Design, query and evaluate information retrieval systems.
While there is a widespread perception that the internet has made information retrieval easier, it has also added exponentially to the supply of available information while simultaneously facilitating the proliferation of inaccurate and even dangerous sources. Boyce, Kraft & Meadow note, "nothing in the entire process is more important to outcome than the user's own understanding of his or her information need and ability to recognize what satisfies that need (2000, 6). While achieving “some” result may be easier, achieving accurate and precise results in many cases still requires specialized searching skills and a user that can connect to the right resources. Often there is a gap between the user's "information need" and the recognition of the best methods for satiating that need. With this in mind retrieval systems must necessarily be evaluated on how well they connect users with their needs.
The ability to effectively query retrieval systems is in part based on an understanding their underlying architecture. Working knowledge of how systems are structured and the basic tenets of authority, facet structure and subject assignment is important to understanding how to best utilize a system in order to find precise results. The preponderance of information available has had the side effect of increasing the demand for effective retrieval system design, the creation of intuitive architectures, and cross-system searching. This demand necessitates professionals who understand the basis of information organization on both the architectural and retrieval sides of systems. With this in mind, the ability to adequately evaluate retrieval systems will be valuable in many forums especially if the evaluator has the unique perspective of viewing functionality through the eyes of both users and designers. Herein, lies a crucial concept in information retrieval design; that it must commence "from at least two different points. One starting point is the user, or prospective user...At the other starting point a person or agency makes a decision to collect information" (Boyce, Kraft & Meadow, 2000, 4). Retrieval systems that keep its end users in mind and cater to that groups information goals are as important as systems that are architecturally sound.
Evidence
In my work experience, assisting the public at reference desks in multiple public libraries, I have had the opportunity to provide effective information retrieval service on a variety of ILS platforms, databases and internet search engines, often utilizing Boolean functions and creating advanced queries. I have also had the opportunity to teach the public how to perform more effective queries in both ad hoc sessions and in formal classes. These formal classes have included teaching Spanish speakers to query effectively in their native language, using appropriate search engine tools, as well as, introducing basic cross-lingual tools for developing search terms.
At my current library position I was tasked with training staff on the basic steps needed to create custom search engines through Google’s CSE portal. This basic tutorial was geared toward allowing staff to acquire the skills necessary to create basic search portals that are themed around a certain narrow area of interest. At the time the most pressing concern was the creating of a County specific employment CSE, utilizing the human resources websites of many of the county’s largest employers. This piece of evidence is demonstrative of an ability to create a valuable search function based on user need, using a customized set of information sources.
In LIBR 202, Information Retrieval, I helped design a simple functional retrieval system using Textbase, which cataloged the contents of a junk drawer of the type everyone seems to have in their house. I was responsible for imagining the conceptual framework for the junk drawer, many of its design elements and all of necessary data entry. I offered to do most of the tedious and precise labor needed for the database due to my own interest in learning more clearly how database systems function. This project too had the intent of tying a information retrieval system to an appropriate need.
A further component of my LIBR 202 class was learning to effectively evaluate search engines. For this project I looked at the search engine Gigablast, analyzing its functionality, its interface and its results display. I also used this assignment as an opportunity to compare the Gigablast search to that of Google, the clear market leader in the field. This comparative analysis was illuminating not only for uncovering some of Gigablast’s deficiencies, but because it taught me more about Google’s functionality as well.
In LIBR 244, Online Searching, I turned my focus to evaluating database retrieval systems, specifically looking at how they deal with Spanish language orthography. This analysis uncovered glaring flaws in major proprietary databases, which are nonetheless marketed to institutions to search Spanish Speakers. The failure of many of the underlying retrieval systems to manage the peculiarities of the Spanish language on its own terms leads to a number of errors where retrieval and authority records are concerned.
Also in LIBR 244 I was tasked with creating complex search queries using a variety of online searching tools and databases, varying from commercial databases to popular search engines. The various search creation assignments in this class helped my knowledge of databases structure come full circle. Allowing me to query more deeply within the types of systems I had previously looked at for design or evaluated for functionality. The evidence displays here involves the use of a wide variety of information retrieval systems.
- Custom Search Engine Tutorial Presentation
- Junk Drawer Database Description
- Gigablast Search Engine Evaluation
- Spanish Language Retrieval Problems in Major Databases
- Online Searching Exercise 4
References
Boyce, B. R., Kraft, D. H. & Meadow, C. T. (2000) Text Information Retrieval Systems. San Francisco: Academic Press.
