 Management Issues
5.1 Personnel
5.1.1 UCSB - Robert Nideffer and Nathan Freitas joined the User Interface Development Team. Robert Nideffer researches, teaches, and publishes in the areas of technology and culture, contemporary social theory, and electronic intermedia. He holds a Ph.D. in Sociology, and an MFA in Computer Arts, and is a founding editor of SPEED, an online journal devoted to the study of technology, media, and society. Presently, Robert works with the Alexandria Digital Library where he is director of inter face design and implementation, and is a co-principal investigator of a two year initiative to research and develop online public spaces.
Nathan Freitas is the lead programmer for the Interface Design and Implementation Team at the Alexandria Digital Library. He graduated from UCSB in June of 1996 with a Bachelor of Arts, Creative Studies (Music Composition). Nathan has participated in t he development of a variety of online projects including "Bodies INCorporated" and SPEED: Technology, Media, Society, and "Life In the Universe with Stephen Hawking", a CD-ROM produced by the EAT Lab at UCSB, and MetaCreations. He is currently involved in an initiative to research multiuser visualization tools using Java and VRML.
5.1.2 University of Colorado - Faculty Researcher Rene Reitsma has joined the Colorado segment of the ADL evaluation team. Rene is a Research Associate at the Center for Advanced Decision Support for Water and Environmental Systems (CADSW ES) and Adjunct Faculty in the Geography Department at CU Boulder. Rene will work on statistical analysis of transaction log data, and spatial interaction modeling of use patterns for the Web interface.
6.0 Partnerships and Resources
6.1 Interactions with Partners
6.1.1 U.S. Navy - The Parallel Processing team has been working with Navy NRaD to develop a load balancer for the SWEB server on their CONVEX parallel machine. They also support us to develop the client-server partitioning techniques and Java based color image browsing.
6.1.2 Hughes Information Technology Systems - Xun Cheng design a relational database model for the gazetteer based on the proposed gazetteer content standard.
Thesaurus of feature types created to categorize gazetteer entries.
6.1.3. USGS - In March, Barbara Buttenfield spent a day at USGS Headquarters in Reston Virginia, working with Jay Donnelly reviewing USGS' plans for the interface design for the USGS National Atlas. She also met with John Moeller and Mike Domaratz of FGDC to discuss opportunities for funded research.
6.2 Additional Resources to Support Research and Development - Goodchild is PI of a three-year, $600,000 award from the National Imagery and Mapping Agency titled "Uncertainty in geospatial information representation". Co-PIs are Dan Montello an d Keith Clarke (UCSB) and Kate Beard (Maine). The research conducted under this award will help us to deal with the problems that arise when georeferenced data sets retrieved through ADL fail to register, or when methods must be found to describe uncertai nty effectively in ADL metadata.
A proposal to develop a spatial data infrastructure and digital library capability in Ecopetrol, the Colombian national petroleum company, was submitted in October. Funding will likely begin in September. It will support graduate students and Colombian researchers in Santa Barbara for a period of two years. The value of the proposal is approximately $270,000.
In addition, Linda Hill submitted a proposal to the FGCD to fund the development of an ingest system for a spatially-defined gazetteer based on the proposed content standard for gazetteer information.
7.0 ADL Presentations and Promotional Activities - A new initiative of the National Center for Geographic Information and Analysis, titled "Interoperating GISs", will bring many of the issues faced by ADL to the fore in a combined international con ference and workshop planned in Santa Barbara in December 1997. This effort is being conducted jointly with the Open GIS Consortium, and Terry Smith and Mike Goodchild are both participating in the organizing committees.
Plans for a book on user-centered evaluation to be co-edited with Nancy Van House continues We have added Ann Bishop as a co-editor, and decided to focus the third Allerton meeting (hosted by Illinois) specifically on the book. To accomplish this, we have moved the meeting dates back somewhat. The third Allerton meeting on User-Centered Library Evaluation Methods will be held in early March, 1998, at a University of Illinois retreat center in Monticello, Illinois. The meeting will comprise authors o f book chapters, to be invited. Chapter drafts will be produced prior to the meeting, whose goal will be to critique, discuss and revise writing in small and focused discussion. Barbara Buttenfield is a member of the Steering Committee which has organiz ed the conference.
The Image Processing team presented work at the NASA workshop on Information Systems, March 1997.
