This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Purpose - The purpose of this paper is to review the interface design aspect of library-based information retrieval system. Some recent design guidelines were synthesized under the usability view.
Design/methodology/approach - This paper introduces the information system retrieval concept firstly. These include the online database and search engine thought. The literature review on the interface designs are discussed next. Then we summarize a few recent salient design guidelines for information retrieval system.
Findings - We found out that interface design for information retrieval has not been entirely acknowledged by most of people participating in designing system. There are just a few researchers investigated into this matter and no consensus to which design guideline would be the best for users under their usability view. Meanwhile, human computer interaction techniques could help to improve the design interface for information retrieval system.
Practical implications - The study suggests that user-centered design would be a good choice for researcher and designer who want to approach a better design interface for information retrieval system.
Originality/value - This paper provides an overview of empirical researches of interface design guideline for information retrieval system. It helps researchers and designer come close to some latest design guideline in which user-centered design was considered as the core element to be developed.
Keywords Interface design, design guideline, information retrieval
Paper type Literature review
Life becomes more modern day by day and among many other humanity interests, people likely want to get more information whilst the time looking for it have to be reduced. According to the Information Today's report, the number of online searches had also increased from approximately 7.5 million in 1982 to 90 million in 1998. The numbers of online databases, producers and vendors also have expanded considerably from 301 to 12,417; 200 to 4,017; and 105 to 2,891 respectively since 1975 . Additionally, in the 5th of these surveys (April 1996), 32.4% of the 11,700 respondents to the question 'What are the main problems with using the Web?' selected 'Not being able to find a page I know is out there' from a range of prompted replies . That arises a need for the improvement of online information retrieval (IR) system where the interface design, performance, accurateness are indispensable key elements. For many people, especially novice users, the current web interfaces still are too difficult for them to learn, remember and use even though this requirement is one of most fundamental functions for any websites.
This paper introduces a study about some latest interface design guidelines for IR system proposed by IR researchers. They started realizing the effect of usability evaluation of human computer interaction (HCI) technique on the IR system and trying to apply those methods to improve its design. However, among a few well-proven design guidelines, there is no ideal solution for this matter yet even the design was approached from the user-centered aspect.
2. Information retrieval concept
Information retrieval is the science of seeking for needed information in documents within the database, whether relational stand alone databases or networked databases such as Internet. In the recent year, hundreds of millions of people join in information retrieval every day when they use a web search engine to find out the information located around the world that they want. The field of information retrieval is deal with providing users in browsing or sorting document selection or further processing a set of retrieved documents. Given a set of documents, collecting is the task of coming up with a tidy grouping of the documents based on their contents. It is similar to arranging the books according to their subject on a bookshelf.
The aim of IR system is to provide documents from the database that are concern with user information need, communicated to the system. In other words, the system is intended to identify which documents the user should read in order to satisfy their information needs which is the user wish to know more, and is differentiated from a query, that is what the user conveys to the computer in an attempt to be in touch the information need.
A fundamental characteristic of information retrieval system based on the web is that to provide the end-user can easily access the searching. These systems generally guide users to access to facilitate multiple interactive search strategies, a different type of databases, multiple manipulations of output, offer interactive help, and provide iterative movement of links (Xie and Cool, 2000a). Even though the system can develop these features, web-based interfaces are still not easy to learn, use and recognize that are reported by some recent studies (Ahmed, 2006; Ahmed et al., 2006a). The enhance IR interface design will help user searching.
There are a few studies has compared about web and non-web interfaces in different aspect of IR system. Dial-up access and web access method are compared by Koehler and Mincey (1996). Then, they were concluded that web-based FirstSearch was a major improvement over the dial-up access. Consequently, the design of web-based IR interfaces needs to facilitate human-computer interaction for system role and consider user need.
