Digital Libraries: International Collaboration and Cross-Fertilization

Archiving Web content is an important topic for digital libraries and especially for deposit libraries. Web archiving systems usually collect Web resources using search robot software and/or by human labor. However, these resource gathering methods have disadvantages: for example, it is difficult to collect all historical versions of a resource or to collect hidden or dynamic resources. This paper proposes a Web archiving system which is designed to collect resources in accordance with a resource archiving policy determined by the person or organization which provides the resources on the Web. This paper presents the model of the Web archiving system and a prototype system implemented based on the model.

CALIS Electronic Thesis and Dissertation Digital Library (CALIS-ETD-DL) is a distributed heterogeneous digital library system funded by CALIS with goals to share the ETD resource among the universities in China and speed up knowledge spreading. It is a prototype system based on the advanced digital library technology and open standards. This paper focuses on the following issues: 1) architecture of CALIS-ETD-DL and its implementation that is centralized metadata repository and distributed digital object; 2) interoperability protocol and standard interface used in CALIS-ETD-DL including international widely used protocol such as OAI, METS, OPENURL, Web Service etc. and self-defined protocol such as ODL-MAP protocol etc; 3) ETD Digital Right Management System for ETD copyright protection; 4) CALIS-ETD metadata specification for describing and administrating ETD resource; 5) ETD persistent identifier and long-term preservation format of ETD. Conclusions are drawn and future works of CALIS-ETD-DL are introduced in the end.

The digitization of library collection is one of the most important tasks for electronic collection development of a library. Digital collection development tends to be more complex than electronic resource licensing and subscription. There are many factors to be considered in the digitization of collection, such as selection policy, organization, access, preservation, funding, and so forth. In this article, the authors discuss some issues associated with collection digitization, including copyright control, intellectual nature, added value, users and their needs, human resource management, technical requirement, funding, and digital content preservation. Several collection digitization principles emerge from the discussion, with respect to standardization and specification, extensibility and integration, security and reliability, preservation and service, collaboration and sharing. as well as evaluation of the digital collection. Finally, the theoretical discussions are further demonstrated by using digital dissertation development at Zhongshan University Libraries, P.R. China, as a pragmatic instance.

This toolkit is designed for the Chinese Mathematics Digital Library (CMDL) System of Tsinghua University. The aim of CMDL is to preserve digitalized Chinese Mathematics ancient books available and provide mathematics research and education service to the public. This toolkit fulfills the cataloging and preservation functions of CMDL. METS encoded documents are used as the final storage format of metadata, including descriptive metadata, structural metadata and administrative metadata. The final XML documents with original digital resources can be packaged into an electronic book record extended from OEB, and submitted to a management system based on Fedora Open Source System. This software is a pure-Java application for platform-portable consideration, and JAXB is used to process XML for its convenient and powerful performance at XML processing.

In order to support various requirements from the user’s perspective, digital library (DL) systems may need to apply a large variety of services, such as query services for a specific DL, mapping services for mapping and integrating heterogeneous metadata records, or query modification and expansion services for retrieving additional relevant documents. This paper focuses on exploiting an extended Service-Oriented Architecture – Peer-based SOA (PSOA) for DL development with the goal of alleviating the weaknesses in the basic SOA infrastructure, especially in the aspects of scalability and interoperability. We also present our work in how to combine the Semantic Web and Web Services together to support interoperability over heterogeneous library services. A query service example is also presented.

This paper presents a part of the research result, according to the research of the International Digital Library Project (IDLP)— CMNet (Chinese Memory Net – US-Sino Collaborative Research Toward A Global Digital Library in Chinese Studies) founded by the America. The Open Archives Initiative (OAI) Protocol for Metadata Harvesting presents one promising method by which metadata regarding Chinese Culture resources can be shared and made more interoperable with metadata from other sources. In this international cooperative project, Shanghai Jiaotong University mainly undertakes the research of key theories, methods and technologies for the Chinese folk music DL and the major content of the research are digitalization processing methods for the Chinese folk music, metadata standards and criteria, metadata interoperability for the Chinese folk music and so on. This paper introduces the OAI-based metadata interoperation framework for Chinese cultures resources, OAI-based system framework for CMNet, OAI protocol for metadata harvesting, administration and value-added services of metadata harvesting for Chinese Folk Music DL (CFMDL) and so on.

With an increasing amount of information being digitized or directly created and subsequently existing only electronically, and coupled with an ever increasing variety of file formats and integrated document functionalities, long-term preservation solutions become crucial. While different approaches, such as Emulation, Migration, or Computer Museums were developed, neither of them excels in all circumstances, and the selection of the most appropriate strategy poses a non-trivial task. In this paper, an adapted version of Utility Analysis is presented, which can be used for choosing an optimal preservation solution for each individual situation. This analysis method comes from infrastructure projects and is here used to combine the wide range of requirements which are to be considered in order to choose a suitable preservation strategy. The evaluation metric will be presented and demonstrated with the help of two practical examples.

Based on the analysis of the characteristics of both OPAC and e-learning resources respectively, some key issues of sharing and integration, such as data exchange format, searching interface, and supporting platform, were discussed, and a total model for integration of the two resources was proposed. Moreover, a prototype of sharing search system named Fsearch was developed, with which users can search for both OPAC system and e-learning system simultaneously. The result shows that Fsearch system can effectively improve the efficiency of acquiring knowledge for users and bring forward a preferable application prospect.

The China-US Million Book Digital Library Project (MBP) is a cooperated project of universities and institutes in China and USA, with funding from the Ministry of Education of China (MOE) and National Science Foundation of USA (NSF). The objective of MBP is to create a free-to-read, searchable collection of one million books, available to everyone over the Internet. This task will be accomplished by scanning the books and indexing their full text. Based on the plan and experience in MBP, six technical issues are presented in this paper, such as content selection, copyright clearance, data production, metadata creation, digital preservation and access, and quality control. The current status of the project and the future plan are discussed.

Academies are culture educational institutions, where ancient Chinese scholars carry out various activities concerning books which include book storing, reading, teaching, lecturing, carving to accumulate, create, and impart culture. Relying on our cultural background of the academy of classical learning, we utilize the extant original academic literatures and consult national and international scholars’ research on Chinese traditional culture of the academic of classic learning to construct Academy Culture Database, characterized by the distinctively Chinese traditional culture. This database is also built on the basis of modern information technique and complied with the international universal format of the database. We aim at revealing the essence of Chinese traditional culture to the world through the modern information processing technique and digital measures, as well as providing a convenient public research platform for home and abroad experts of academy culture studies, education history studies, Confucian studies and of culture history studies.

