CHAPTER 5: A PROTOTYPE HYBRID KNOWLEDGE ORGANISATION SYSTEM MODEL FOR VIDYANIDHI DIGITAL LIBRARY This chapter renders the output of research study and presents an overview of VDL, and its current features, State-of-the-Art technology. It also provides details of evaluation of KOS tools, methodology adopted to develop the prototype hybrid KOS model for VDL. The Proof of Concept demonstrates the outcome of the integration of KOS features in to the VDL along with the screenshots and prototype demo video are furnished at the end.
5.1 INTRODUCTION The growth of digital information over the Web including DLs and ETDs is tremendous in the last decade and leading to more complexity in organization of information. Due to this complexity, information organization, its retrieval and the interfaces for showing search results have posed challenging environment to the information scientists in these research areas (Gunjal & Urs, 2010a). A review of the literature reveals that KOS deployed in DL and ETD vary and differs in their model adopted and its services. There are different levels of KOS which they come under by drawing the line by adopting different KOS aspects. In Vidyanidhi Digital Library (VDL http://www.vidyanidhi.org.in/) also the level of KOS shows there is scope for improvement to move to the next level of KOS. Hence, in this research, the study is focused on KOS aspects and to map them for an ETD case i.e., VDL. The goal of this research is to design and develop a prototype hybrid KOS model for the future development of VDL at the University of Mysore, India.
To achieve this goal, a gap analysis study was carried out to the existing VDL eTheses Project to understand the scope for improvement that can enhance the information retrieval (IR) process. Hence, various Open Source and Commercial tools are evaluated. Selection criteria of tools were based on the features such as – tool should be Java based, can enable Information visualization and Concept maps aspects of KOS. Based on the selection criteria four potential tools are identified and evaluated such as IHMC CMapTools (CMap), JavaScript InfoVis Toolkit (JIT), SpaceTree and Visual Understanding Environment (VUE). After thorough evaluation and review of tools selected for prototype, the researcher selected JavaScript InfoVis Toolkit (JIT) tool to integrate with the VDL tool i.e. DSpace based on the existing features. This tool enable users to create interactive data visualizations and to Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
180
Chapter 5: A Prototype Hybrid KOS Model for VDL
visualize the existing VDL communities like a Concept map structure showing relationship among communities without navigating to the new page. To achieve this prototype, a java based wrapper application is built in between DSpace and JIT. JIT tool will call the servlets which are residing in the application server i.e. Tomcat of DSpace. As a response, XML format will be send to the JIT tool. This XML format will be recognized by JIT tool which is internally using SpaceTree. With this XML format, JIT tool will display the content in a graphical format to visualize the Concept map.
Finally, this chapter concludes with the Proof of Concept (PoC) that demonstrates the outcome of the integration of KOS features in to the VDL along with the demo and sample screens of the prototype hybrid KOS model that makes use of Concept maps which can help users to visualize the communities and its relationships. It represents a new form of information visualization by using appropriate tool along with DSpace and thus to a new approach of the application of human-computer interaction (HCI) research in web-based information retrieval.
5.2 VIDYANIDHI DIGITAL LIBRARY (VDL): AN OVERVIEW This section provides a brief overview of VDL and its current features with state-ofthe-art technology.
Vidyanidhi Digital Library: A Profile Vidyanidhi (Meaning 'Treasure of Knowledge' in Sanskrit) is India's premier Digital library initiative to facilitate the creation, archiving and accessing of doctoral theses. Vidyanidhi is an information infrastructure, a digital library, a portal of resources, tools and facilities for doctoral research in India. Vidyanidhi is envisioned to evolve as a national repository and a consortium for e-theses through participation and partnership with universities, academic institutions and other stake holders. Vidyanidhi enhances access to Indian theses and enlarges the reach and audience for Indian doctoral research works.
