Computer Science Technical Report Research Articles

Cataloging Department Participation in Digital Initiatives

ABSTRACT The Cataloging Department at Carnegie Mellon University Libraries is involved in a number of digital projects. The libraries' projects to digitize out-of-copyright monographs, the Posner Family Collection, and computer science technical reports; to catalog electronic honors theses in the library management system; and to begin participation in the DLF/OCLC digital registry have all had an impact on the work of the department. We have established new procedures and workflows to accommodate the various project materials that enter and leave the department. Our successful participation in these projects depends on a number of factors–key among them is flexibility.

A framework for distributed digital object services

The following paper was written by the authors over a period of approximately 16 months during the period November 1993 to May 1995 in an attempt to explore a set of open research issues and to integrate them with certain ideas of a small group of researchers who had been briefed on the notions inherent in the Digital Object Architecture that one of the authors (Kahn) had been developing at the Corporation for National Research Initiatives (CNRI). This research group had been organized by CNRI as part of its "Computer Science Technical Reports" (CSTR) project that was funded by the Defense Advanced Research Projects Agency (DARPA). The CSTR project had several objectives, of which one was to digitize existing collections of technical reports from five leading computer science departments at universities in the U.S and make the reports available on the Internet. A second objective was to fund research at these institutions on digital libraries, especially research that made good use of their local collections. A third objective was to link heterogeneous electronic libraries, such as were being developed in the program. The difficulties inherent in this third objective ultimately led to this paper. In particular, the motivation arose from a set of concerns first expressed by these researchers at a meeting at CNRI on October 25, 1993 when the Digital Object Architecture was first presented to the research group by Kahn. In the discussion that resulted, many issues emerged, some of which we ultimately decided must be resolved in any architecture, and some of which-perhaps just as crucially-could be deferred for the time being. For example, questions arose about the role of semantics in identifiers; we decided to address them early on. Another issue that proved controversial was how to deal with situations in which the digital objects were actually mobile programs in the network. Since a digital object could contain other digital objects, this led directly to the notion of a mobile repository. The basic question was how best to describe this kind of situation and explain how to access information that was not at a known specific location or IP address on the net. We decided not to cover this aspect of the architecture in the interest of getting closure on the paper. A critical part of this effort was dealing with various intellectual property issues; and the architecture was designed to take this important aspect into account. We were fortunate to have the help of Ms. Patrice Lyons, an experienced intellectual property lawyer, in the formulation of certain key elements of the architecture dealing with terms and conditions for access to information, and in the drafting of the paper. Although her name does not appear on the paper, her insights and contributions played an important role in the preparation of the paper. One key component of the architecture is a general-purpose resolution system, known as the Handle System®, an implementation of which has been operational on the Internet since 1994. At present, close to 50 million identifiers are assigned to digital objects and resolvable by the system, which has been available 7x24 for many years. It is anticipated that the number of resolvable identifiers in the system will grow rapidly and could easily exceed a billion in the not too distant future.

GreedyDual ∗ Web caching algorithm: exploiting the two sources of temporal locality in Web request streams

The relative importance of long-term popularity and short-term temporal correlation of references for Web cache replacement policies has not been studied thoroughly. This is partially due to the lack of accurate characterization of temporal locality that enables the identification of the relative strengths of these two sources of temporal locality in a reference stream. In [ACM Sigmetrics’00, June, 2000 (to appear), Computer Science Technical Report BUCS1999-014, Boston University], we have proposed such a metric and have shown that Web reference streams differ significantly in the prevalence of these two sources of temporal locality. These findings underscore the importance of a Web caching strategy that can adapt in a dynamic fashion to the prevalence of these two sources of temporal locality. In this paper, we propose a novel cache replacement algorithm, GreedyDual ∗, which is a generalization of GreedyDual-Size. GreedyDual ∗ uses the metrics proposed in [ACM Sigmetrics’00, June, 2000 (to appear), Computer Science Technical Report BUCS1999-014, Boston University] to adjust the relative worth of long-term popularity versus short-term temporal correlation of references. Our trace-driven simulation experiments show the superior performance of GreedyDual ∗ when compared to other Web cache replacement policies proposed in the literature.