To be in line with The Chinese University of Hong Kong’s (CUHK) mission of assisting in the preservation, creation, application and dissemination of knowledge, the CUHK online institutional repository, known as the ‘SIR’ (Scholarly Information Repository), is designed to serve the University community and its constituent four Colleges, the people of the Hong Kong Special Administrative Region (SAR) and China, and academics and researchers in the wider world. This paper reviews the planning and implementation of the CUHK SIR model. Discussions on what materials needed to be included in the SIR database are highlighted – for example, not only preprints, published articles, theses and dissertations, but also University and College publications, audio and video clips, etc. The paper details certain system design parameters – it runs on DSpace, and a simple interface is available for faculty to submit citations of their publications and other documents for digitization and inclusion. Finally, the challenges facing the future growth and use of the SIR are noted – not least in securing both the intellectual property rights and the financial and human resources to sustain and develop a major online initiative of this kind.

Semantic architecture is crucial for a digital library application especially in a distributed system environment. It provides various approaches to overcome semantic interoperability problems and usually consists of metadata solution with open system architecture. The design of the digital library system for the China Cultural Celebrities’ Manuscripts Library (CCCML), which is a branch of Shanghai Library, has taken into account a lot of the main aspects from the requirement of semantics, including the metadata profiles, encoding consistence, authority control, ontology functioning, semantic integration, etc. We argue that it is very important to establish an articulated layered semantic architecture for digital libraries in the semantic web environment. And it becomes more and more clear that the semantic services can be settled with the Semantic Web Services technologies, which is supported by and consisted of a wide range of standards and protocols. And a lot of mainstream interoperability architecture, such as OAI, OpenURL etc., can be conformed or implemented by Semantic Web Services. This paper gives some major considerations and overviews on the design of semantic architecture for CCCML, which shows a lot of similarity in typical digital library systems.

In order to help researchers of humanities study the cultural ancient resources from temporal and geographical perspectives, we build WebGIS-RBDL at Peking University. In WebGIS-RBDL, metadata records and historical map layers are linked together, and therefore users can get more spatio-temporal information about rare books through folded map layers, including historical maps, modern maps and landform images. In this paper, the architecture of WebGIS-RBDL is introduced. Then, the principles, rules and methods for metadata extracting and correlating in WebGIS-RBDL are discussed. Furthermore, efficient tools to extract spatio-temporal information automatically from rare-book metadata are developed. The WebGIS-RBDL can be extended to include other ancient collections easily. Our efforts show that using GIS in digital libraries is a new and promising method to build more effective user-friendly interfaces in digital libraries.

The results from a questionnaire survey distributed to Australian institutions hosting digital collections, are presented. It was found that few metadata schemas are being used, although several others are being contemplated; that content standards tend to follow format standards; and that in-house standards are often required to supplement these. A fairly low level of resource and metadata exchange is taking place. Standards were considered important for a variety of reasons, a leading reason being interoperability. However, when it came to practice and new initiatives, there was a range of approaches taken to achieve interoperability. Some institutions concentrate on interoperability across their own systems, others focus on cross-institutional projects. Some institutions stress adherence to metadata standards, others are implementing new technologies that can handle different metadata formats and content.

Text categorization is a crucial task of increasing importance. Our work focuses on the study of Chinese text categorization on the basis of Boosting model. We chose the People’s Daily news from TREC5 as our benchmark datasets. A minor modification to AdaBoost algorithm (Freund and Schapire, 1996, 2000) was applied for this hypothesis. By way of using the F1 measure for its final evaluation, the results of the Boosting model (AdaBoost.MH) is proved to be effective and outperforms most of other algorithms reported for Chinese text categorization.

We address the problem of integrating documents from a source catalog into a master catalog. Current technologies for solving the problem deem it as a flat category integration problem without considering the useful hierarchy information in the catalog, or deal with it hierarchically but without a rigorous model. In contrast, our method is based on correctly identifying relationships among categories, such as Match, Disjoint, SubConcept, SuperConcept, and Overlap, which come from the relations of sets in Set theory. Compared with traditional Match/NotMatch relationship in literature, our approach is more expressive in defining the relationship. The relationships among categories are first learned in a probabilistic way, and then refined by considering the hierarchy context. Our preliminary experiments show that it can help to correctly identify category relationships, and thus increase the accuracy of document integration.

The 21st century has witnessed an overwhelming research output on GSI but due to publication scattering and want of organized database, the outcome of these publications have not been visible to the policy makers and gender scientists. Further, the traditional document management system put a heavy burden on the user in the process of information access, extraction and interpretation. Besides, the absence of domain-specific semantic index in the discipline GSI results in irrelevant retrieval of records resulting in high recall but low precision. In addition, the existing system makes the process of literature-search difficult and time consuming. The exponential growth of literature, the information overload, and the time wasted in searching and browsing in the field of GSI demanded a semantic information processing system. This situation calls for the development of Knowledge Management System (KMS) in GSI by building dynamic (i) knowledge repositories, (ii) semantic index and (iii) knowledge maps and thereby alleviates the problems encountered by the user in the traditional document management system.

The conventional design of digital repository systems provides little flexibility and is often difficult to scale and customize. The authors analyzed the causes for inflexibility and difficulties in managing digital repository systems and proposed an ontology-based design in managing digital repository systems in this paper. The ontology focuses on the system controls that manage a wide array of functions such as object data, user access control, and rights management. The authors used an instance in the Object class to demonstrate how the ontology can be used to reduce the difficulties in conventional design of repository systems.

In recent years, we have seen huge volumes of research papers available on the World Wide Web. Metadata provides a good approach for organizing and retrieving these useful resources. Accordingly, automatic extraction of metadata from these papers and their bibliographies is meaningful and has been widely studied. In this paper, we utilize a bigram HMM (Hidden Markov Model) for automatic extraction of metadata (i.e. title, author, date, journal, pages, etc.) from bibliographies with various styles. Different from the traditional HMM, which only uses word frequency, this model also considers both words’ bigram sequential relation and position information in text fields. We have evaluated the model on a real corpus downloaded from Web and compared it with other methods. Experiments show that the bigram HMM yields the best result and seem to be the most promising candidate for metadata extraction of bibliographies.

We describe the motivation, design and implementation of an infrastructure to support metadata quality assessment within a specialised Open Archives Initiative (OAI) sub-domain, the Open Language Archives Community (OLAC). While services for structural validation of metadata are widely used, there is little corresponding work regarding services which evaluate the semantic and syntactic content of metadata from a qualitative perspective. We posit that any measure of metadata quality benefits from both contextual and referential assessment – metadata on a per record and per collection basis is legitimately assessed against the baseline of broader community practice, as well as for compliance to any external standard. In this paper we describe the implementation of a metadata quality assessment scheme, and the corresponding interfaces to the evaluation tool.

The traditional methods of feature selection and weighting make the best of document information, but despise or ignore the category information. The new feature selection and weighting methods use category information as a factor, which make up the disadvantages of traditional methods. Using new methods, the features distributed equally on a single category are more important than using old methods. It is proved by the experiment that four famous classifiers based on new feature selection and weighting methods are more effective than those based on traditional methods.