VDL: State-of-the-art technology adopted Vidyanidhi is a collaborative, distributed architecture, participatory Digital Library (like ADT model - Australasian Digital Theses Program http://adt.caul.edu.au/) that helps to establish a distributed database of digital versions of theses produced by the research students. It seeks the partnership of universities and other institutions and sign MOUs to Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
181
Chapter 5: A Prototype Hybrid KOS Model for VDL
formalise the partnership. Further Vidyanidhi’s technical details comprising digital library of e-Theses i.e. developed and hosted on the Microsoft Small Business Server 2000. The Web server was deployed using Windows Server 2000, Internet Information Server (IIS) 5.0 using Microsoft SQL Server 2000 and ASP (Microsoft case study, 2004).
Vidyanidhi: Content and Collection Currently VDL provides access to (a) Indian Theses Database - provides search interface to collections* such as (a) Indian Theses Database (Metadata database) and (b) Indian ETD collection - on DSpace (Full text of theses). * Statistics provided here are as per the available current statistics.
a. Indian Theses Database (Metadata http://www.vidyanidhi.org.in/home/firstpage.htm)
provides search interface to Indian Theses Database
is growing and continuously updated
access to more than 1,50,000 Indian theses metadata
contains records of bibliographic descriptions of theses
has implemented the Unicode standard for Indian Languages and scripts.
multilingual database with records in English and Indian Languages (Kannada, Telugu, Urdu and Hindi).
efforts are in place to include full-text theses of all Indian languages
being Unicode compliant, it includes 1500 records in Kannada (for Kannada Theses) and 4000 records in Hindi Language (for theses in Hindi).
also provide University database listing 281 universities and Ph.D rules and regulations of 74 Universities.
in the future, the organization plans to implement support for Tamil, Telugu and Marathi, which is currently under development.
provides various search options such as - Simple Search, Advanced Search, Field Specific Search, Author Search, Supervisor Search, Subject Search and University Search
offers Search in Indian Languages : Kannada and Hindi
Simple Search provides o Multilingual Search options – English, Hindi, Kannada o
It gives option for – Also Search link
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
182
Chapter 5: A Prototype Hybrid KOS Model for VDL
b. Indian doctoral theses (Full text http://dspace.vidyanidhi.org.in:8080/dspace)
Indian ETD collection uses DSpace
access to 12000 full-text doctoral theses – to view, access and download (as per restrictions by the authors)
available in full text searchable PDF format
viewable only and digitally signed
full-text theses are available pertaining to the Indian languages such as English, Kannada, Hindi, Telugu and Urdu
users can Browse by - Communities & Collections, Titles, Authors and By Date
Simple and Advanced Search
5.3 APPROACHES ADOPTED FOR THE DEVELOPMENT OF PROTOTYPE MODEL Different kinds of approaches, phases and the approach adopted for development of Prototype Model are as follows:
Kinds of approach Different kinds of approaches to KOS aspects such as Classification, Categorisation, Concept map, Search strategy, Information Visualisation, Thesaurus, Taxonomy, Ontology, Semantic Web, Artificial Intelligence and other aspects have shown their evolution, trend and development along with implications on DLs and ETDs in order to get fruitful results. Hence, in this research, the study is focused on KOS aspects such as Concept map and Information Visualisation to map them for an ETD case and to develop a prototype hybrid KOS model for the future development of VDL at the University of Mysore, India.
Phases of Prototype development i.
Customisation of Commercial tools
ii.
Integration with Open Source tools through coding
iii.
Selection of KOS aspects for prototype
iv.
Practical and Theoretical aspects of prototype
v.
Design and development aspects of prototype
vi.
Solutions obtained for the challenges faced
vii.
Implementation of the developed prototype
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
183
Chapter 5: A Prototype Hybrid KOS Model for VDL
Approach adopted for Prototype Besides the aforesaid approaches, the following approach and methodology framework adopted to develop and implement the prototype. i.
The researcher adopted Integration with Open Source tools through coding approach to develop an interactive prototype model.
ii.
Evaluation and selection of tool to integrate with the existing VDL tool i.e. DSpace
iii.
Design and development of prototype hybrid KOS model to VDL with the help of integration of evaluated tools by coding.
iv.