The Information Systems team presented the paper titled: "Scalable Access within the Context of Digital Libraries" at the IEEE forum on Research and Technology Advances in Digital Libraries., ADL 97, Library of Congress, Washington, D.C.
ADL contributed a content based search demo at the NSF exhibits center in Washington, DC.
Mary Larsgaard attended the Dublin Core 4 meeting in Canberra, also one day International Metadata Seminar; first meeting on 2-4, 2d on 5th; wrote "Dublin Core Coverage Element" strawman for the meeting; chaired committee after meeting that is doing "f inal" Coverage description. Mary also wrote "Spatial Data in Digital Form for Urban Planning" for a conference of the Council of Planning Libraries, San Diego; Larry Cruse (Map Librarian, UC San Diego at La Jolla) kindly presented the paper for Mary.
Mary Larsgaard also attended the Association of College and Research Libraries Conference, Nashville; 4 handouts prepared for spatial-data-in-digital-form halfday workshop at ACRL; a paper on Alexandria was prepared and presented at same conference on 4/13 - "The Alexandria Digital Library".
In the coming months, Barbara Buttenfield has been invited to teach a course on spatial data delivery at the University of Vienna, Austria. She will deliver several research colloquia during her residence. In Colorado, work will continue on the LTER pilot project, concerning granularity of metadata. As this work develops, it will be reported in quarterly reports.
Demonstrations of ADL were given to the following UCSB visitors:
3/97
Beth Sandore, Associate Professor and Coordinator for Imaging Projects; University of Illinois
Kimio Hosono, Professor, School of Library and Information Science; Keio University
Journal of Geography group
Orlie Brewer, Advanced Computing Technologist, Research and Technology Division, Boeing Information & Support Services
4/97
ADL Demo at DARPA
Nordic Metadata Project (BIBSYS, Norway; http://www.bibsys.no/english.html and Helsinki University Library - National Library of Finland; http://linnea.helsinki.fi/gabriel/en/countries/finland.html)
In addition, Mike Goodchild made the following presentations:
"NCGIA's Varenius Project: Advancing Geographic Information Science". University of Newcastle-upon-Tyne, May 1997.
"GIS, Spatial Analysis, and the Geographical Key". European Research Conference, "Socio-Economic Research and Geographic Information Systems". Il Ciocco, Italy, May 1997.
"The GIS Research Agenda". University College, London, Centre for Advanced Spatial Analysis and Department of Photogrammetry and Surveying, May 1997.
"From GIS to Geographical Information Science". London School of Economics, May 1997.
"Research and Development: What has been achieved so far?" and "Spatial
Information and Its Impact on Society". Conference: 10 Years After Chorley: The future for geographic information. The Royal Society, London, May 1997.
"Communicating the Results of Accuracy Assessment: Metadata, Digital Libraries, and Assessing Fitness for Use". First CASSINI International Workshop, Paris, April 1997.
"The Geolibrary". AGI Guest Lecture, GISRUK '97, Leeds, April 1997.
"Enabling Technologies and Educational Impacts: GIS". Atlantic
Institute Think Tank V, "Global Education Paradigms in Geomatics/Geoinformation", Paris, April 1997.
"Modern Geographic Information Systems and Model Linking". Wageningen Agricultural University, Netherlands, February 1997.
"Keynote: Geographic Data and the New Information Profession". ALISE, Washington DC, February 1997.
8.0 Publications
Andresen, D., Yang, T., Watson, D., Poulakidas, A., "Dynamic Processor Scheduling with Client Resources for Fast Multi-resolution WWW Image Browsing," Proceedings of the 11th International Parallel Processing Symposium (IPPS'97), Geneva, April, 1997.
Andresen, D., T. Yang, O. H. Ibarra, O. Egecioglu, "Adaptive Partitioning and Scheduling for Enhancing WWW Application Performance" Submitted to the Journal of Parallel and Distributed Computing (JPDC), 1997.
Beard, M.K., Smith, T.R., "A Framework for Meta-Information in Digital Libraries." In Managing Multimedia Data: Using Metadata to Integrate and Apply Digital Data. A. Sheth and W. Klaus (Eds.) McGraw Hill. (forthcoming).
Beard, M.K., Sharma, V., "Multidimensional Ranking in Digital Spatial Libraries." Special Issue of Metadata. Journal of Digital Libraries. (forthcoming).