A transaction log analysis of a digital library

As experimental digital library testbeds gain wider acceptance and develop significant user bases, it becomes important to investigate the ways in which users interact with the systems in practice. Transaction logs are one source of usage information, and the information on user behavior can be culled from them both automatically (through calculation of summary statistics) and manually (by examining query strings for semantic clues on search motivations and searching strategy). We have conducted a transaction log analysis on user activity in the Computer Science Technical Reports Collection of the New Zealand Digital Library, and report insights gained and identify resulting search interface design issues. Specifically, we present the user demographics available with our library, discuss the use of operators and search options in queries, and examine patterns in query construction and refinement. We also describe common mistakes in searching, and examine the distribution of query terms appearing in the logs.

Content locality in distributed digital libraries

In this paper we introduce the notion of content locality in distributed document collections. Content locality is the degree to which content-similar documents are colocated in a distributed collection. We propose two metrics for measurement of content locality, one based on topic signatures and the other based on collection statistics. We provide derivations and analysis of both metrics and use them to measure the content locality in two kinds of document collections, the well-known TREC corpus and the Networked Computer Science Technical Report Library (NCSTRL), an operational digital library. We also show that content locality can be thought of temporally as well as spatially and provide evidence of its existence in temporally ordered document collections like news feeds.

Application of computer algebra techniques to affinity binding equations

In this paper we apply computer algebra techniques to obtain information on the influence of free ligand concentration on the capacity of an affinity chromatography matrix when the protein to be absorbed shows cooperative binding. In the case of a dimeric protein it was derived, in particular, an explicit expression for the (total) concentration of soluble ligand that results in maximum binding to the matrix [J.P. Benett et al., Bath Computer Science Technical Report 1990, 90-44]. In the case of a tetrameric protein, we derive an expression that relates the yield to the parameters of the system and solves it numerically.

The Internet for Sociologists

s and downloadable article preprints is that of the high-energy physics community (as well as some other branches of physics and mathematics) hosted at the Los Alamos National Laboratory (see http://xxx/lanl.gov). Computer scientists have organized a searchable index of the technical report series of over 70 major academic computer science departments in North America and western Europe-the Networked Computer Science Technical Reports Library (http:// www ncstrl.org/Dienst/htdocs/lnfo). . . . . . . . . . Hew SOC10 LOglStS partlclpate m 1rlstltutlona L Ly organized arrangements for sharing printed preprints or working papers, although many sociologists share their working manuscripts ancl preprints informally with selected colleagues. Those sociologists, like David Zaret, who want to make these documents in electronic form do so on their own web pages. Scholats are also finding that the WWW can be useful for increasing the readership of theit published work. For example, Howie Becker has organized his WWW page (http:/ /weber.u.washington.edu/ hbecker/) to include articles that he has published in an eclectic set of specialty journals ancl books. Anyone interestecl in following Becker's writing would not have been likely to locate many of these articles by relying upon disciplinary journal indicess such as Sociofiler Electroliic jourllals: tables of colltellts alld abstracts illdices Some editors of journals and books are creating on-line archives of tables of contents and abstracts that may be browsed or searched. For exallple, the Anmlal Review of Soaology has organized the last 12 years of tables of contents ancl the last several years of abstracts for on-line review (http:// www. annurev.org/soc/holne. htln) . However, the full text of luaterials is available in paper foml, and readers must obtain them via routine Ineans. These decentralized inclices are more culnbersonle for comprehensive searching than agglomerated indices such as Sociofile; but they allow more convenient browsitlg of specific publications. In addition, they may offer other kinds of value. For example, the editors of Socilogical MeSolon gnttp:/ /weber. u.washington.edu/ soctneth2/) offer abstracts andfree copies ofthesoftware that autlaors mention in their book chapters. Aere are a huge variety of ways to organize electronic publicatior s or electronic supplements to paper publications. Journal editors ancl otlaers who are interested in expanding the range of materials in electronic fonnats coulcl usefully examine various strategies triecl in other fields. For example, there is a huge collection of medical and bioscience journals that illustrate various electronics andfree copies ofthesoftware that autlaors mention in their book chapters. Aere are a huge variety of ways to organize electronic publicatior s or electronic supplements to paper publications. Journal editors ancl otlaers who are interested in expanding the range of materials in electronic fonnats coulcl usefully examine various strategies triecl in other fields. For example, there is a huge collection of medical and bioscience journals that illustrate various electronic fonnats collected on MedWeb (http:/hAzww./gen./ elnory. edu/MEDWEB/alphakey/Electronic SublicationKs.htlnl) . There LS an i nt e res t ing s e t This content downloaded from 157.55.39.96 on Sat, 03 Sep 2016 04:05:38 UTC All use subject to http://about.jstor.org/terms