A java based wrapper application is built in between selected tool for integration and VDL tool i.e. DSpace.
v.
Implementation of the developed prototype
vi.
Output is expected to visualize the communities and its relative entities in a single window
vii.
Result should display the content in a graphical format to visualize the communities in a Concept map form.
Development of a prototype of hybrid KOS model for VDL i.
A gap analysis study was made in the existing VDL ETD Project to understand the scope for improvement that can enhance the information retrieval (IR) process.
ii.
In this context various Open Source and Commercial tools were evaluated.
iii.
The selected tool was integrated with the existing VDL tool i.e. DSpace.
iv.
The Proof of Concept (PoC) demonstrates the outcome of the integration of KOS features in to the VDL
v.
Demo and sample screens of the prototype study of hybrid KOS model for VDL are added in Appendices.
5.4 KOS STUDY IN VDL The VDL has adopted many tools and standards, techniques to expend the search interface. The following KOS features in VDL are identified in Table 5.1. .
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
184
Chapter 5: A Prototype Hybrid KOS Model for VDL
Table 5.1 Study of KOS features in VDL Classificatio n
N/A
Visualization/ Search Interface / Conceptualizati Auto-Categorization on N/A VDL provides multilingual search capabilities in two Indian Languages i.e. Kannada, Hindi and English. VDL provide different type of search features such as Simple Search, Advanced Search, Field Specific Search, Author Search, Supervisor Search, Subject Search, University Search and Directory search Simple Search provides (Figure 5.1) Multilingual Search options – English, Hindi, Kannada It gives option for – Also Search link Advanced Search provides (Figure 5.2) with all of the words with exact phrase with atleast one word with out the words
Miscellaneous Unicode implementation for Indic Languages and scripts First ever multilingual portal for Indian theses providing access in Hindi and Kannada Facilitates complex multilingual searches Enable remotely upload theses University partners get value added services
Tools and Standards used in VDL Various tools and standards have been used in VDL to provide content services in both database and eTheses projects. The present research study focuses on VDL eTheses. For the VDL eTheses Project, following tools and standards have been used. Vidyanidhi is hosted on DSpace software using Dublin core metadata element set for metadata. It follows Metadata Standards by ETD-MS: an Interoperability Metadata Standard for Electronic Theses and Dissertations – version 1.00 (http://www.ndltd.org/standards/metadata/etd-ms-v1.00rev2.html) as shown in Figure 5.3.
Search Interfaces in VDL
Figure 5.1 Simple Search in VDL
Figure 5.2 Advanced Search in VDL
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
185
Chapter 5: A Prototype Hybrid KOS Model for VDL
Metadata Elements used in VDL The metadata elements followed and managed by Vidyanidhi are: i. dc.title ii. dc.creator iii. dc.subject iv. dc.description v. dc.publisher vi. dc.contributor vii. dc.date viii. dc.type ix. dc.identifier x. dc.language xi. dc.coverage xii. dc.rights xiii. thesis.degree
Figure 5.3 Metadata Submission form of VDL
For Indian languages, VDL provide facility data input and automatic transliteration approach with a small plug in software has been developed based on ITrans Standard for transliteration that will automatically convert the data input in Indian Languages into Transliterated Roman Script Records (Urs, 2004). 5.4.1 Limitations of KOS in VDL Based on the analysis of comparative study of DLs and ETDs, one can assume the current level of KOS of a particular DL or ETD helps to aim for the next level of KOS for its subsequent development of the next stage. That is to say, when the books were simple, the enumerative scheme of classification was adequate to classify most of the books, but when the subjects turned complex, the faceted classification systems came into use with more specificity to express each facet of the subject. Similarly it applies to the cataloguing and indexing practices. A study of a live ETD site is also made to examine the KOS in the DL.