There are two different types of online information retrieval (IR) systems: online databases and Web search engines. Web search engines include directories, search engines, meta-search engines, and specialized search engines. These facts pointed that three elements are essential to users in the evaluation of online IR system which are interface design, system performance and collection coverage. The user who has already got experience in using the system, the design of system, the volume of information covered in the system, interface design all are affect on the system performance. The design of IR system is implemented by based on the discussion of advantages and problems of web search engines and online databases.
Online databases are an Internet-based resource for all the various type of information that the users desire to know. Online database have more experienced that having a good understanding of the way interfaces than web search engines. Most of user either novice or expert prefer that the simple, usable, useful and intuitive interface like Google, otherwise they may complain about the complicated design rules and codes in using online database.
Search engine for large commercial networks and web are generally based on database model of text retrieval in which documents from the network are copied to a source database, where it is guided and made searchable. Three web search engines with their search capabilities and retrieval performance are evaluated by Chu and Rosenthal (1996). These search capabilities contain Boolean logic, field search, word and phrase search, truncation and retrieval performance that consists of precision and response time that included effectiveness and usability. Search engines can help you simple ways to understand how to search using basis English unlike online databases using annoying codes that do not work. As an example: Google is the most popular among search engines. It was highly praised for: Its simplicity and ease of use. One subject well described it: "Google is my preferred search engine web site. I usually use it for everything search I want and need. I like it because it is not complicated and so simple. You can search for Web, Images, Maps, News, Books, Translate, Gmail, Groups, Document or Blog by simply clicking the corresponding tab. Simplicity is priceless".
3. Design principles for information retrieval system
3.1 Literature review
Several IR search systems like Dialog and ESA/IRS were firstly released in the early 1970s. However, because of the complexity of the interfaces, they are not familiar with most of the users except the skilled intermediates who are able to interact with the command query. Later, some simplified interfaces based on menu were introduced in the 1980s. BRS/After Dark, Dialog's Knowledge Index systems are some examples that follow this trend. In the 1990s, non-Boolean searching and web-based systems like Dialog's TARGET, Dialog, ESA/IRS, Orbit, Mead Data Central's FREESTYLE, and Westlaw's WIN were launched to provide a more accessible interface to end-users. Besides the skilled intermediates, naÃ¯ve users also feel simple to interact with the system through simple query.
A few studies have focused on the differences between IR interfaces. Canter et al.  found out a menu-based interface is more useful than command or natural language for naive users. Spavold  explained that students understood menu choices quickly and felt secure in navigating databases. Xie and Cool  concluded that some functions of the web interface help user perform better than the non-web interface. To compare novice and expert students' searching, Sutcliffe et al.  and Ahmed et al.  showed that the search performance was poor generally but expert student's had better result than novice one. This specifies the need of improving the interface for IR system in supporting naÃ¯ve users. Their needs and system functions should be considered carefully while designing the interface for IR system. The simplier of interface system got, the easier users could remember, learn and use.
User interface guidelines in information retrieval
Several guidelines of interface design for IR system have been presented so far but the value of them or the confidence of applying those guidelines to a real system still is a controversial issue. For example, Denning et al and Ahmed et al design guidelines were reported in 1998 and 2006 respectively. Both of them are a short list of guideline to design interface for IR system but it is uncertain that those small sets of guideline are better than the large ones. The designer skills and experience are still required in adapting the small guidelines to real interface. Meanwhile, those skills and experience are not so necessary for designer in implementing large guideline. In either way, prototyping, user testing and new created interface are three essential elements to facilitate the user interaction on web-based IR system. Furner  suggested that in order to be a successful interface, IR system should be easy to learn and to use. Specifically, the interface should be implemented to fulfill the following purposes:
â€¢ Assistance of the user in their formulation of a query.
â€¢ Allowing the user to control the issue of a query to the retrieval mechanism.
3.2. Denning et al design guideline
Denning et al  introduced 6 principles to build and evaluate the interface design for web-based IR system named Enhanced Resource Finder. This is a part of Eisenhower National Clearinghouse for Mathematics and Science Education (ENC) project that was implemented to provide online access to curriculum materials (books, kits, software, videos, etc.) for a wide range of users' experiences including K-12 math, science educators, parents and students. Those guidelines were built on user requirements and human factors principles analysis.