Metadata repositories are loosely-controlled environments where each discrete metadata dataset retains its independent identity, multiple standards were applied, and records were contributed on a voluntary-based system. As more integrated digital libraries take such an approach in order to provide comprehensive collections, quality of the metadata records describing these collections’ items has drawn increased attention, since they function as a ‘window’ of a digital library collection and its services and directly involve in the success or failure of any information seeing performance. This poster reports the preliminary findings from a project funded by the National Science Foundation to study the quality of the metadata records in the National Science Digital Library (NSDL)’s Metadata Repository. (Figure 1. Project work space).

Discusses multilingual and culture-dependent metadata issues affecting the GREEN digital library and proposes a robust approach to internationalizing digital library collections.

Content-based music information retrieval and theme mining are two key problems in digital music libraries, where “themes” mean the longest repeating patterns in a piece of music. However, most data structures constructed for retrieving music data can not be efficiently used to mine the themes of music pieces, and vice versa. The suffix tree structure can be used for both functions, nevertheless its height is too large and its maintenance is somewhat difficult. In this paper, a kind of index structure is introduced, which adopts the idea of inverted files and that of

N

-gram. It can be used to retrieve music data as well as to mine music themes. Based on the index and several useful concepts, a theme mining algorithm is proposed. Also, two implementations of a content-based music information retrieval algorithm are presented. Experiments show the correctness and efficiency of the proposed index and algorithms.

This paper presents our implementation techniques for an intelligent Web image search engine. A reference architecture of the system is provided and addressed in this paper. The system includes several components such as a crawler, a preprocessor, a semantic extractor, an indexer, a knowledge learner and a query engine. The crawler traverses web sites in multithread accesses model. And it can dynamically control its access load to a Web server based on the corresponding capacity of the local system. The preprocessor is used to clean and normalize the information resource downloaded from Web sites. In this process, stop-word removing and word stemming are applied to the raw resources. The semantic extractor derives Web image semantics by partitioning combining the associated text. The indexer of the system creates and maintains inverted indices with relational model. Our knowledge learner is designed to automatically acquire knowledge from users’ query activities. Finally, the query engine delivers search results in two phases in order to mine out the users’ feedbacks.

As a result of the recent information explosion, there is an increasing demand for automatic summarization, and human abstractors often synthesize summaries that are based on sentences that have been extracted by machine. However, the quality of machine-generated summaries is not high. As a special application of information retrieval systems, the precision of automatic summarization can be improved by user relevance feedback, in which the human abstractor can direct the sentence extraction process and useful information can be retrieved efficiently. Automatic summarization with relevance feedback is a helpful tool to assist professional abstractors in generating summaries, and in this work we propose a relevance feedback model for fractal summarization. The results of the experiment show that relevance feedback effectively improves the performance of automatic fractal summarization.

Searching digital images on a networked environment is rapidly growing. Despite recent advances in image retrieval technologies, high-precision and robust solutions remain hampered by limits to knowledge about user issues associated with image retrieval. This paper examines a large number of queries from a Web image search engine, and attempts to develop an analytic model to investigate their implications for image retrieval technologies. The model employs the concepts of uniqueness and refinement to categorize successful and failed queries. The results show that image requests have a higher specificity and may often contain queries refined by interpretive, reactive, and perceptual attributes. Based on the proposed model, the study further investigates feasible technical solutions integrating both content-based and concept-based technologies to deal with real image query types. The initial study has provided useful results that enhance the understanding of digital image searching and suggests implications for the improvement of image retrieval systems.

A robust character region identification approach is proposed here to deal with cover images using a differential top-hat transformation (DTT). The DTT is derived from morphological top-hat transformation (TT), and efficient for feature identification. This research is considered as a fundamental study for auto-classification of printed documents for organizing a Digital Library (DL) system. The entire procedure can be divided into two steps: region classification and character region identification. In the first step, a source gray image is segmented by a series of structuring elements (SE) into sub-images using the DTT. Since the widths of regions are relative to the scales of the characters, the different scales of characters are classified into the series of sub-images. The character region identification processing is composed of feature emphasis, extraction of candidate character regions and region reconstruction processing. Feature emphasis processing reduces noises and emphasizes characters in the sub-images, and then the candidate character regions are extracted from the gray scale sub-images by a histogram analysis. Lastly, a morphological image reconstruction algorithm based on conditional dilation is introduced to make the extracted character regions distinct from noises. To demonstrate the robustness of the proposed approach, 30 gray scale cover images were tested in the experiments, which revealed that an average extraction rate of 94% has been achieved.

In this paper, we propose a novel approach for multilingual story link detection. Our approach uses features such as timelines and multilingual spaces for giving distinctive weights to terms that constitute linguistic representation of events. On timelines term significance is calculated by comparing term distribution of the documents on a day with that of the total document collection. Since two languages can provide more information than one language, term significance is measured on each language space, which is then used as a bridge between two languages on multilingual (here bilingual) spaces. Evaluating the method in Korean and Japanese news articles, our method achieved 14.3% improvement for monolingual story pairs, and 16.7% improvement for multilingual story pairs. By measuring the space density, the proposed weighting components are verified with a high density of the intra-event stories and a low density of the inter-events stories. This result indicates that the proposed method is helpful for multilingual story link detection.

The need for fast and vast dissemination of research results has led a new trend such that more number of authors post their documents to personal or group Web spaces so that others can easily access and download them. Similarly, more and more researchers use online search for accessing documents of interest in Web, instead of paying a visit to libraries. Currently, to locate and download an online copy of a particular document

D

, one typically (1) uses Search Engines with the citation information and browses through returned web pages (e.g., author’s homepage) to see if any contains

D

, or (2) uses searching facilities of an individual Digital Library (e.g., CiteSeer, e-Print) looking for

D

, and if not found, repeats the search in another Digital Library. However, the scheme (1) involves human browsing to get to the final online copy, while the scheme (2) suffers from incomplete coverage. To remedy these shortcomings, in this paper, we present a system, named as

PaSE

, which can effectively locate online copies (e.g., PDF or PS) of scientific documents using citation information. We consider a myriad of alternatives in crawling and parsing the Web to arrive at the right document quickly, and present a preliminary experimental study. Using some of the best alternatives that we have identified, we show that PaSE can locate online copy of documents more accurately and conveniently than human users would do at the cost of elongated search time.

A data model supporting the unification of the management of temporal XML document databases and version control, i.e., a

Temporal Version Data Model (TVDM)

, is proposed. TVDM is temporally change-centric and represents the change between two consecutive versions by means of

temporal delta (tDelta)

. A tDelta employs a

temporal XML expression (tXex)

which is an extension of ordinary XML elements by incorporation into them temporal as well as ordinary variables. In addition to valid time, TVDM introduces version time and associates it with each version. All versions including the base or the current one can be uniformly represented by tDeltas. A technique to implement TVDM is also presented.