In this context as already mentioned the VDL is used as a case here. Hence, on the basis of KOS study conducted on VDL, it is observed that there is scope to move VDL to the next level of KOS. This led the researcher to propose a hybrid KOS model suitable for DL architecture which can give the precise results. Similarly, one of the limitations of VDL is Metadata harvesting where other Dublin Core elements from associated universities cannot be supported.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
186
Chapter 5: A Prototype Hybrid KOS Model for VDL
5.4.2 Research on VDL and Gap Analysis Various researches have undertaken in VDL w.r.t different aspects of KOS. Few of them are explained here.
Urs and Raghavan (2001) explicate the building blocks of the Vidyanidhi are two fullcontent databases and the metadata database; the search engine; the user interface with multilingual interfaces layered and embedded within. Similarly Urs (2003, 2004) provide an overview of VDL, its architecture; and its database along with its content and features.
Further Angrosh and Urs (2007) elucidate Ontology-based information retrieval systems developed using the case of Agricultural ETDs present in VDL to enhance the search experience (Urs & Angrosh, 2006).
So for, the studies reveal that KOS deployed in DL/ETD vary and differs in their model adopted and its services. There are different levels of KOS which they come under by drawing the line by adopting different KOS aspects. In VDL also the level of KOS shows there is scope for improvement to move to the next level of KOS. Hence, in this research, the study is focused on KOS aspects and to map them for an ETD case and to propose a design of hybrid KOS model for the future development of VDL at the University of Mysore, India. An effort has been made to integrate the existing ETD collection in DSpace with other Open Source or Commercial tools to deploy KOS features such as Concept Maps or Information Visualization. The outcome of this prototype study has showed it helps to visualize the VDL communities through which it enhances the user experience and provide better results in more user-friendly, interactive and intuitive ways.
Gap and scope for improvement in VDL In VDL eTheses Project, the current observation shows the following issues and hence, it indicates that there is scope for improvement in these areas to provide enhanced services to its clientele.
currently users have to navigate through various pages to reach for the required document.
this process involves 7-8 steps consecutively moving to individual web pages each time.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
187
Chapter 5: A Prototype Hybrid KOS Model for VDL
users need to spend substantial amount of time
there is a need of single user interface (UI) showing various relationship in a Concept map or Information visualization.
there are ample opportunities for enhancing user experience.
5.5 STUDY, EVALUATION AND SELECTION OF KOS TOOLS Based on the KOS study conducted in DLs and ETDs, the outcome of the study was helped to identify the gap and scope for improvement in VDL. This led to conduct further research to develop and implement a prototype hybrid KOS model for VDL. As explained above in section 5.2, the current status of KOS aspects in VDL helped to understand the gap and to decide on the attributes for the prototype.
Related Works Over a period of time, many research efforts were noticed and recorded under the literature review (Chapter 2) and research trends in KOS (Chapter 3). It is evident that, many researchers adopted various KOS aspects in DLs to enhance its services to their clientele. This research study caters to user interface (UI) part of DL dealing with providing easy browsing mechanism for users of VDL.
Marchionini (1999) believes that most of the digital library research and development to date has centered on issues related to technology and content of digital libraries. However Geisler and others (2002) states basic finding aids such as search and browse are common in most of today’s digital libraries. On the other hand, Furnas and Rauch (1998) found that in searching for information a “one-shot query” is very rare. Jones and others (1999) state people do not just search for items in digital libraries, but also browse for them. In this regard, Gutwin and others (1999) discuss the browsing in digital libraries, but tend to focus on how user interfaces they develop can support browsing, rather than considering what browsing is. Likewise Zhang and others (2009) proposed various types of principles for design of friendly user interfaces and offer detailed and specific suggestions that are of practical importance.
Urs and Angrosh (2006) analysed the past research efforts in ontology-based KOS in DL. Ontologies, specifically when based on knowledge representation formalisms do add
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
188
Chapter 5: A Prototype Hybrid KOS Model for VDL
value and enhance the search experience. On the other hand Zhou and others (2007) build a prototype system to fully explore technologies that can support collaborative concept mapping effectively and provide the following features: visualization of concept map, flexible collaboration style, supporting natural interaction, knowledge management and history management.
