Open Source and Digital Libraries

Open Source software (OSS) in libraries allows library users to modify how an application works. This can be done by modifying the source code. Source code is the code that programmers can modify to alter how a program or application works. In open source software, the average user can modify software to suit their needs. Some advantages of open source software are that it is cheaper than commercialized software, the source code is open to libraries so that it can be modified, libraries do not have to pay licensing fees to modify the software to their specifications, and it helps libraries convert to digital form (Kamble, Raj, & Sangeetal, 2012). OSS provides many advantages to libraries, however there are several disadvantages as well. OSS systems require knowledge and expertise to maintain and upgrade, unlike commercial software companies OSS does not provide technical support to assist with problems that may arise, because software is personalized by users (Kamble, Raj, & Sangeetal, 2012)..

Open Source Software in a library setting generally provides libraries with modules for circulation, cataloging, acquisitions, and an interface for online public access catalogs. Libraries must choose software that meets their specific needs, as well as the needs of their patrons’. Staff must also be trained to use the software to achieve maximum effectiveness and efficiency.

References
Kamble, V., Raj, H., & Sangeeta1. (2012). Open Source Library Management and Digital Library Software. DESIDOC Journal Of Library & Information Technology, 32(5), 388-392.

Multimedia and XML

Extensible Markup Language (XML) is a markup language that allows users to share resources in different formats (Banerjee, 2002). The goal of XML is to create a markup language that is easily usable over the internet, can support many different applications, is compatible with other markup languages such as SGML, and is compatible with easily creatable programs that process XML documents (“Extensible Markup Language (XML) 1.0 (Fifth Edition)”, 2013). This has become a much needed tool in libraries because of the increase of the use of electronic resources in the last decade. “XML establishes rules for defining new formats. In regular markup languages such as the Hyper-text Markup Language (HTML), authors must use certain tags that make the text bold, create click-able links, draw tables, apply style sheets, etc. With XML, authors are free to make up their own tags and attributes if they don’t feel that those that have been created by others meet their needs (Banerjee, 2002).” XML is a very useful tool because libraries can store electronic materials using XML in a standardized format (Banerjee, 2002). Users can also define their own display when using XML (Banerjee, 2002). The OCLC has supported an initiative called the Open Archives Initiative that allows users to send a query to databases through the web and receive results in XML format (Banerjee, 2002). XML is a potentially revolutionary tool because of its compatibility and ease of use. XML and new markup languages similar to XML could be the future of libraries and how we share electronic information on the internet.

Banerjee, K. (2002, September). How Does XML Help Libraries? Retrieved July 6, 2014, from http://www.infotoday.com/cilmag/sep02/Banerjee.htm

Extensible Markup Language (XML) 1.0 (Fifth Edition). (2013, February 7). Retrieved July 6, 2014, from http://www.w3.org/TR/REC-xml/

Web Retrieval

spiderindex

Before the advent of the internet and search engines such as Google, a person relied on memorization and the use of friends or experts to find the answers to their everyday questions (Wegner, 2013). People also shared information with members of their social group in hopes that they would remember a portion of the information being shared. Internet search engines now allow the user to retrieve information at will without remember specific facts, relying on members of ones social group to remember, or relying on an expert such as a repair man (Wegner, 2013). While the Internet is useful tool for retrieving information on any subject a person can imagine, there are some challenges. There is a vast amount of volatile data that includes links or file names that have changed or disappeared altogether (Baeza-Yates, 2011). Distributed data on different networks varies in terms of reliability because of the amount of available bandwidth,  which is due to an  inconsistent topology of computers linked together (Baeza-Yates, 2011). In addition, there is a large volume of date available on the web, which makes scaling difficult. Data on the web is also redundant and duplicated, or it can be inaccurate or outdated. In order for a user to retrieve the data they are seeking the must express an effective query, and be able to interpret the results effectively so that they can find the most reliable information possible.

 

References

Baeza-Yates, R., & Ribeiro-Neto, B. (2011). Modern information retrieval: The concepts and technology behind search (2nd ed.). New York, NY: Addison Wesley.
Wegner, D. M., & Ward, A. F. (2013). How Google Is Changing Your Brain. Scientific American, 309(6), 58-61.