Technical Report Literature on the World Wide Web

This article reviews and analyzes the current state of Web access to technical report literature. Besides locating sites for bibliographically identifying technical reports, it also looks at the ways the full-text of technical reports are being distributed. Where appropriate, it compares the electronic bibliographic control products with their print or formerly print equivalents. Most widely available at this point are federal government reports and computer science technical reports. Information on coping with the various file formats found at the sites is also provided. URLs to the various technical report sites and to sites dealing with the software applications needed to access the reports are included.

The New Zealand Digital Library Project

The New Zealand Digital Library project is a research programme in the Computer Science Department at Waikato University, whose aim is not to set up new libraries but to develop the underlying technology for digital libraries and make it available publicly so that others can use it to create their own collections. We are concerned with large collections of electronic, predominantly textual, documents, physically dispersed on computers the world over, and aim to make them accessible through a uniform interface that allows information to be located and accessed. This article describes the NZDL project, illustrated by a large example collection of Computer Science Technical Reports. The system is freely available on the World-Wide Web. The migration of information from paper to computers promises to change the whole nature of research, and in particular the methods by which people locate information. The goal of the New Zealand Digital Library project is to explore the potential of Internet-based digital libraries, by which we mean large collections of electronic, predominantly textual, documents, physically dispersed on computers the world over, which are accessible through a uniform interface that allows information to be located and accessed. Our vision is to develop systems that automatically impose structure on fundamentally anarchic, uncatalogued, distributed repositories of information, thereby providing users with effective tools to locate the information they need and peruse it conveniently and comfortably. As a geographically isolated but technologically advanced nation, New Zealand stands to gain markedly from effective deployment of information resources that are freely available on international computer networks. The New Zealand Digital Library project is a research programme in the Computer Science Department at Waikato University, funded in part by the NZ Foundation for Research, Science and Technology and the NZ Lotteries Grants Board. Our aim is not to set up new libraries, but to develop the underlying technology for digital libraries and make it available publicly so that others can use it to create their own collections. Not surprisingly, the technology required varies greatly depending on the kind of collection and the source of the information. Consequently, we are making several different large collections of public-domain text available on an experimental basis as test cases. These allow us to investigate the technical problems of gathering and indexing the material, to assess the useability of our interfaces, and to collect information on external usage so that we can improve the facilities offered. We call this testbed the “NZDL” to distinguish it from our research programme. This article begins by presenting the facilities offered by the NZDL. We then discuss the indexing and retrieval strategies that it supports, and go on to explain how the collection is built and maintained. Finally we indicate likely directions that our project will be taking in the future.