Bergstrom and Atkinson, (2009) build a prototype to develop and test the effectiveness of a web-based application, PaperCube (http://papercube.peterbergstrom.com/), which was designed to augment a scholar's interaction with a digital library and explore bibliographic metadata using a defined set of visualizations. Similarly Wiederkehr (2009) explain PaperCube research that helps user to browse the data of CiteSeerX (Scientific Literature Digital Library and Search Engine) using different views i.e. Treemap-type visualization shows citation or reference hierarchy. The application is built as a RIA using HTML, CSS and the Sproutcore JS framework and runs in all the modern web browsers.
Kohlhammer (2010) explains the common approach to support user navigation and retrieval in digital libraries and data centers, operates on meta-information, that is manually appended to the datasets which has limitation due to scalability. To overcome this contentbased search, i.e. search on the data itself is addressed by a cooperative research project of the German National Library of Science and Technology in Hannover (TIB) together with the Technical University of Darmstadt and the Fraunhofer Institute for Computer Graphics Research (IGD).
Search V/s Browse Many researchers worked on the benefits of browse or navigation compared to search. Users opt for browse rather than search as it depends on the accuracy of the search results that meets users’ expectations in IR process. Some examples to quote are:
McKay and Cunningham (2003) convey that Browsing and searching are often interleaved during information seeking and DLs often support full-text search, but are not so helpful in supporting browsing. They describe here a novel browsing system created for the Greenstone software used by the New Zealand Digital Library that supports users in a more natural approach to the information seeking process. Likewise, Hagedorn, Chapman and Newman (2007) developed a prototype that would showcase methods for search and retrieval Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
189
Chapter 5: A Prototype Hybrid KOS Model for VDL
using automated clustering of subject metadata. Results indicated precise clustering of records was achieved and the prototype interface proved that faceted classification could be powerful in helping end-users to find resources.
Dalmau and Others (2005) confirms that structured forms of browse and search can be successfully integrated into digital collections to significantly improve the user’s discovery experience. It establishes ways in which the technologies used in implementing enhanced search and browse functionality can be abstracted to work in other digital collection environments. Collins, Mulholland and Zdrahal (2005) propose Semantic browsing of digital collections. Here the use of knowledge level resource descriptions to identify relevant resources and create structured presentations. Similarly, Bertini and Others (2010) demonstrates a novel web-based tool that allows a user friendly semantic browsing of video collections, based on multimedia ontologies, concepts, concept relations and concept clouds.
5.5.1 Evaluation of tools selected for Prototype
Existing tools at VDL DSpace: DSpace (http://www.dspace.org) is a digital repository platform jointly developed by Hewlett-Packard and MIT Libraries. DSpace provides the basic functionality required to operate an institutional digital repository, and is intended to serve as a base for future development to address long term preservation and access issues. On November 4, 2002, the system was launched as a live service hosted by MIT Libraries, and the source code made publicly available according to the terms of the BSD open source license, with the intention of encouraging the formation of an open source community around DSpace (Tansley, et al., 2003).
Selection criteria and Evaluation of Prototype Tools The following selection criteria adopted for evaluation of KOS tools that should possess attribute such as: (i) tool should be Java based and (ii) can enable aspects of KOS such as Information visualization and Concept maps.
The purpose here to opt for Java based tools due to the following reasons:
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
190
Chapter 5: A Prototype Hybrid KOS Model for VDL
DSpace is a Java based tool; hence, developing and integrating a prototype model using Java based KOS tool would be much easier. This demand by the platform i.e. DSpace required the researcher to look for a Java based KOS tool for evaluation.
Advantages of Java based tool o is the common programming language used by developers and also helps to create dynamic apps for multiple purposes. o as most of Java technologies are under the GNU General Public License. o one characteristic of Java is portability i.e. programs written in the Java language can run on any hardware/operating-system platform.