 

Indexing and Searching “The Chicken or Egg”

Today, I’d like to review an article that reexamines how users search for images. This article is called “Chicken or egg theory: do we truly know how the search?” by Elaine Menard. In this study Menard attempts to examine the search habits of image searchers so that he can obtain knowledge for developing a bilingual taxonomy using image indexing. This approach could be used by users who speak one language, or users that are multilingual.

Menard first examined a seemingly obvious question of whether indexing or retrieval comes first. Even though the obvious answer seems to be indexing, Menard contends we must totally reexamine how users search. Text represents a barrier for users of different cultures or languages. Using visual images could potentially solve this problem, however methods for retrieval using pictures is still too complex for the average user (Menard, 2011).  Users usually use two methods to form their queries when searching for a picture. First they submit a drawing, and the tracking system attempts to locate it based on the drawing’s physical features (Menard, 2011). Systems using these type of low-level features are not readily available on the internet (Menard,2011). This forces users to use text queries for picture searches even though the pictures usually don’t have any text associated with them. Today’s technology for picture retrieval is still based on manual indexing since automated image content analysis is not widely available ( Menard, 2011).

Another difficulty that presents problems is when the query is in a different language from the index terms. Cross-language image retrieval (CLIR) is a helpful tool because “the language of the texts used to annotate the images should not affect retrieval: that is, users should be able to formulate a query to search for images in their native language, making the target language transparent” (Menard, 2011). Menard also believes a taxonomy should be created to help users retrieve information faster and easier. In order to create a taxonomies Menard examines how users search for images. Menard concludes that one of the main problems with image searching is the indexing of images is not suitable for image searching (Menard, 2011). He concludes three approaches for indexing images should be used, “the terminological attributes referring to the lexicographical aspects of the indexing terms; the perceptual attributes examining the physical, functional and identifying features of the object represented in the image; and the structural attributes identifying the relationships between the words” (Menard, 2011).

The results of the research on query behavior concluded that English speaking participants on average made more queries than French speakers. Chinese and Russian speakers expressed fewer complaints. Russian users formulated the longest queries, and Chinese user formulated the shortest. Menard believes that in order for image retrieval to advance their needs to be an established framework, since many users attempt to browse or use textual category labels to search for images, a new browser interface designed appropriately would greatly improve the search results of users.

References

Chicken or the egg [Drawing]. (n.d.). Retrieved from http://http://www.theblaze.com/wp-content/uploads/2013/01/chicken-or-the-egg_7.jpg

Ménard, E. (2011). Chicken or egg theory: do we truly know how they search?. Indexer, 29(4), 150-156.

 

Aboutness and Relevancy

The  term “aboutness” refers to a summation of the subject of a document. This is usually determined through the abstract of the document.

The term “relevancy” refers to whether a document meets the needs of the user’s search. The relevancy of a search is determined by the terms the user uses in their search. If the user uses keywords that aren’t keywords in the type of document they are searching for, then this will produce a failed searched or one that produces a partially successful result. Failing to use the correct subject heading when the database uses a controlled vocabulary will also produce a failed search.

Query Languages

The type of query a user is able to formulate is dependent on the type of IR model that is used for retrieval. Full text queries will not produce the same results as a keyword ranking system. A query for data retrieval and information retrieval are very different. The difference between data retrieval and information retrieval is that in data retrieval, information is not ranked based on relevancy (Baeza-Yates & Ribeiro-Neto, 2011, p.255).

Keyword-based searches are the main type of queries used in basic IR models. “Keyword-based queries are popular because they are intuitive, easy to express and allow for fast ranking” (Baeza-Yates & Ribeiro-Neto, 2011, p. 256). Keyword searches can be based on single word queries, which retrieve documents containing the word or words used in the query. In the Vector space model, it supports ranking of the results retrieved in the search. It does this by term frequency in a document or inverse document frequency which counts the number of documents in which the word appears (Baeza-Yates & Ribeiro-Neto, 2011, p.257). The third way of interpreting search queries is conjunctive form in which documents only match the query if all of the words are present in the document. Many systems include the option for context queries as well, which means if the words appear close to each other in a document, the document is ranked with higher relevancy (Baeza-Yates & Ribeiro-Neto, 2011, p.257). Boolean queries are  another type of keyword query. Boolean queries allows users to use the operators OR, AND, or BUT to specify the documents they want retrieved (Baeza-Yates & Ribeiro-Neto, 2011, p.258).