As a scalable alternative to traditional server-based architecture, peer-to-peer (P2P) computing has become a popular distributed computing paradigm. However, efficient content search is absent, which hinders the wider deployment of the peer-to-peer systems. Earlier peer-to-peer systems such as Napster and Gnutella suffer from unscalability. Structured peer-to-peer networks achieve good scalability and high reliability, and yet they can only support a single-key based lookup instead of content search by means of DHTs (Distributed Hash Tables). In this paper, we propose a text-based peer-to-peer content search solution, which uses a hierarchical architecture. The heterogeneity of the popularity of terms both in documents and queries and the heterogeneity of hosts are both examined herein. There some techniques are employed to cope with the troubles arising from these heterogeneities. The experimental results show that the solution is feasible and efficient.

This paper presents a system for the segmentation of a printed document image into word images, which can be used effectively for document image retrieval based on keyword spotting. The system is composed of three image manipulation modules: skew correction, document layout analysis, and word segmentation. To enhance the practical applicability and flexibility of our research results, we test the system with 50 images of Korean papers and 50 images of English papers provided through full-text image retrieval services by the Korea Information Science Society and the Pattern Recognition Society, respectively. Currently, the accuracy of word extraction ranges from 90 to 95%, depending on the language of the document.

One of the key issues within the research of Media Assets is the description and intelligent recognition of content-based features. Upon the basis of massive fundamental research, we present in this paper a configurable content description and recognition model based on a Gaussian-Fuzzy membership system. Multiple and partial memberships and compatibility for various algorithms are supported in this architecture.

In this paper, we propose a novel blind MPEG video watermarking technology to hide copyright information by a slight modification of discrete cosine transformation domain in I-frames. It takes full advantage of both stillness and motion information of video content to guarantee the high perceptual invisibility. An effective characteristic extractor of block is delicately designed based on the properties of stillness and motion information, namely the texture, the luminance and the difference between adjacent frames. An eigenfunction is proposed to modify the extracted block characteristic. We illustrated the robustness of the scheme to several video distortions.

Traditional information retrieval (IR) systems are developed based on the “best match” principle which assumes that users can specify their needs in a query and documents retrieved are relevant to users. However, this objective measure of relevance is limited as it does not consider differences in experts’ and novices’ knowledge and context. This paper presents initial work towards addressing this limitation by investigating subjective relevance (that can include topical, pertinence, situational, and motivational relevance) features that can be incorporated into digital library interfaces to help experts and novices search and judge relevance more effectively. A pilot study was conducted to elicit initial subjective relevance features from experts and novices. The paper concludes with a discussion of elicited design features and their implications for user-centered digital libraries.

An effective and fast shape description and retrieval method is presented for huge image databases. As a shape representation for deformable objects, a multi-scale skeleton representation is proposed in order to preserve the consistency of the skeletons and to reduce the effect of the structural changes. Incorrect matches due to the boundary noise in a segmentation process are avoided by including multiple coarse skeletons of different scales. A fast computational method for the similarity of skeletons is also proposed by using the moment invariants. Experimental results on animal databases showed that the proposed method gives prominent accuracy in retrieval.

Form-based Web queries are commonly used by Web users for extracting information from the Web. However, some hidden information cannot be retrieved from the Web interface, which results in the limited query answerability through the Web. We present a system that can increase the amount of information to answer users’ queries. The design of our system includes (i) the introduction of a Web query rewriting model that handles any form-based query, and (ii) the evaluation of input queries using answers to previous, relevant queries. Experimental data shows that our query rewriting approach achieves more expected answers.

Existing library retrieval systems present users with massive results including irrelevant information. Thus, we propose SURM, a Retrieval Model using “Subject Classification Table” and “User Profile,” to provide more relevant results. SURM uses Document Filtering technique for the classified data and Document Ranking technique for the non-classified data in the results from keyword-based retrieval system. We have performed experiment on the performance of filtering technique, updating method of user profile, and document ranking technique with the retrieval results.

Question Answering has been the recent focus of information retrieval research; many systems just incorporate a search engine as a black box and most effort has been devoted to the question analyzer and the answer identifier. In the context of QA, however, passage provides an ideal medium between the document collection and an exact answer. And passage retrieval is a finer-grain approach than the traditional document retrieval both for the answer identifier and a human reader. In this paper, distinctions are first made between document retrieval and passage retrieval. And the Hot-Spot Passage Retrieval algorithm, which takes into account the measures of

blurred BM25, coverage

and

height

, is examined in detail. For evaluation, an isolated test is conducted and the algorithm gains 18.3% better answer redundancy and 4.8% better coverage rate than Okapi’s original passage retrieval algorithm.

Information overload has led to a situation where users are swamped with too much information, resulting in difficulty sifting through the material in search of relevant content. In this paper, we address this issue from the perspective of collaboration in query formulation. We describe a search assistant that helps users with query formulation by finding related previously submitted queries through mining query logs. The search assistant runs as a reusable software component and can be incorporated into various search engines. We report our approach to designing and implementing the software and evaluation results.

When confronted with a query, question answering systems endeavor to extract the most exact answers possible by determining the answer type that fits with the query and the key terms used in the query. However, the efficacy of such systems is limited by the fact that the terms used in a query may be in a syntactic form different to that of the same words in a document. In this paper, we present an efficient semantic query expansion methodology based on a question category concept list comprised of terms that are semantically close to terms used in a query. The semantically close terms of a term in a query may be hypernyms, synonyms, or terms in a different syntactic category. The proposed system first constructs a concept list for each question type and then builds the concept list for each question category using a learning algorithm. When a new query is given, the question is classified into the node in question category, and the query is expanded using the concept list of the classified category. In the question answering experiments on 42,654 Wall Street Journal documents of the TREC collection, the traditional system showed in 0.223 in MRR and the proposed system showed 0.50 superior to the traditional question answering system. The results of the present experiments suggest the promise of the proposed method.

As parts of global information infrastructure, digital libraries will likely be accessed by people all over world. Ontologies and their association can alleviate the heterogeneity and particularly the diversity of languages. This paper proposes a solution to cross-lingual information retrieval problem via ontology alignment. We elaborate two original techniques, i.e. primitives’ association based on CL-LSI and mapping configuration optimization, to augment existing ontology mapping technology. As a result, multilingual collections can be bridged by this mapping, and searching across them can be achieved by three tractable steps: querying against local ontology, routing to target ontologies, and harvesting contents there online.

This paper reports our research in the Web page filtering process in specialized search engine development. We propose a machine-learning-based approach that combines Web content analysis and Web structure analysis. Instead of a bag of words, each Web page is represented by a set of content-based and link-based features, which can be used as the input for various machine learning algorithms. The proposed approach was implemented using both a feedforward/backpropagation neural network and a support vector machine. An evaluation study was conducted and showed that the proposed approaches performed better than the benchmark approaches.

Topic-driven Web Crawler or focused crawler is the key tool of on-line web information library. It’s a challenging issue that how to achieve good performance efficiently with limited time and space resources. This paper proposes a focused web crawler wHunter that implements incremental and multi-strategy learning by taking the advantages of both SVM (support vector machines) and naïve Bayes. On the one hand, the initial performance is guaranteed via SVM classifier; on the other hand, when enough web pages are obtained, the classifier is switched to naïve Bayes so that on-line incremental learning is achieved. Experimental results show that our proposed algorithm is efficient and easy to implement.