Based on aforesaid selection criteria, the following four potential candidate tools were identified and short listed for evaluation:
a. IHMC CMapTools: The IHMC CMapTools (http://cmap.ihmc.us/) program empowers users to construct, navigate, share and criticize knowledge models represented as concept maps. It allows users to, among many other features, construct their Cmaps in their personal computer, share them on servers (CmapServers) anywhere on the Internet, link their Cmaps to other Cmaps on servers, automatically create web pages of their concept maps on servers, edit their maps synchronously (at the same time) with other users on the Internet, and search the web for information relevant to a concept map. The IHMC CmapTools client is FREE for use by anybody.
b. JavaScript InfoVis Toolkit: The JavaScript InfoVis Toolkit (JIT http://www.thejit.org) provides tools for creating Interactive Data Visualizations for the Web. JIT was created by Nicolas Garcia Belmonte student of Computer Science at the Buenos Aires Institute of Technology, in Argentina. The toolkit implements advanced features of information visualization like TreeMaps, an adapted visualization of trees based on the SpaceTree, a focus context technique to plot Hyperbolic Trees, a radial layout of trees with advanced animations called RGraph and other visualizations.
c. SpaceTree: SpaceTree (http://www.cs.umd.edu/hcil/spacetree/) is one of the HCIL Products available for licensing by the Office of Technology Commercialization at the University of Maryland. It is a novel tree browser that builds on the conventional layout
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
191
Chapter 5: A Prototype Hybrid KOS Model for VDL
node link diagrams along a single preferred direction. It adds dynamic rescaling of branches of the tree to best fit the available screen space and the use of preview icons summarizing the topology of the branches that cannot be expanded. In addition, it includes integrated search and filter functions.
d. Visual Understanding Environment: The Visual Understanding Environment (VUE
http://vue.tufts.edu/) is an Open Source project based at Tufts University. It is a free, open source concept and content mapping application written in Java, developed to support teaching, learning and research and to organize, contextualize, and access digital information. It provides a flexible visual environment for structuring, presenting, and sharing digital information along with their metadata. Table 5.2 Evaluation of tools selected for Prototype Tool Name SpaceTree VUE CMap JIT
Type Commercial Open Source Open Source Open Source
Features tree browser concept mapping interface concept mapping interface creates interactive data visualizations
After thorough evaluation and review of tools selected for prototype, the researcher selected JIT tool to integrate with DSpace used by VDL based on the existing features. This tool enables to show the visualization aspect of communities of VDL in tree structure. To achieve this feature over DSpace, the outcome of the proposed model showed that it has enabled users to visualise the communities without navigating to different pages using various steps.
5.6 DEVELOPMENT OF PROTOTYPE HYBRID KOS MODEL FOR VDL Based on the aforesaid selection criteria, evaluation of KOS tools were conducted for prototype study. After thorough evaluation and review of these tools, the researcher selected JIT tool to integrate with DSpace used by VDL based on the existing features. This tool enables to create interactive data visualizations that help users to visualize a Concept Map like structure for the existing VDL communities in a tree structure showing relationship among communities without navigating to a new page.
VDL uses DSpace to store its content i.e. ETDs. To enhance the user experience through visualisation of communities, JIT, an information visualization tool is integrated
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
192
Chapter 5: A Prototype Hybrid KOS Model for VDL
using wrapper application to show the communities as Concept maps and to visualize the communities with its relative entities in a single window showing their levels and its relationships as shown in Figure 5.4.
Figure 5.4 Architecture of prototype hybrid KOS model for VDL
Figure 5.5 Logical workflow in prototype hybrid KOS model
The JIT tool will call the servlets which are residing in the application server i.e. Tomcat of DSpace. As a response, XML format will be send to the JIT tool. This XML format will be recognized by JIT tool which is internally using SpaceTree. With this XML Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
193
Chapter 5: A Prototype Hybrid KOS Model for VDL
format, JIT tool will display the content in a graphical format to visualize the Concept map as shown in Figure 5.5 & 5.8.
This is achieved through developing a java based wrapper application with the required code to communicate between DSpace and JIT. To the existing system, a link called “Visualize” is added on VDL homepage as shown in Figure 5.6.