Principle 1 - Help the user develop an understanding of the operation of the interface and the search process.
In other words, the interface assists users map their knowledge of the topic into the system. It will:
â€¢ Provide categories to help the user understand how the search can be structured. Information seeker might use explicit categories to form a search. Categories also make users feel easier to match their own picture of topic to the expression used by designer.
â€¢ Reduce banner and logos size and let them blend into background to avoid distracting the user's attention. Only the important functions are arranged saliently on the screen.
â€¢ Reduce the amount of required navigation by making the functions for displaying and adding search terms available on the screen used to build the search. Google instant, YouTube instant and eBay are some websites that have applied this guideline into their interface.
Principle 2 - Provide information to help users judge the value of continuing a search path.
To encourage users digging the search result, the browse list should be presented in alphabetical order, hierarchy list of subject-related content or both of them. It will help users reduce time of continuing a search path. For example, if users wanted to find a mathematic textbook, they could follow either "m" section in the alphabetical list or "science" in the hierarchical list. Besides, the number of records related to search keywords is also useful for users to narrow down the result.
Principle 3 - Assist the user in refining the search query or search topic.
Several techniques could be useful for designer here. A clear and simply instruction or a Help page for users interested in query construction would helpful for users. Additionally, displaying and changing the query on the results page could help users recall what was searched and imply what might be changed.
Principle 4 - Avoid complex navigation.
Should the webpage screen contain the long query or users have to scroll down to complete it? The designer should arrange the page content with minimized scrolling and paging. Google interface implements this guideline by locating 2 search boxes on the top and down of search result.
Principle 5 - Make the system actions explicit to the information seeker.
Grouping related controls together likely guide users what to do to complete a form. Place functions near the object acted upon . Textboxes might suggest users to type into while buttons suggest users to click. Then placing the Search and Reset buttons near the query would be a good hint of applicability.
For some users, they interested in how the Boolean operators worked in the system. This information could be explained on the Help pages but there is also a need to adapt this requirement on the result page to avoid users keep guessing.
Principle 6 - Provide verbal labels suggestive of meaning.
Verbal labels should have suggestive meanings and functions of their related objects. Each windows heading, each button label, and each textbox label are supposed to describes the type of information that those controls connected to. Sometime, global and local of the same controls also needs to be treated differently according to its context. For example, the "Help" link on the home page might have other purpose than the one located on the Result page. Users are easier to get confused if the labels are not clear in meaning.
3.3 Ahmed et al design guideline
Ahmed et al  presented 9 general principles to increase the effective of IR system after evaluating the interface design of Web of Science (http://wok.mimas.ac.uk)
Principle 1 - Strive for consistency
User interface of IR system is the first element needed to be consistent. Different interfaces likely make users (especially novice users) get distracted and lost while navigating. Besides, some other components like layout, color, and font also should be consistent across interfaces. By following the consistency guideline, users would learn, remember and use easier even on multiple interfaces.
Principle 2 - Support both novice and experienced users
It is believed that many group of users participated into IR system. Each group has different background, knowledge and experience. Therefore, the interface should be designed to assist all of those groups. Novice users would prefer a search system with simple interface and basic functionalities. Meanwhile, the expert users would like to have more options and controls to utilize more functions of system.
Principle 3 - Make the interface actions visible to the users
The query formulation should be displayed on the interface. Some users also interested on how Boolean operators worked to create the result. Thus, expression of a query becomes a general standard for results lists.
Principle 4 - Assist users in refining the search query
Firstly, the designer needs to offer ways to change the query as well as reset it. The easier steps users have to perform, the more satisfied users are. Secondly, the result page should include the summarized query so that users can recall easily what have been searched for. The query can be displayed at the top or/and bottom of the results so they might change it, this technique would help users avoid complex navigating they likely to get if they moved back to the previous screen.
