Case Study | Google Search Appliance
Parliament of Thailand enhances search experience in its Digital Library with the Google Search Appliance
LI T
Legislative Institutional Repository of Thailand
At a Glance What they wanted to do • Help users find relevant legislative documents through increased search accuracy • Combine search results from a variety of media sources and formats into accurate, unified results • Reduce redundancy and duplication of efforts during search What they did • Implemented the Google Search Appliance (GSA) to rapidly search diverse media and overcome complexities due to Thai scripting What they accomplished • Sped up the search process by up to 10 times • Increased traffic to the Digital Library by 200% • Increased knowledge-sharing in the academic society as new users from universities and law practitioners use the system for their learning and work purposes “I firmly believe that there is no other solution currently available in the market that is comparable to what Google offers in terms of usability and product life span continuity.“ —Dr. Adisak Sukul, Ph.D., Department of Computer Science, KMITL
Background The Parliament of Thailand houses an extensive collection of over 350,000 official documents dating back from as early as 1932 – the year Thailand’s Constitution began. With the number and size of documents growing dramatically every year, the National Assembly sought a reliable and scalable system to handle the massive load. This led them to embark on a project to build the country’s largest Digital Library of its legislative collection. Members of the public can access this collection on the public website http://dl.parliament.go.th/. The objectives were to digitise all information on a single platform that would be accessible to government, academic users and members of the public, in addition to the long-term preservation of these important documents while reducing valuable storage space for print copies. Challenges The National Assembly’s library stores vast amounts of valuable legislative information in various types of legislative documents that are in different formats such as documents, journals, newspapers, e-books and videos. For each type of legislative document, the Dublin Core metadata, an industrystandard set of vocabulary terms was used to standardise and refine the search experience. However, due to the complexity of the metadata, even though a search engine optimisation system was previously built using Thailand Parliament’s own repository software, users found it difficult to pull out accurate search results quickly. This was because documents frequently cross-referenced one another and were inter-related, hindering the search process significantly. Dr. Adisak Sukul, Ph.D., from the Department of Computer Science, Faculty of Science, KMITL who served on the working group as a consultant for digital collection development explains, “The differences in metadata structure for each document type made it difficult for a single search system to generate accurate search results. The Enterprise PageRank algorithm, which is a part of the Google Search Appliance, was useful in ensuring that results are accurate. The Thai language can also be complex because of how the characters are joined together. One of the drawbacks of the previous search engine was its inability to recognise the Thai script accurately, which resulted in many inaccurate search results or failed to produce any results at all.” Solution The National Assembly sought a customisable solution that could deliver a simple yet powerful search experience for their end-users. The solution also needed to integrate well with the main back-end module – DSpace Repository software – an open source, in-house system that was popular with Thailand’s academic libraries. While the DSpace repository software was able to collate the academic materials onto a single platform, one of its main drawbacks was its limited search functionalities.
About Google Search Appliance
Google.com puts the Internet’s wealth of information at your fingertips. The Google Search Appliance does the same for all your organisation’s information. The Google Search Appliance is an integrated hardware and software product designed to give organizations the productivityenhancing power of Google search. It’s an enterprise search solution as simple, powerful and comprehensive as Google itself. The latest version includes improved relevance, access to more content and enhanced security options. For more information, visit www.google.com/enterprise/search “Working with the Thai language on search engine optimisation is highly challenging. However, Google’s dedicated and intelligent Search teams worked hard to ensure that this challenge would be overcome.“ —Dr. Adisak Sukul, Ph.D., Department of Computer Science, KMITL
“We chose the GSA for the Digital Library’s search engine front-end because of Google’s reputation in the search engine space,” said Sukul. “I firmly believe that there is no other solution currently available in the market that is comparable to what Google offers in terms of usability and product life span continuity.” After switching to the GSA in 2011, the search experience for Digital Library users has improved tremendously. Consolidating content onto one single platform powered by the Google search engine sped up the search process by as much as 10 times. Additionally, the GSA’s user-friendly interface also attracted new users to the Digital Library once the system was made publicly available online. “In the past, only members of the parliament house could access the system. As we kept receiving increasing amount of enquiries from the public for legislative-related information, the Parliament of Thailand has since decided to make available non-sensitive documents on a public domain. Now we have increased traffic flow from new users such as lawyers, students and interests from non-governmental organisations,” says Sukul. The GSA was also able to overcome the Thai language barrier with its built-in dictionary and function that intelligently learns how to recognise and segregate words automatically. Results With the implementation of the GSA, information stored on multiple platforms was condensed, streamlining the search process. This has made the Digital Library more attractive to users and to date, Thailand’s Digital Library has witnessed a 200% increase in user traffic. “Now that we are only using one platform, searchability is ten times better than it used to be,” says Sukul. “We also noticed that searches conducted using Google.com or other search engines were directed to the documents in the Digital Library. This means that we are now reaching out to more people, and that a wide variety of users are benefitting from our library’s extensive collection.” The GSA’s auto complete and spell checker functions also benefit users who can now generate more accurate results in a quicker amount of time. “Auto complete and spell checker are incredibly useful functions because of how the Thai language is structured. It is not uncommon for words to be spelt wrongly and the GSA is able to recognise the most commonly used Thai words and offer more accurate suggestions making it simple and quick to search,” said Sukul. Various challenges due to the large scale of the project, complexity arising from complex metadata and the intricate nature of the Thai language, required close collaboration between the National Assembly and Google teams. “Working with the Thai language on search engine optimisation is highly challenging. However, Google’s dedicated and intelligent Search teams worked hard to ensure that this challenge would be overcome. I am very pleased with how successful the library project has been and a large part of this success goes to Google,” said Sukul.
© 2013 Google Inc. All rights reserved. Google and the Google logo are trademarks of Google Inc. All other company and product names may be trademarks of the respective companies with which they are associated. SS2082-1304