Ensuring Retrieval Effectiveness in Distributed Digital Libraries

We find that dissemination of collection-wide information (CWI) in a distributed collection of documents is needed to achieve retrieval effectiveness comparable to that of a centralized collection. Complete dissemination is unnecessary. The required dissemination level depends upon the content-skew of the distributed collection, i.e., how documents are allocated among sites. Low dissemination is needed for random document allocation, but higher levels are needed when documents are allocated on the basis of their content. We define parameters to control dissemination and document allocation and present results from four document collections. These results provide insight into the necessary technology underlying digital libraries. We also describe the architecture of the Networked Computer Science Technical Report Library (NCSTRL), a concrete example of a system that fits our model of a distributed archive.

Creating a Networked Computer Science Technical Report Library

Creating a Networked Computer Science Technical Report Library

Wide Area Technical Report Service

Wide Area Technical Report Service (WATERS) is a distributed database of computer science technical reports. Contributors are departments of computer science that make their reports, stored locally at their sites, available through World-Wide Web and a WAIS search engine. By using WATERS, anyone with access to the Internet can, using a client such as Mosaic browse, search, obtain abstract and bibliographic information, and retrieve technical reports online. The latter assumes that the user's WWW client is set up to launch viewers such as ghostview for Postscript files, xdvi for dvi files, and xtiff for TIFF page images.

Dienst: an architecture for distributed document libraries

As one of the five universities participating in the ARPA-sponsored Computer Science Technical Report project, we at Cornell have developed a digital library architecture called Dienst. Dienst is a protocol and implementation that provides Internet access to a distributed, decentralized multi-format document collection. The collection is managed by a set of interoperating Dienst servers distributed over the Internet. These servers provide three digital library services: repositories of multi-format documents; indexes into the document collection and search engines for these indexes; and user interfaces for browsing, searching, and accessing the collection.

Generalized Quantifiers and Logical Reducibilities

We consider extensions of first order logic (FO) and least fixed point logic (LFP) with generalized quantifiers in the sense of Lindstrom [Lin66]. We show that adding a finite set of such quantifiers to LFP fails to capture all polynomial time properties of structures, even over a fixed signature. We show that this strengthens results in [Hel92] and [KV92a]. We also consider certain regular infinite sets of Lindstrom quantifiers, which correspond to a natural notion of logical reducibility. We show that if there is any recursively enumerable set of quantifiers that can be added to FO (or LFP) to capture P, then there is one with strong uniformity conditions. This is established through a general result, linking the existence of complete problems for complexity classes with respect to the first order translations of [Imm87] or the elementary reductions of [LG77] with the existence of recursive index sets for these classes. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-92-85. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/348 Generalized Quantifiers and Logical Reducibilities MS-CIS-92-85 LINC LAB 240

Wide area technical report service—technical reports online

Wide Area Technical Report Service (WATERS) is a distributed database of computer science technical reports. Contributors are departments of computer science that make their reports, stored locally at their sites, available through World-Wide Web and a WAIS search engine. By using WATERS, anyone with access to the Internet can, using a client such as Mosaic browse, search, obtain abstract and bibliographic information, and retrieve technical reports online. The latter assumes that the user's WWW client is set up to launch viewers such as ghostview for Postscript files, xdvi for dvi files, and xtiff for TIFF page images.

VP Ellipsis and Semantic Identity

No abstract.