The people, documents, and other entities from a domain persons know, or are in other ways associated with, influence their decision making and the types of recommendations that serve them best. For example, recommending persons to meet in a conference or a paper to read from a digital library collection does not only depend on the task, interests, and skills of a user, but also on the persons and works they are already familiar with. In order for personalization services to reflect this dependency, extended user models that consider users’ network of related domain entities in addition to other user characteristics, are required. Based on a unified context model, we present the Personal Web Context approach that models the typed relationships a user is involved in. Based on a Resource Network which can, for example, be built from the information collection and the associated meta data managed by a digital library, domain-specific rules are used to suggest valuable extensions of this ”neighborhood” of a user. Such work can form the basis for new types of digital library services.

Research in recommender systems focuses on applications such as in online shopping malls and simple information systems. These systems consider user profile and item information obtained from data explicitly entered by users. – where it is possible to classify items involved and to make recommendations based on a direct mapping from user or user group to item or item group. However, in complex, dynamic, and professional information systems, such as Digital Libraries, additional capabilities are needed for recommender systems to support their distinctive features: large numbers of digital objects, dynamic updates, sparse rating data, biased rating data on specific items, and challenges in getting explicit rating data from users. In this paper, we present an interest-based user grouping model for a collaborative recommender system for Digital Libraries. Also, we present several user interfaces that obtain implicit user rating data. Our model uses a high performance document clustering algorithm, LINGO, to extract document topics and user interests from documents users access in a Digital Library. This model is better suited to Digital Libraries than traditional recommender systems because it focuses more on users than items and because it utilizes implicit rating data.

In large academic libraries, the organization and management of diverse digital content for easy access has always posed a challenge. Users have varying information needs and are often confounded by having to access information residing on disparate systems via different interfaces and authentication credentials. To improve the user experience, the new Library Portal was implemented using enterprise portal technology to deliver personalization and single sign-on capabilities. It leverages the widely-adopted J2EE standard for application integration using an industry-leading application server and award-winning portal framework on low cost Linux platform. By utilizing standard-based technologies, the library is able to manage a scalable and highly-configurable portal which integrates with INNOPAC, the library‘s automated system, and the campus LDAP directory servers to provide convenient single sign-on to library services. Personalized access to INNOPAC, library contents and digital library resources is delivered through user profiling, group-based entitlements and a rule engine.

Digital libraries have been rather successful in supporting learning activities by providing learners with access to information and knowledge. However, this level of support is passive to learners and interactive and collaborative learning cannot be easily achieved. In this paper, we study how digital libraries could be extended to serve a more active role in collaborative learning activities. We focus on developing new services to support a common type of learning activity, field study, in a geospatial context. We propose the concept of personal project space that allows individuals to work in their personalized environment with a mix of private and public data and at the same time to share part of the data with team members. To support the portability of the resources in our digital library, the selected resources can be exported in an organized manner.

Recent years, electronic resources and electronic services provided by academic libraries in the Chinese mainland are becoming ever-increasingly diverse and multiplex. Libraries patrons strongly expect to obtaining one kind of more quicker, more easier, one-stopped, and personalized service to access libraries collections and use libraries services. Under the condition, Library of Shanghai Jiaotong University initiated the portal project. Shanghai Jiaotong University Library portal project defines its portal as a personalized, customized and integrated service of all kinds of resources and services provided by library. This article details the project background, general understanding of this portal service, functional requirements, design strategies, and some key and difficult problems in developing such a service in China.

Through a research and development of some key techniques and related technologies of digital library for ten more years, an important technical foundation for developing a practical digital library has been established. A dozen of digital libraries are summarized in three main types as the following: a DL of special collection; a service-orientated DL; a DL with commercial database. The architecture of a service-orientated DL has been developed in Shanghai Jiao Tong University (SJTU): It consists of various E-resources and an information service platform. Generally, the E-resources for this type of DL includes three parts: a) digital special collections of their own; b) commercial online e-publications or databases (including locally mirror-sited databases); and: c) some selected information resources on the Internet. The information service platform consists of an unique information searching platform and a VRS platform. E-resources are searched through a unique interface. Using the virtual reference system patrons is given advice and guidance concerning their problems in searching resources or other problems.

A good library can make significant contributions to library users. These users are likely to return to the library again and again where service is excellent and materials are accessible easily. The research purpose for this paper is to adopt the European Customer Satisfaction Index (ECSI) on a survey performed on library users to measure users’ satisfaction and loyalty factors. A questionnaire is designed to integrate measurement into the library’s management system for visual analysis of what users expect from the library. Results from the survey show that users expect more improvement in the “library environment” and “collections of printed publications” section. Conclusions from the study, we were able to make the following suggestions to satisfy users’ expectations: (1) increase seating areas for readers, (2) increase collections of printed publications and periodicals to meet readers’ demands, and (3) increase the number of duplications on popular literatures.

The advent of digitization has enabled individuals, institutions, and communities to create and disseminate digital representations of their cultural heritage in digital libraries (DL). These digital collections are increasingly broader in scope and reach, spanning geographical and cultural boundaries. They serve users from their originating culture and often also play the role of cultural ambassadors by serving users from other cultures. While cross-cultural use of DLs is present, little research is available on the influence of culture as it pertains to the design and use of DLs. In this paper, we define cross-cultural usability guidelines for DLs and apply these to the heuristic evaluation of six “national” DLs. Results of a usability evaluation study of five of the DLs are also reported. The national DLs were selected based on their mission, funding model, and originating organization and they all represent their respective cultures. Results of the evaluation show that current DL UIs and content do not serve international users very well. These results were confirmed in a user study in which users were asked to browse “national” digital libraries from countries other than their own. Based on our results, we propose guidelines for the designers of DLs to support their international users and truly serve as “cultural ambassadors” for their originating countries. While the results of this study were gained through the evaluation of DLs using the Latin alphabet, the general design guidelines are proposed for DLs of all cultures.

Using Nielsen’s Heuristic Evaluation, this paper reports a user study with six usability-trained subjects to evaluate PAPER’s access features in assisting users to retrieve information efficiently, part of an on-going design partnership with stakeholders and designers/developers. PAPER (Personalised Adaptive Pathways for Exam Resources) is an improved version evolving from an earlier implementation of GeogDL built upon G-Portal, a geospatial digital library infrastructure. After two initial evaluations with student and teacher design partners, PAPER has evolved into a system containing a new bundle of personalized, interactive services with four modules: mock exam; personal coach (practice and review); trend analysis and performance review. This paper highlights lessons learnt in the design of PAPER using Nielsen’s heuristics, and discusses implications for the design of access features in digital libraries in general.