Added a link to visualize VDL Communities
Figure 5.6 VDL Homepage in Prototype model
5.6.1 Benefits of Prototype hybrid KOS model The prototype hybrid KOS model helped to overcome the limitation of KOS in VDL and also provide solution to the gaps that are identified in Section 5.2.5. Some of the features and achieved benefits of the prototype hybrid KOS model for VDL are:
Visualisation aspects with different set of views (see below screens)
Concept maps show relationship among the communities
Metadata and Content can be fetched from VDL (DSpace) in XML format and this can be passed through other Open Source tool using wrapper application to visualise the communities as concept maps like tree structure.
Users can visualize the VDL communities within the same browser.
Reduced the number of steps to reach the required community
Easy navigation by reducing movement of screens to different windows
Brings normalisation among variations in KOS and enhance user experience.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
194
Chapter 5: A Prototype Hybrid KOS Model for VDL
In VDL eTheses Project, the current observation (Figure 5.7 and Appendix IX) shows users have to navigate through various pages to reach the required community or document. This process involves 7-8 steps consecutively moving to individual web pages each time. Hence, users need to spend substantial amount of time and there is no single user interface (UI) to see the below communities and its relationships.
Figure 5.7 Existing Communities in VDL
Figure 5.8 Visualization of Communities
As a solution to overcome the above limitations, the prototype hybrid KOS model is developed and integrated JIT tool with DSpace to achieve the Visualization of Communities (Figure 5.8). Thus, this prototype helped to attain the goal to enhance user experience and IR process of VDL in more user-friendly and intuitive ways.
5.6.2 Brief overview of Prototype hybrid KOS model Details of different views of communities are explained here with the screenshots. Steps to view the tree animation are as follows:
A static JSON Tree structure is used as input for this animation.
Click on a node to select it.
Users can select the tree orientation by changing the select box in the right column.
User can change the selection mode from Normal selection (i.e. center the selected node) to Set as Root.
Drag and Drop the canvas to do some panning.
Leaves color depending on the number of children they actually have.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
195
Chapter 5: A Prototype Hybrid KOS Model for VDL
Different views using Normal selection mode Users can select the mode i.e. Normal or Set as Root and choose the tree orientation in to Left, Top, Bottom and Right views. This will enable users to view the tree structure in different views of visualisation of communities as shown below.
Figure 5.9 Visualization in Top view
Figure 5.10 Visualization in Bottom view
Figure 5.11 Visualization in Left view
Figure 5.12 Visualization in Right view
Sample views in both Normal and Set as Root selection mode Another advantage of this prototype is it also enables users to see tree structure in different views and facilitate users to select community in two modes i.e., Normal and Set as root to see the tree structure. All these features can be viewed in a single user interface without navigating to a new web page. Below screenshots show the sample views of both selection modes i.e. Normal or Set as Root and with different tree orientation. In case of Normal mode, the child will be shown as sub-set of parent where as in Set as Root mode, the selected node will become a root (like parent node) and all other nodes will be pointed as sub-set under the selected node as shown below. Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
196
Chapter 5: A Prototype Hybrid KOS Model for VDL
Some organizations using the JIT toolkit are the White House and the Mozilla Foundation to create interactive data visualizations for the Web such as TreeMaps, SpaceTrees, Hyperbolic Trees, RGraphs and other visualizations.
General Studies as Child node
Figure 5.13 Visualization of communities as Concept map in Left view with Normal mode
General Studies as Parent node
Figure 5.14 Visualization of communities as Concept map in Left view with Set as Root mode
5.6.3 Enhancing user experience through browsing With the proposed hybrid KOS model, a new service can be easily added to the pool of components available to the user interface to enhance the IRS. This integration has enhanced the user experience by branching out all the VDL communities in to a tree like structure automatically in an intuitive manner within the same window. Thus, users can easily visualize all the VDL communities as Concept maps that show relationship among the Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
197
Chapter 5: A Prototype Hybrid KOS Model for VDL
communities in a single user interface without navigating to new pages involving fewer steps. This is just a prototype and the same can be used for a practical implementation in future.