Analysis of dynamic congestion control protocols

We present an approximate analysis of a queue with dynamically changing input rates that are based on implicit or explicit feedback. This is motivated by recent proposals for adaptive congestion control algorithms [RaJa 88, Jac 88], where the sender's window size at the transport level is adjusted based on perceived congestion level of a bottleneck node. We develop an analysis methodology for a simplified system; yet it is powerful enough to answer the important questions regarding stability, convergence (or oscillations), fairness and the significant effect that delayed feedback plays on performance. Specifically, we find that, in the absence of feedback delay, the linear increase/exponential decrease algorithm of Jacobson and Ramakrishnan-Jain [ Jac 88, RaJa 88] is provably stable and fair. Delayed feedback on the other hand, introduces oscillations for every individual user as well as unfairness across those competing for the same resource. While the simulation study of Zhang [Zha 89] and the fluid-approximation study of Bolot and Shanker [BoSh 90] have observed the oscillations in cumulative queue length and measurements by Jacobson [ Jac 88] have revealed some of the unfairness properties, the reasons for these have not been identified. We identify quantitatively the cause of these effects, via-a-vis the systems parameters and properties of the algorithm used. The model presented is fairly general and can be applied to evaluate the performance of a wide range of feedback control schemes. It is an extension of the classical Fokker-Planck equation. Therefore, it addresses traffic viability (to some extent) that fluid approximation techniques do not address. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-91-18. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/385 Analysis Of Dynamic Congestion Control Protocols: A Fokker-Planck Approximation MS-CIS-91-18 DISTRIBUTED SYSTEMS LAB 5 Amarnath Mukherjee John C. Strikwerda Department of Computer and Information Science School of Engineering and Applied Science University of Pennsylvania Philadelphia, PA 19104-6389

Structure and Intonation

Rules for assigning phrasal intonation to sentences are often assumed to require an autonomous level of structure, distinct from what is usually thought of as surface syntactic structure. The present paper argues that the requisite notion of structure can be subsumed under the generalised notion of surface structure that emerges from the combinatory extension of Categorial Grammar. According to this theory, the syntactic structures and the intonational structures of English are one, and can be captured in a single unified grammar. The interpretations that the grammar provides for such constituents correspond to the entities and open propositions that are concerned in certain discourse-related aspects of intonational meaning that have variously been described as theme and rheme, given and new, or presupposition and focus. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-90-45. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/571 Structure and Intonation MS-CIS-90-45 LINC LAB 174

Gapping as constituent coordination

A number of coordinate constructions in natural languages conjoin sequences which do not appear to correspond to syntactic constituents in the traditional sense. One striking instance of the phenomenon is afforded by the gapping construction of English, of which the following sentence is a simple example: (1) Harry eats beans, and Fred, potatoes Since all theories agree that coordination must in fact be an operation upon constituents, most of them have dealt with the apparent paradox presented by such constructions by supposing that such sequences as the right conjunct in the above example, Fred, potatoes, should be treated in the grammar as traditional constituents, of type S, but with pieces missing or deleted. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-89-76. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/820 Gapping As Constituent Coordination MS-CIS-89-76 LINC LAB 162

Overview of the University of Pennsylvania CORE system

The CORE System is a proposed standard for a device-independent graphics system. The concept of a deviceindependent system was first developed in 1977 by the Graphics Standards Planning Committee (GSPC) of ACM Siggraph and later refined in 1979 [1,2]. The CORE System design has received favorable reviews and has been implemented by various vendors at several universities, and other computing facilities (e.g. [3,7]). The main objectives of the CORE System are to provide uniformity, compatibility, and flexibility in graphics software. Three advantages that the CORE system provides over non-standard graphics systems are device independence, program portability, and functional completeness. A large number of different graphics hardware devices currently exist with a wide range of available functions. The CORE System provides device independence by shielding the applications programmer from specific hardware characteristics. The shielding is at the functional level: the device-independent (DI) system uses internal routines to convert the application programmer's functional commands to specific commands for the selected hardware device driver (DD). The progammer describes a graphical world to the CORE System in device-independent normalized device coordinates. The programmer also specifies the viewport on the logical view surface (output device) where a picture segment is to be placed. As the CORE System becomes the accepted standard graphics package, program portability will become more feasible. Program portability means the ability to transport application programs between two sites without requiring structural modifications. The CORE System was designed for functional completeness so that any graphics function a programmer desires is either included within the system or can be easily built on top of CORE System routines. Disciplines Computer Engineering | Computer Sciences Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-82-17. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/1004