Information and communication technologies (ICT) facilitate the process of identification, collection, storing, processing and disseminating of information. The library and information science professionals are utilizing ICT to keep pace with the problem of information explosion. The benefit of instant access to digital information is the most distinguishing attribute of the information age. In this paper, the authors tried to highlight the libraries/information centers in the 21

st

century, the different components of providing digital information services, and Information and Communication Technology’s role in modernizing libraries. In Bangladesh, the use of ICT is yet to take off. In absence of strong telecommunication the Internet facility is not spread out successfully. Financial constraints are the major hindrance for its growth at the national, local and organizational levels. As a result, the implementation of ICT facilities for libraries is not receiving adequate support from their parent body. Most of the libraries in Bangladesh do not have computer facilities. The print media is still a major source of information in libraries. However, the situation is changing, and the library professionals should be ready for everything to cope with the new ICT used in libraries/information centers. In this paper, the authors intended to emphasize the expected change in libraries/information centers in respect of user services and how to cope up with.

Electronic pathfinders/subject guides help library patrons learn about library resources and develop research strategies. By reviewing the existing literature and the current status of reference service in digital libraries in China, the authors suggest that integrating subject specific pathfinders in digital libraries would benefit both users and librarians. This paper also offers a format and a construction guide to aid librarians in preparing readable and usable electronic pathfinders. An appendix presents pathfinder format guidelines and a sample pathfinder.

In this paper, we discuss current issues on the copyright of digital libraries from a database designer perspective, and present a scheme for copyright protection of indexed digital contents stored in digital libraries that are associated with keyword-based retrieval operations.

With the rapid development of technology and the wide dissemination of information, many universities and institutions suffer from information overload and have to apply information management to deal with this information chaos in the digital world. During the 9th National 5-year Plan, CALIS started the project of Internet resource navigation database. In this paper, the goal, principle and contents for constructing an Internet resource navigation database on key disciplines are presented. The resource selection criteria and standards for web-based resource description are also described. The progress of the navigation database construction is introduced.

Information technology presents a challenge as well as an opportunity to reference service. This paper examines the reasons for the declining patronage of reference service, and outlines the benefits of offering live digital reference service. While covering the present status of this service, this paper also attempts to look ahead of its future development.

This paper proposes a structural matching algorithm for 3D retrieval. It decomposes the 3D object into meaningful patches, and performs similarity calculation by bipartite matching. The retrieving performance of the proposed algorithm will remain almost no any decreasing for connected objects. While the performance will be a great improvement if the database contains disconnected objects.

Web page classification is one of the essential techniques for Web mining. The approach proposes a framework for Web page classification, that is a hybrid architecture using the PCA features selection approach and the SOFM with a combination of some conventional statistical methods. The proposed hybrid architecture consists of four modules as following:

The page-page-preprocessing module is used to extract textual features of a document, what is divided into

stopping

and

stemming

. The

stemming

is a process of extracting each word from a document by reducing it to a possible root word. The

stopping

is a process of deleting the high frequent words with low content discriminating power in a document, such as ‘to’, ‘a’, ‘and’, ‘it’, etc.

Metasearch has been proved to be a quite effective method to integrate information spaces provided by underlying search engines. But how to admit new resources dynamically was almost neglected. Current solutions include software recoding or resource rebinding by metasearch engine developers. Besides extra cost in time and money, both tactics might be futile for resources in a controlled network (e.g. a campus network using virtual IP address) or fail to keep pace with the emerging of new resources.

In this paper, we design and implement a certificate-based authentication and authorization architecture in digital library. Our certificate-based authentication architecture consists of resource server, directory server, policy engine and log server. The heart of this system is the policy engine, which gathers and verifies certificates and then evaluates the user’s right to access to the requested resource based on these certificates. Our system uses two types of persistent certificates: X.509 user identity certificates and attribute certificates. The identity certificates are generated and managed by certificate authorities, such as the Netscape CA server. These certificate authorities provide a Web interface that allows the creation or revocation of certificates. A directory server can be used to provide the certificates for use by applications and Web browser to manage the certificates for the user. The resulting certificates can be stored in directories chosen by the user that are accessible via a Web server, a directory server, or an MSQL database. The client first has to pass the Kerberos authentication. He has to provide the identity and the password. If he has passed the authentication, then the certificate authority issues the certificate for him. We implement the Kerberos algorithm and DES algorithm in the architecture. When we have got certificates and want to get services from service provider, we have to get authorization. In our model, authentication and authorization is separated. The authorization architecture consists of browser, service provider and authorization directory server. When an individual within the consumer community requests information from a remote service provider, the browser sends the individual’s digital certificate. The service provider validates the individual’s certificate and uses it to locate the institution’s authorization server. The authorization server checks the validity of the service provider’s certificate. If valid, it returns attributes concerning the individual’s status to the provider in the form of a list of attribute names and values. From the authorization server’s response, the service provider decides to whether deliver the service to the individual. Finally we compare the difference of certificate-based architecture and the user identity-password architecture. The excellence of the certificate-based architecture and the shortcomings of IP address, identity-password architecture are analyzed

The China Culture Grid is proposed to effectively organize, classify, and integrate the heterogeneous resources for the purpose of providing pervasive, proactive, on-demanded, and personalized services for various people with different backgrounds, capabilities and expectations, under different time and venue. This paper describes the architecture and key technologies of the China Culture Grid, including Uniform Content Locator, Ontology and Personal Content Locator. An experiment study and a prototype system are also presented.

Face detection algorithms have primary factors that decrease a detection ratio: variations in lighting effect, location and rotation, distance between objects, and complex background. Variations in illumination, background, visual angle and facial expressions make the face detection difficult [1], [2]. We propose a face detection algorithm for color images in the presence of varying lighting conditions as well as complex background. Our method detects skin regions over the entire image, and then generates face candidate based on the spatial arrangement of these skin patches. The algorithm constructs eyes, mouth, nose, and boundary maps for verifying each face candidate.

Three-dimensional volume data as a new type of resources for digital library poses a challenge for user to visualize the data remotely with ordinary network and user terminal. This paper presents a decoupling architecture with a front-end viewer and a back-end server. Multiple reference images are produced by the server and warped and composed by the front-end. This approach enables remote visualization at interactive speed and also protects original data.

Nowadays an educational digital library usually contains instructional materials such as class lectures, seminar presentations, and various training materials. These materials accompany video and audio media. In this case, multimedia streaming and delivery techniques are concerned in the construction of a novel multimedia digital library. In this paper, we focus on multimedia streaming techniques in digital library applications. To achieve scalability and deliver high quality streams, we present the multimedia proxy caching scheme between the server and client path, and a prototype design and implementation of a proxy caching system is proposed. By validate our implementation, the experimental results show that our proxy system can contribute to improve streaming media quality, reduce network transferring latency and decrease streaming service response time.

This paper describes a multi-document summarizer in Chinese, ACRUX, which contains three new techniques: a fuzzy classification method based on KNN (FAMKNN), Subject-Oriented Multi-document Summarization (SOMS), and Multi-document Summarization with Link Analysis.

This paper proposes a simple method of querying between heterogeneous Ontology-based information sources using association matrix. It introduces the definition and calculation of association matrix and presents a method using concept vector and association matrix to rewrite query.