Principle 5 - Offer informative feedback
Informative feedback is not only simple as the returned search result but it also reflects what happened once the search is completed and why they happened to users. In either ways, the content of search result needs to be clear and focus on the information users were looking for. The search keyword should be also highlighted in the result as users would find it easier for them to track their search and choose which contents might be related to their interested information.
Principle 6 - Offer simple error handling
Because users' mistakes are unavoidable for any IR systems, they need to be handled carefully. The error messages would be better to indicate the way to diagnose search failures and give users several suggestions how to overcome it. They also need to be clear, uncritical and not so technical to users. A simple way out from the system would help users increase their confidence of being in control and promote them explore more as they have a way out even if they try unknown options.
Principle 7 - Permit easy reversal of actions
Undoing several previous actions would be a good feature for IR system as users might recover from their mistakes without any consequences. They could refine the search after excluding some aspects that are not relevant to their desired information. Additionally, user understands the searching topic more deeply. History of queries and the ability to return to previous search result without retyping the search query are some good choices to implement this feature.
Principle 8 - Avoid complex navigation
The navigation controls like navigation buttons, menu with a few different search options should always be visible to users while they build the query. Those controls would help users reduce the number of required navigation they have to perform. Besides, arranging functions on the search screen instead of other pages also a nice sample to avoid complex navigation.
Principle 9 - Reduce short-term memory load
Describing the required format for input data and showing Boolean operators as an option are two of many techniques to reduce short-term memory load for users while they are using IR system. Those techniques help users choose items through recognition and it would be easier for them rather than try to recall.
Denning et al design guideline  highlighted some high level concepts that were introduced to improve the usability and effectiveness of IR systems. Those design concepts was presented early and still needed to be enhanced but they acted an important role in guiding the current researchers and designers.
Ahmed et al design guideline [100-2006a] was released 8 years later than Denning et al work. Therefore, some principles such as assisting user in refining the search query and avoiding complex navigations were inherited by earlier efforts. However, Ahmed et al [100-2006a] have developed several new helpful principles for designers. Those guidelines are not aimed to be the exhaustive principles but again, they introduced some high level concepts that were proved to improve the satisfaction of both novice and expert users. Besides, Ahmed et al also addressed that among many design guidelines, just "few presented a well-documented, proven methodology for IR usability engineering". A user-centered methodology for designing IR interfaces was suggested. This methodology was based on sequentially performing:
"(1) A competitive analysis of an existing IR system to perform usability testing
(2) A user task analysis based on activities during usability test.
(3) An initial prototype design drawn from task analysis.
(4) A heuristic evaluation of the initial prototype design.
(5) An interactive prototype design, incorporating input from heuristic evaluation.
(6) A formative evaluation of the interactive prototype using task scenarios.
(7) A revised prototype design based on formative evaluations, and finally.
(8) A summative evaluation of the final prototype design and a comparison of the results with the results of competitive analysis for performing the same tasks" [100-2006a]
There is no a perfect solution for the interface design of IR system yet. The design should be kept iteratively changing based on the testing result from real users. Prototyping, user testing and releasing new design are several sequential essential steps in the design process. The final version of interface design should be achieved after a few iterations.
The designer should consider the user-centered methodology to be the core element of improving the usability of interface design for IR system. Although several approaches on the design guidelines have been proposed after investigating users searching, the need of developing a better design framework still remains. Users still encounter difficulties while using IR systems. It is believed that this is not only due to the lack of users' searching skill, but also by the shortage of interface design in which usability evaluation acts as a very important factor in enhancing. Denning et al  and Ahmed et al  introduced a few wide frameworks of integrating these HCI techniques into interface design for IR system. Those could be the good practices for the followed designers and researchers to study, to develop further and to apply into their work. Ultimately, both of them aim the same target: users will be the center of the interface design, not other ones.