Chemical Literature (Chem 184/284)

Lecture 3: Using Online Databases, Part I:
Web Search Engines vs. Traditional Article Databases vs Library Catalogs

Locating Information: The Evolving Paradigms

• At one time, centuries ago, a well-educated scholar could be expected to be familiar with all the published scholarship in his or her field, to know, and perhaps even own, all the important books in that field. But as scholarship, especially in the sciences, grew, tools had to be devised to enable a researcher to locate published information held in libraries and elsewhere.
• Libraries published catalogs of their holdings, first in book form, then, starting in the late 19th century, on index cards filed in cabinets -- the card catalog.
• Scientific journals started in the 17th century. Eventually, the volume of published literature there required special tools as well. Scholars created bibliographies -- lists of references to published works in a particular subject area in a particular time period. By the mid-19th century, regularly published indexes to the journal literature enabled users to locate published articles by topic or by author.
• With the growth of digital electronic computing in the 1950's, the then-new electronic database technology began to be applied to library catalogs and scholarly indexes in the 1960's. By the year 2000, electronic catalogs and indexes had almost completely supplanted their print counterparts, and even the journals and other primary publications themselves were moving to electronic media.
• At the same time, networks of computers were increasingly linked together, growing from the U.S. military-supported ARPAnet into the Internet of today. The use of the Internet as a medium for sharing information really took off with the invention of the protocols underlying the World Wide Web in 1989. As the number of publicly-available Web pages grew into the billions, search engines were devised to allow users to locate the needles of useful information in the giant haystack of the Web.
• Today, virtually every college student is familiar with searching for information electronically, and, far and away, the most familiar search tool is the web search engine -- most likely Google™. However, powerful as it is, Google™ is not the be-all-and-end-all of scholarly searching. Catalogs and article indexes are still extremely important, and it is useful for the searcher to understand how these tools differ from one another.

Web Search Engines vs. Library Catalogs vs. Article Index Databases

General Characteristics of Electronic Catalogs and Bibliographic Databases

Command Line vs. Menu Driven

• The earliest electronic catalogs and bibliographic databases all required entering a typed command, say, f kw diels-alder, where f, standing for find told the system what function to perform, kw, standing for keyword told it which index to search, and diels-alder is the desired search term.
• Most catalogs with a Web interface use a more menu-driven approach. That is, the screen displays the options available to you and you select them by pointing and clicking. The only things typed in are the user's search terms. Note that, in most cases, the menu interface taps into a "most commonly used" subset of the command features.
• Frequently, Web catalogs and databases will offer multiple search screens -- usually a Basic Search, one or mor Advanced Search screens, which offer more options and allow more complex searches, and not infrequently, an "Expert" or other version which is really the command line form of the search system.
• Generally speaking, command line systems are more powerful and flexible, but have a steeper learning curve, since you have to memorize commands and learn system syntax before you can use the system.

Available Indexes -- What You Can Search

• Library catalogs in the English-speaking world all use the same basic record format (the MARC format mentioned above). But the individual catalog implementation decides which parts of the record are indexed so that they are searchable.
• Nearly all catalogs will have some type of author index, a title index and a subject index. Some will have additional indexes. Common added indexes include some sort of keyword index (which searches most fields in the record) and call number indexes.
• Be aware that not all author or title or subject indexes function the same way. Some search for complete phrases, some for a word or words within the field. Some automatically check for all the possibilities of an author's initials, some don't.
• Remember that in most library catalogs, the subject search searches assigned Library of Congress subject headings. If you are not sure of your terminology, use keyword (where available) or title word searches first, then check for subject headings.
• "Browse" vs. "Search" -- Many catalogs allow for browsing of indexes. Rather than retrieving a set of records that match the search, browsing lets the user look at the alphabetized list of items (authors, titles, subject headings, call numbers) starting at a chosen point. You can then select a heading from the list and view the books associated with it. Browsing can be advantageous when dealing with items that may be in a different format than you would assume, such as author names and LC subject headings. Browsing call numbers can serve as a substitute for browsing the book shelves in a particular subject range.
• Limiting searches -- Some types of information, like language and publication date are rarely searched by themselves, but can be used to limit the results of another search. Check what limits are available and how you can use them.
• Search histories -- Some systems, but not all, allow you to view previous searches in a given session. This can be handy for retracing your steps.
• Combining searches -- Some, but not all, allow you to take two or more separate searches and combine them. This adds flexibility in searching.

Speaking of Subject Headings...