Beyond keyword searching there are other forms of queries such as pattern matching, which is “a set of syntactic features that must be found in a text segment” (Baeza-Yates & Ribeiro-Neto, 2011, p.260).  In addition, natural language queries allow the user to ask a question in everyday language to retrieve an answer to their query. Structured queries are another type of query but it is mainly useful when documents have a consistent fixed structure, such as a mailing list. Hypertext queries have been made possible by Webglimpse, which allows the user to navigate and search the content of the current node (Baeza-Yates & Ribeiro-Neto, 2011, p.264).

As you can see, there are many different types of queries present in IR, however keyword searches using Boolean logic are the most commonly used by the average person.

Reference

Baeza-Yates, R., & Ribeiro-Neto, B. (2011). Modern information retrieval: The concepts and technology behind search (2nd ed.). New York, NY: Addison Wesley.

Controlled Vocabulary vs Natural language And Digital Languages: Markup language and Metadata

Controlled Vocabulary vs Natural Language

Controlled vocabulary is an artificial language in which words and phrases are tagged so that information on a specific topic can be more easily retrieved (Chu, 2010) . Natural language in IR is a query or search without the use of a controlled vocabulary, using common everyday language that people speak (Chu, 2010). Controlled vocabulary is commonly used in a library setting via subject headings. Searching using subject headings allows the user to effectively retrieve materials or information on a particular subject, if the user has prior knowledge of subject headings and is trained to use them properly. Natural language searches can be helpful for the untrained user. In natural language searches, a user can use questions in their search query such as “why is the sky blue?” This type of search would not be possible using a controlled vocabulary.

If a user was searching for the answer to the question “why is the sky blue” through the subject search of the Hillsborough County Public Library System at HCPLC.org,  the search engine  does not produce any hits. Instead to find materials that provide the answer to that question the user  would have to provide subject headings such as: optics, light–scattering, or atmospheric physics to retrieve the desired information. Using Google to answer the same question would retrieve 179,000,000  hits, and Google allows you to use natural language. As you can see if the user is skilled with the use of subject headings, a subject heading based search using a library database could be a more effective for them to narrow down more relevant and reliable information. If the user is not familiar with subject headings, using natural language searches via internet search engines may be their best option.

 Metadata and Markup languages

Metadata is most simply described as data about data. In a library setting, the most commonly used form of metadata is a MARC record. These records are used to catalog the materials  of libraries across the world. MARC records have their own tagging systems, which describes every aspect of the material being cataloged. An important form of metadata on the web is called Resource Description and Framework  (RDF). “The framework allows one to describe Web resources to facilitate automated processing of information” (Baeza-Yates & Ribeiro-Neto, 2011, p206).

Markup language is used to describe the formatting, structure, and attributes of digital information. HTML or hypertext markup language is one of the markup languages used on the web. Information is organized and structured by marks called tags (Baeza-Yates & Ribeiro-Neto, 2011, p208).

Reference

Baeza-Yates, R., & Ribeiro-Neto, B. (2011). Modern information retrieval: The concepts and technology behind search (2nd ed.). New York, NY: Addison Wesley.

Chu, H. (2010). Information representation and retrieval in the digital age (2nd ed.). Medford, N.J: Published for the American Society for Information Science and Technology by Information Today.

 

IR Models: Vector Space Model

 

simplistic-term-vector-mode

In the Vector Space Model, search terms are defined as a dimension and queries or documents are expressed as a vectors. (Chu, 2010, p.115) A vector consists of values that rank an item’s importance. Term weighting in the vector space model can be ranked by term frequency or by the user’s perception of the term’s importance.

Strengths of the vector space model include: Boolean logic is not required to perform a search, terms are weighted by their relative importance, a user can limit the size of the retrieval output, and relevance feedback can improve retrieval performance (Chu, 2010, p. 116).

Weaknesses of the Vector Space Model include: terms are assumed to be independent from each other,  a difficulty expressing phrasal relationships due to the absence of Boolean logic, the weighting mechanism is subjective and complex, and several terms are needed to represent a query for good information retrieval.