Beard, K., Buttenfield, B.P., "Detecting and Evaluating Errors by Graphical Methods." Chapter: Goodchild, M.F. Maguire, D. and Rhind, D. (Eds.) Geographic Information Systems: Principles and Prospects. London: Longman, 1996.
Chandrasekaran, S., Manjunath, B.S., Wang, Y.F., Winkeler, J., Zhangk, H., "An Eigenspace update algorithm for image analysis," (CS TR, April 1996) to appear in CVGIP: Graphical models and image processing, 1997.
Dolin, R., Agrawal, D., El Abbadi, A., "Classifying Network Architectures for Locating Information Sources," Proceedings of the Fifth DASFAA Conference," Melbourne, Australia, April, 1997, 31-40.
Goodchild, M.F. and Buttenfield, B.P. National Research Council Mapping Science Committee, Sugarbaker, L.A., Chair, The Future of Spatial Data and Society: Summary of a Workshop. Washington, D.C.: National Academy Press, 1997.
Houston, A.L., Chen, H., Schatz, B.R., Sewell, R.R., Tolle, K.M., Doszkocs, T.E., Hubbard, S.M., Ng, D.T., "Exploring the Use of Concept Space, Category Map Techniques, and Natural Language Parsers to Improve Medical Information Retrieval." Submitted t o Decision Support Systems, Special Issue on Decision Support for Health Care in a New Information Age, 1997.
Leitner, M., Buttenfield, B.P., "Cartographic Guidelines for Visualizing Attribute Accuracy." Proceedings, AUTO-CARTO 13, Seattle, Washington, April, 1997.
Ma, W.Y., Manjunath, B.S., "Tools for Texture/Color Based Search of Images," SPIE Int. Conf. 3106, Human Vision and Electronic Imaging II, pp. 496-507, San Jose, CA, February, 1997.
Ma, W.Y., Manjunath, B.S., "A Texture Thesaurus for Browsing Large Aerial Photographs," to appear in the Journal of the American Society for Information Science, special issue on AI, 1997.
Orwig, R.E., Chen, H., Nunamaker, J.F., "A Graphical, Self-Organizing Approach to Classifying Electronic Meeting Output," Journal of the American Society for Information Science, Volume 48, Number 3, March, 1997.
Poulakidas, A.S., A. Srinivasan, O. Egecioglu, O. Ibarra, T. Yang, "A Compact Storage Scheme for Fast Wavelet-Based Subregion Retrieval," (3rd Annual International Computing and Combinatorics Conference), to appear, August 1997.
Prabhakar, S., D. Agrawal , A. El Abbadi and A. Singh "A Brief Survey of Tertiary Storage Sytems and Research" Proceedings of the 1997 ACM Symposium on Applied Computing (ACM SAC 97) Feb 28 - Mar 2, 1997, San Jose, CA.
Prabhakar, S., K. Abdel-Ghaffar, D. Agrawal and A. El Abbadi "Cyclic Allocation of Two-Dimensional Data" TechReport, Dept. of Computer Science, UCSB, TRCS97-08.
http://www.cs.ucsb.edu/TRs/techreports/TRCS97-08.ps To Appear in the 14th International Conference on Data Engineering, ICDE 98 , Orlando Florida, Feb 23-27 1998.
Skupin, A., Buttenfield, B.P., "Spatial Metaphors for Display of Information Spaces." Proceedings, AUTO-CARTO 13: 116-125, Seattle, Washington, April, 1997.
Tolle, K.M., Chen, H., Chow, H., "Applying Neural Networks to Pharmacokinetic/ Dynamic Data Sets," submitted to Decision Support Systems, Special Issue on Decision Support for Health Care in a New Information Age, 1997.
Unpublished Documents Available
Buttenfield, B. P. and Hill, L. "Report on the Second Alexandria Design Review Workshop." workshop held at UCSB 2/97, document on the Web at http://alexandria.sdc.ucsb.edu/~lhill
Plewe, B. Cartographic and GIS Representations of Gradation. Ph.D. Dissertation, Department of Geography, SUNY-Buffalo; defended 24 March 1997.
Prabhakar, S., D. Agrawal and A. El Abbadi "Impact of Media Exchanges on Robotic Libraries" TechReport, Dept. of Computer Science, UCSB, TRCS97-07.
http://www.cs.ucsb.edu/TRs/techreports/TRCS97-07.ps
I certify that to the best of my knowledge
Terence R. Smith
Director, Alexandria Digital Library Project