• It is possible to provide subject access to documents by keywords (words used by the author in titles, chapter headings, abstracts and/or full text) only. Some databases use that approach. However, most catalogs and indexes to journal literature use some form of standardized, human-assigned, subject headings. Why?
• The problem with keywords by themselves is that different authors may use different terms to describe the same concept. For example, one researcher might speak of double optical refraction and another of birefringence, both referring to the same phenomenon. Similarly adrenalin and epinephrine refer to the same compound, and California condor and Gymnogyps californianus to the same bird. If you search for one term, you may miss relevant documents that use the other term.
• In order to pull related concepts together, indexers have devised standardized subject heading lists. When an indexer using a given system reads a document, the indexer checks the concepts that describe the document against the standard headings, and picks the one(s) that best describe the content of the document.
• Choosing the right heading can be tricky, though. If I'm indexing a book on "California condors", and I have a choice of using the headings "Gymnogyps californianus" or "New World vultures" or "Birds", which do I use? All are correct, and depending on how general the interests of the potential researcher are, any might be used as a search term. Some systems would have the indexer add search terms for all levels of the concept; others rule that the indexer should use the most specific term that describes the content of the document. Still others organize the subject headings into a thesaurus which shows the relationships among broader, narrower and related terms. Some electronic search systems use a thesaurus to allow the user to automatically search the whole hierarchy of terms for a given subject.
• But standardized subject headings have their limitations, too. Standard lists always lag behind the terminology of fast-developing fields. And since terminology may change over time, the standard terms themselves are forced to evolve -- older documents may not be indexed the same way as newer documents. Finally, how do you, the researcher, know what the standard subject headings are for your area of interest?? In the old world of printed indexes, you frequently had to consult a printed list of subject headings, or index guide or thesaurus to find the proper terms.
• In electronic catalogs and article databases, however, you can usually take the expedient of searching first using the terminology you have in hand as a keyword search. Inspect the resulting hits, pick the ones that best match your interests and view their full records in the catalog or index. There, you can see the assigned subject headings that the system used for those "good" documents. If the subject headings match your keywords, you're golden. If not, revise your search to incorporate the subject heading terms as synonyms to increase retrieval. Or, if you're getting too many irrelevant hits with your keywords, search on the subject heading exclusively to tighten the focus of your search.
• Some web-based catalogs and databases allow you to click on a subject heading and automatically generate a search without having to key in the words.
• Library catalogs in the United States use a subject heading list generated by the Library of Congress. These headings use subheadings extensively to narrow down broad topics. Note that subject headings in the sciences, especially chemistry, are not always what a chemist would have thought of first. Use the "Search keyword; check subject headings" approach for best results.

Other Search Features

• Truncation -- Most online catalogs and databases allow some kind of truncation, that is, replacing part of a word with a symbol to search for multiple words with the same root. For example, organo? might search for organochlorine, organohalogen, organometallic, etc. Some systems allow you to truncate single characters, some allow you to truncate internally, e.g. wom!n. There is little consistency as to which characters are used for truncation: * # ? ! $ are all used in various systems for various types of truncation.
• Boolean searching -- Generally speaking, most systems use the operators of Boolean algebra: OR, meaning either "term A" or "term B"; AND meaning both "term A" and "term B" must be present; and NOT meaning "term A" is present, but records with "term B" are excluded. However, not all systems are identical. Be aware of the usage on the system in question.
• Proximity -- If you enter multiple terms in a search window, some systems treat them as separate terms, some search them as phrases. Some allow you to specify the relationship of terms with proximity operators. Example: "term A" NEAR5 "term B" meaning that in a record the two terms have to be within five words of each other.
• Stopwords -- Usually words that are very common and lack subject meaning are not indexed, such as "a", "an", "the", prepositions, etc. In library catalogs, sometimes "a" "an" or "the" at the beginning of titles are omitted.

Display Features

• Be aware of how the system displays results. Usually there will be some sort of short default record display. Check on how to display full records and how to display holdings, circulation information, etc.
• Sorting -- What order is a list of records displayed in? Some systems go alphabetically by author or by title; some use reverse chronological order (most recent first.) Some systems allow you to choose the sort order.
• Marking records -- Can you mark records as you go along to select out a set for viewing or printing/downloading/e-mailing? Can marked records be used as a subset for further refinement?
• Printing/downloading/e-mailing -- Some Web interfaces simply use the print/save/send functions of the Web browser; others have special functions to handle these tasks.

Personalization features

• Many systems allow you to use some sort of password to create or access personalized functions -- such as viewing your circulation records or setting up update searches.
• These features vary widely from one system to another, as does the method for getting or generating a password. Some use on the spot registration, some use ID's like library card numbers.

This page created by Chuck Huber (huber@library.ucsb.edu).
Updated: 11/27/07 04:44:24