Few examples that enhance user experience through integration are: Ginsburg (2004) describes a prototype, the Technical Report Visualizer System (TRV), which uses a wide variety of open standards to show DL classification metadata in the navigation interface. The author displays the topic metadata in a Java hyperbolic tree and makes use of XML conceptual product and implementation product standards and specifications, such as the Dublin Core and BiblioML bibliographic metadata sets, XML Topic Maps, Xalan and Xerces, to link user navigation activity to the abstracts and full text contents of the articles.
Warren and Alsmever (2005) examine the main research challenges in the field of digital libraries and identify the extent to which semantic knowledge technology can be used to respond to these challenges. Functionality developed within the SEKT project (http://sekt.semanticweb.org) is used to enhance an existing digital library.
Deng and Li (2008) analyze the changes of user environment in digital library and the role of the user experience; the scheme to enhance the user experience of DL is discussed from several aspects, such as polishing up the Web site design, enriching the library collection, improving the functionality of the retrieval system and so on. Similarly, Cassella and Calvi (2009) in their conference report on ECDL 2009 (The 13th European Conference on digital libraries http://www.ionio.gr/conferences/ecdl2009/index.php) provide details on the theme about Enhancing Digital Libraries Users' Experience.
5.6.4 Suggestions for Further research in KOS There are several possible directions available for further research work in KOS and DL arena. As indicated before, aspects of KOS and its features are broad in nature, thus it is difficult to cover all aspects of KOS within the scope of research study. An attempt has been made here to cover the most important aspects of KOS features such as Concept Maps or Information Visualization. Other uncovered KOS concepts like Classification, Categorisation, Taxonomy, Ontology, Thesaurus, Search capability, User Interface and other aspects need to be explored in order to get the fruitful results and may be planned for further research.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
198
Chapter 5: A Prototype Hybrid KOS Model for VDL
This research study was carried out to demonstrate the selected KOS aspects for only one ETD i.e. VDL. Similarly other ETDs, Digital Libraries, Institutional Repositories, Open Access Repositories, etc. can be considered and the prototype of hybrid KOS model may be integrated to test for further research.
5.7 SUMMARY As many research projects have made effort to integrate tools to provide features handling browsing and searching, to the best of our knowledge, none of them predominantly explores services in DLs and ETDs using Open Source tools such as DSpace in addressing KOS features. This attempt has been for an ETD case viz., VDL to integrate to showcase the KOS features that can enable users the effective information retrieval. Hence, an attempt has made to provide KOS features such as Concept maps or Information visualization to enhance user experience for VDL users. This research investigates using a simple visual interface in order to integrate with an existing digital library i.e., VDL to facilitate KOS features for easy information retrieval (Gunjal & Urs, 2012).
The output of the preliminary study shows that KOS plays an important role in DL architecture to enhance IRS and its performance. ETDs being DLs are mostly metadata driven and are facilitated search with either own or by third-party search engines. But KOS concepts like Thesaurus, Visualisation, Auto-Categorisation, Text Categorisation, Search Strategy need to adopt in order to get the fruitful results. This requires further study to make an in-depth understanding on Thesaurus, Text Categorisation and Visualisation aspects and mapping of these concepts to the DL and come up with a hybrid KOS model for ETD case viz., VDL for further development.
Thus, the hybrid model with different aspects of KOS deployed in DL helps to visualize the VDL communities through which it enhances the user experience and provide better results in more user-friendly, interactive and intuitive ways and this model may be a suitable solution for the further development of VDL in future. The proposal intends to address need of a KOS model for ETDs to support for effective and comprehensive access to multilingual, multi-disciplinary, distributed knowledge resources of any ETD database. The leveraging of larger-scale KOS is a critical missing factor. The application cluster will span the lifecycle from creation to use of knowledge in all the areas of ETDs.
Knowledge Organisation Systems in Digital Libraries: A Case of ETDs
199