The project

Digital Library Architecture and Application Platform

(DLAAP)is a state hi-tech project approved by the State Planning Commission in 2001 and undertaken by the Project Team of Shenzhen Library. The General Plan and Technical Implementation Plan of the project had gained the experts’ recognition in 2002 and the research of the architecture of the system and application platform development had made some achievements. At present a distributed digital library application platform had been developed, which includes digital resources process and inquiry system, library business processing system, reader service and management system, which will be fully put into use at the opening of the new Shenzhen Library.

In resent years, several digital libraries have established large repositories of scientific literature, such as ISI SCI®, CiteSeer.IST [2] and CORA[1] etc. These projects pay much attention on how to interlink scientific articles via reference indexes. However, few of them richly utilize the semantics behind reference links and people get limited supports in searching, reviewing and analyzing scientific literature from scholarly perspectives.

In digital library, the provided personalized service responds to the user’s interests. Therefore, it is important to trace and understand the changes of the user’s interests, so as to supply pertinent service. This paper depicts a hybrid model, which is based on concepts correlations, to deal with users’ interest change. Using this model, the change of the user’s interest can be quickly found and combined with previous user’s interests. So, a better personalized service can be provided.

This paper introduced a system-WEBDL (Web data Digital Library) which addressed the problems of effectiveness and efficiency of searching from the Web with the hope of making full use of available information on the Web The user could search the web information just like query in a local traditional Digital Library.

The SJTU library has subscribed 225 databases, including some 400,000 e-books, 1,7000 full-text e-journals, proceedings, thesis and other citation databases, etc. These e-resources are allocated in different databases, covering a variety of disciplines and overlap each other in contents in certain disciplines. It is necessary to create a unique gateway for these resources.

With Comparing metadata-based method, semantic web-based method and ontology-based method for integrated management of digital resources, the authors think that ontology-based method will be the developing direction for integrated management of digital resources.

For the description and management scientifically and in order and for the establishment of digital library information service system with our own features conform to international standards, we design and develop a set of system platform on the basis of multimedia resources. The multimedia resources metadata managing system supports the function of digital collection, processing, bibliography description, storage, organization and publishing of text, pictures or images, audio and video frequency and other multi-media information. The system must first define a set of metadata standard system according to the features of different resources and design metadata model describing these them, order metadata system templates and establish metadata application model through its models. The function of the system templates is to process, describe information resources when bibliographers establish some new multimedia resources on the basis of metadata system defined in advance by the system; secondly, the system describe the metadata of digital resources through metadata description template, identifying subject, classification and related data; thirdly, the system check, and publish data information through data checking and publishing model. During the period of the research of the multimedia metadata model and management system module, we use the domestic and overseas experience for reference. The choice of metadata, description of attributes, validity inspection of DTD and Schema, the standard of description classification, quotation of authority, XML and XSL data format and the upload data through the FTP all conform to the domestic and international standards. The system sets up multimedia resources element attribute descriptive frame on the base of the DC and also provides standard Chinese Library Classification system and authority name control. The bibliography description process can be easily realized with the use of visual operational interface. The pull function should be convenient for bibliographers. The system is highly integrated with other digital library system such as virtual reference service system and MELINETS (Modern Electronic Library Information Net System) of BUPT (Beijing University of Posts and Telecommunications). Mature techniques have been employed during the developing period. Open and standard interface are provided and various data exchanged are supported. The metadata database and object database can be effectively managed with the system management module. The realization of the design and techniques, the standardization and automation of multimedia resources in bibliography description are strengthened and the working efficiency improved.

The continuing evolution of the Australian Digital Theses (ADT) Program, with the aid of an additional grant from the Australian Federal Government, sees it redeveloping the existing central metadata repository to increase its coverage and utility to the national and international research community. The repository’s content will expand to include metadata about all Australian higher degree theses, whether in digital form or not. This goal responds to a clearly stated desire of the Australian research community, and brings together the university, government and corporate sectors, in building an open access service which exposes Australian research.

With its diverse cultural and linguistic heritage, India today produces significant volume of digital material in Indian languages. This has been facilitated by increasing availability of word processing systems supporting Indian languages and their use in various areas including e-governance; education and research; and mass media. There is need for digital library software for organizing and provision of access to this material. Such software has to meet two prime requirements: Indexing and searching of documents in Indian languages (full text and metadata), and customizing the collection user interface in Indian languages. Further the software should be able to handle Indian language material in different encoding formats and fonts. Majority of Indian language material available online today seem to follow one of the three encoding strategies: ISO 8859-1 and Windows 1252 series character sets, with custom fonts; ad-hoc (font-specific or user defined) encoding schemes; and Unicode character set. Search and retrieval requirements would include features such as word truncation and alphabetical sorting. Cross-language material searching is an advanced search requirement. Greenstone is a very popular open source software used today for creating digital library collections. Main objective of this study was to assess capabilities of this software in creating Indian language digital library collections with above mentioned requirements for indexing, searching and display. We gathered five sample collections in two Indian languages Hindi and Kannada, in different encoding formats, for this study. For each of these collections, we assessed the multilingual support of Greenstone with respect to collection building; search and retrieval; and interface design. We used the ‘Collector’ approach of Greenstone to build the five collections. We could successfully build the collections. Limitations were found in handling metadata in Indian languages using the ‘GLI’ approach. We present details of internal mapping of character sets carried out by Greenstone during collection building process. We could successfully carry out simple keyword and Boolean searches on these collections. We discuss details of search features. Viewing results requires installation of suitable fonts at the operating system level and configuration of the browser. We found limitations in sorting. Greenstone does not support crosslanguage searching. In terms of users interface, Greenstone has in-built support for customizing the interface for well known languages. It also supports designing customized interface for other languages. We could successfully design desired user interface for the test collections in Hindi and Kannada. Overall, Greenstone appears to be a versatile software for building Indian language digital library collections, with some limitations.

This paper is the summing-up of the study on descriptive metadata for variant resource objects in a national project

Chinese Digital Library Standards (CDLS)

. In this study, the two key issues, metadata structure and extended rules, are focused on while designing the metadata standards aiming at the different objects and their characteristics. They will enhance the interoperability among the metadata standards. This paper takes the metadata standard of ancient atlases as an example to illustrates its structure, elements and extended rules.

In more than 90 years history, plenty of featured scientific literature has been accumulated at Tsinghua Univ. Library. From 2001, we participated in the international cooperative initiative EMANI (Electronic Mathematics Archiving Network Initiative) together with SUB Goettingen (Germany), Cornell University Library (USA), MathDoc (France) and the Springer Press. The major idea of this project is the cooperation can promote the digitization of ancient as well as current publications of Mathematics all over the world, to care about the long term preservation of the content in readable form and provide convenient access to distributed digitized materials for users all over the world (See http://www. emani.org).

The paper firstly explains the definition of Online ILI. Then taken Webtraining of PKU library as an example, authors try to put forwards an operable framework and methods of Online ILI considering the conditions in academic libraries in China.