While the vector Space model provides several options that Boolean can’t provide such as ranking the relevance of documents retrieved and providing relevance feedback, “such systems, however, have not been able to markedly outperform systems based on the Boolean logic model” (Chu, 2010, p. 124).

 

References

Chu, H. (2010). Information representation and retrieval in the digital age (2nd ed.). Medford, N.J: Published for the American Society for Information Science and Technology by Information Today.

Simplistic Term Vector Space Model [Graph]. (n.d.). Retrieved from http://http://moz.com/blog/lda-and-googles-rankings-well-correlated

User Interface for Search: User Search Habits

 

Image

I believe that textbook examples of user search habits differ greatly from my real world experiences of helping patrons retrieve information. In the textbook titled Modern Information Retrieval by Ricardo Baeza-Yates and Berthier Ribeiro-Neto, the authors state that users’ information search methods begin with identifying the problem, articulation of information needs, query formulation, and evaluation of results. (Baeza-Yates & Ribeiro-Neto, 2011, p. 23) While sophisticated users may formulate their searches this way, I feel that inexperienced users try to type in  long phrases instead of limiting their search terms to one to three key terms needed for the search. This will often lead to a failed search, or them not retrieving the exact information they are seeking. According to the text, “One common observation… is that users often reformulate their queries with slight modifications, since this can be easier than trying to specify the query precisely in the first attempt” (Baeza-Yates & Ribeiro-Neto, 2011, p. 25). Instead of reforming their query as the text suggests, patrons at the library in which i work will often give up on the search or seek assistance from a librarian. Librarians should assist users by suggesting potential search terms, and explaining how to formulate a query so that they can obtain the proper results. Librarians can also recommend search engines and databases that could provide more effective results. Hopefully, once the librarian has explained the basics of the search process, the user will be competent enough to effectively search now and in the future.

Most libraries provide basic database search classes that can help patrons formulate a search properly by identifying their information needs, and then use the necessary terms to obtain the information they seek. This along with schools teaching children good search habits at an early age will help a new generation of users perform effective searches.

References
Baeza-Yates, R., & Ribeiro-Neto, B. (2011). Modern information retrieval: The concepts and technology behind search (2nd ed.). New York, NY: Addison Wesley.

Information Representation

In a library setting, information retrieval is the process of  using keywords, subject headings, and other queries to recover relevant information from a database on a particular subject (Baeza-Yates & Ribeiro-Neto, 2011, p. 3) . Libraries have databases of their own materials called the Online Public Access Catalogs (OPAC), which allows the user to search for materials using a controlled vocabulary or subject headings, keyword searches, and Boolean searches. Controlled Vocabulary is ” an artificial language, whose vocabulary, syntax, semantics, and pragmatics are limited.”(Chu,2010, p.19-20) Keyword searches are also commonly used in today’s library OPACs, along with Boolean searches, which enable the user to use the operators AND, OR,  NOT  between keyword search terms to narrow the results of their search (Chu, 2010, p. 5). Another form of searching becoming available on the internet is natural language. This type of search allows the user to form queries with common language, such as “why is the sky blue”. This feature still has not been developed to its full potential, but perhaps one day it will revolutionize how we search for information. However, The challenge today for libraries, is to provide database searches that are user-friendly. The interface, or what the user sees, hears and touches when using an electronic device is often what determines whether a system is “user-friendly”. (Chu, 2010, p.18).

Social media is another way to obtain user feedback about practically anything . Indexing on the internet often leads users to social media sites. In today’s society, social media provides user feedback to all different types of services they wish to obtain. These services include: books, videos, travel services, restaurants etc. Social media is useful in this way for providing users ratings of the quality of services that companies provide.

References
Baeza-Yates, R., & Ribeiro-Neto, B. (2011). Modern information retrieval: The concepts and technology behind search (2nd ed.). New York, NY: Addison Wesley.

Chu, H. (2010). Information representation and retrieval in the digital age (2nd ed.). Medford, N.J: Published for the American Society for Information Science and Technology by Information Today.

 

 

Create your own website, free on WordPress.com.Create your own website.
Start free