The Tun Hussein Onn Library (thereafter known as THOL), collaborates with the Monash University Library Australia and other education partners of Sunway College in their digital library initiative. THOL, as a collaborated digital library, has several unique characteristics:

Extensive amount of heterogeneous resources, with differentiated access to different users.

Ongoing digitization effort for resources from different libraries

Multifaceted access restrictions

Heavy reliance on online library services and education by remote users

The famous Indian librarian, Dr. S. R. Ranganathan, proposed a five laws of library during 1931[1], which was taken as classical theory of library ever since. The laws point out the basic principles of library, clarify the object for effort, and obtain widely agreements during the several decades. Based on these laws, American librarian Michael Goman proposed a new five laws of library. Then in 1975 IFLA conference reached same comprehension: The library possesses four social functions: 1) Preserve human cultural heritage; 2) Open social education; 3) Deliver scientific information; and 4) Develop intelligence resources. All this has made huge contributions in promoting the library activities and still serve as basic principals for that. However, in the new situation of knowledge-based economy, the conceptions are subject to further deepening and developing. The computer technology offers good chances and opportunities for libraries to do it while the network technology makes it possible for libraries to be in the front of the information technology. Electronic collections and their searching abilities are good examples. China’s policy to develop university and colleague’s education also helps the changes. Some new ideas and practices about the library’s functions have been formed from the classic and practice laws. They are “

service- oriented information provider

” comparing the “

custodian of books

”; “

multiple media

” comparing the “

one medium

”; “

library without walls

” comparing the “

own collection

”; “

the library comes to us

” comparing the “

we go to the library

”; service “

just in time

” comparing the service “

in good time

”; “

outsourcing

” comparing the “

in sourcing

” and “

global reach

” comparing the “

local reach

”,etc. Library is transferring from local library to a global library and must immerge into the global information market. As is put in the UNESCO document No. 105, library is a area should pay special attention. In modern academic organization, the word “Library” is given a new meaning. It will not be a place to collect,... The library will become a dialogue center necessary in modern study, teaching and scientific research between information provider and reader. The library,... also provide intelligent environment. This has posed challenges for librarians who should be more than traditional ones. Their responsibilities include the traditional techniques of classification, cataloguing, acquisition but also the new ones to train patrons in searching databases and to use the new technology in providing efficient service to patrons, etc. On the other hand the libraries, especially the academic libraries should become center of learning with the assistance by librarians who should therefore involve research work. These new functions and services could be available all around the campuses and beyond with the facilities of network.

An investigation on digital reference service (DRS) has been carried out amongst 30 world top university libraries. This paper summarizes the recommendable elements of DRS in these libraries. Then the basic characteristics are generalized and an ideal DRS model including eight modules for academic libraries will be introduced.

“service innovation” is the theme that library should study. The library needs kinds of innovation such as mechanism, librarian’s quality, reading circumstance and service function innovation.

In the past 10 years, academic libraries have combined various modern digital technologies with traditional library services to improve their services.

Explicit knowledge is coded knowledge that is saved in various formats, such as books, journals and information in online databases. To incorporate the concept of knowledge management into the management of libraries will change services and management in libraries. In terms of processing, it will increase the division of explicit knowledge and break packets of explicit knowledge into smaller pieces. This will be very beneficial for user retrieval of information as knowledge becomes more complex within branching sub-disciplines, and it will also enhance the effect of the information services in library.

Academic digital library should consider applying modern marketing theory as the significant change in resource form and service means in recent years. Marketing in academic digital library can be regarded as a process of understanding, stimulating and meeting the needs of its users. It must satisfy the selected user market through establishing specific academic resources and services of digital library. The four main principles about assessment of user request, establishing target market, using whole market strategy and emphasizing long-term benefit are very important in marketing academic digital library.

Digital libraries, institutional repositories, electronic journals, aggregator services, and open access publishing are just some of the new digital resources and services that are impacting libraries and information centers around the world. Marketing these new resources and collections is even more important in this new digital age, especially when some are challenging the value and worth of academic libraries. In the past, different audiences and patron groups with different expectations generally required different marketing strategies. But is this true in the digital age? This poster will compare and contrast the marketing strategies of academic libraries in North America and the Asia-Pacific Rim, focusing on shared strategies and new methods of reaching and convincing audiences of the value of academic libraries, which are key players in the new digital age, linking users with an ever more complex array of digital resources and services. The presenters will cover such topics as branding digital resources; imbedding marketing into training and instruction; forms of electronic marketing including newsletters and web logs (blogs); and integrating marketing into user needs surveying.

Digital library development demonstrates its potential in content creation and collaborative partnership. However, it will take time for digital libraries to achieve critical mass. Therefore sustainability is a crucial element in fulfilling their potential.

The requirements of intensive information urge the emergence of portal. Based on the existing research, portal can be described as an integrated system of information, services and applications with the capabilities of single sign-on and seamless link as well as the functions of personalization and customization. Generally speaking, there’re three portal modes : (1)Vertical portal, which provides access to a variety of information and services about a particular area of interest; (2)Horizontal portal, often referred to as “mega-portals”, which targets the entire Internet community. (3)University or enterprise portals.

The Hong Kong Baptist University (HKBU) and Hewlett-Packard HK SAR Ltd (HP) collaborated a pilot project named “Surfing the HKBU e-Campus in 80 Days” from 17 March to 4 June 2004. It is designed to promote e-culture and to enhance students’ e-learning experience through the use of pocket PCs and the University’s wireless network technology. A committee was formed with senior members of the HKBU Information Technology Services Center, the Library, and six faculty members.

The Internet offers unprecedented opportunities for people to communicate and access to information. Libraries are embracing the new digital medium as it allows massive data storage, faster and more effective data retrieval, and easier content navigation and transfer. This communication revolution helps libraries fulfill their important mission in the society as the information repository and disseminator. The legal landscape of the digital environment, however, is very different from that of the traditional print environment due to the nature of the digital technology. This research explores the shift of legal paradigm and its implications upon the digital library development, and also suggests viable solutions to cope with the trend.

We propose an online learning algorithm for digital library. It learns from a data stream and overcomes the inherent problem of other incremental operations. Experiments on RCV1 show the superior performance of it.

The design and construction of a personalized portal based on Library patrons’ requirements and functions of academic libraries with such characters as modularization, personalization and intelligent online help was sketched out in this paper. The framework involves the integration of various information resources and services according to the attributes of subjects. Also, the integration of services and resources based on the way of serving was contained in the system. In addition, this system consists of a trial module that tracks users’ behavior and therefore recommends digital collections. The system which provides information seeking, filtering, organizing and delivering is part of a networked digital library project whose principle goal is to present users a convenient way to select, structure and retrieval digital resources from different repositories.

This paper analyses the characteristics of military equipment information resources and the special demand of military equipment digital library. Based on the general digital library architecture combined the characteristics of military equipment information resource, it builds the architecture of military equipment digital library and the operation mechanism for military equipment digital library is discussed.