University of Maryland College of Information Studies Command Language Comparison: Dialog Classic, Gale, Google, and MdUSA Christina K. Pikas LBSC 750 Due: December 6, 2001 Introduction The purpose of this paper is to examine the differences between the command languages used by different database vendors. The database vendors were selected to provide a broad comparison in command languages and also because these are databases that the author frequently uses. The following vendors (and versions) were selected: The Dialog Corporation. “Dialog Classic Version 2.0.” 2001. http://www.dialogclassic.com, accessed December 2, 2001. Gale Group, Inc. “Gale Info Trac Web.” 2000. http://infotrac.galegroup.com/itweb/prge_main, accessed December 2, 2001. “Google.” 2001. http://www.google.com/, accessed December 2, 2001. “MdUSA.” http://mdusa.lib.umd.edu, accessed December 2, 2001. Although TheThomson Corporation owns both Dialog and Gale Group they are dramatically different. Only a small subset of the databases available through Dialog Classic is available through Info Trac Web. Dialog Classic bills by time and by output, while Info Trac Web is site-licensed for unlimited searches. MdUSA is the University of Maryland system’s shared portal for searching the majority of the databases licensed by the schools. It offers the ability to search databases from multiple vendors at the same time, the Jointly Administered Knowledge Environment (JAKE) tool which allows the user to determine which databases index which journals, and allows for remote access to the databases. Some of the databases open in the vendor’s portal, but the majority use the MdUSA interface. MdUSA is only available to registered students, staff, and faculty of the University of Maryland. Finally, Google (www.google.com) is a popular Internet search engine. The command language and advanced search options are discussed below. Expectations There are several things that all useful command languages will have. These minimum expectations allow users to efficiently find records matching the information need and to move quickly from program to program. The first of these expectations is the ability to use Boolean logical operators including AND, OR, and NOT. The second is the ability to access indexes directly. In other words, full text searching using a keyword can be an option, but should not be the only option. Preferably, access should be given to author, title, publication date, and source (i.e., journal name) indexes as a minimum. Stop words should be specified in the help screens. Phrase searching and nesting functions should be allowed. Truncation and proximity searching are also very desirable. In addition to the above functions, the command language should make sense and be easy to learn. Many vendors have flattened the learning curve by adding drop down boxes so that average users can avoid learning the advanced language idiosyncrasies. Advanced users should still be able to use the commands directly. Internet search engines should also have the features listed above. Additionally, file type searching and image or media searching is desirable. The ability to adjust set size is an important feature for speed and ease of use. 2 Differentiating Features The various features of the four command languages are compared in a table located in the appendix. Dialog Classic A user cannot use Dialog Classic unless she is well versed in the advanced command language. Online help screens are available but it is not clear if the system continues to rack up dial units while the help screens are open. A nice feature Dialog has that was not found in the other command languages was the ability to remove duplicates within the search results. Once the user is trained in the command language and features, Dialog Classic is very powerful. It has all of the required features and also has special output formats to easily export the data for labels or to spreadsheet or citation software. Unfortunately, the data is not formatted attractively for printing as in the case of Info Trac web or MdUSA. Gale Info Trac Web A nice feature of the command language is the ability to browse indexes. This is similar to the expand command in Dialog. Also, the thesaurus is available to search using the “subject guide” feature. The cross-references, broader and narrower terms are hyperlinked with the number of articles with each term listed. Also, a large quantity of the articles has permanent URLs so that they can be bookmarked and retrieved later. The relevance search feature includes automatic stemming and ranking functions. For all searches the keyword search can be over the full record or limited to just the title, citation, and abstract, the equivalent of the basic index in Dialog Classic. Google One major differentiating feature of Google is that it does not offer truncation or word stemming. The basic help page states that this feature is not offered to increase search accuracy. Google will also return pages that do not have the words that are combined with AND in the search. This way, the user will always get results. The advanced user can use the + symbol to require any word. Stop words are mentioned in the help screens but no list of words is offered. On the other hand, Google offers quite a few unique and unexpected features. First, it allows the user to specify how many occurrences in the page the word must have. Second, it has a spell checker for your search. It is not exactly clear what some of the functions in Google do. For example, how are related pages determined? Are they pages that are linked? If so, how is that different from the link: command? Beta or testing features of Google will enhance its use. The image search engine is already very useful but the translation feature could use some work. A German website on turbochargers was completely nonsensical when run through the translator. MdUSA The first major differentiating feature is that MdUSA allows the user to search multiple databases at the same time. This is also available through Dialog Classic, but is not available through Info Trac Web. By creating a Topic, the user can specify a group of individual databases besides the pre-selected groups to search. 3 Another important feature of MdUSA is the ownership index. When the full record is viewed, area institutions owning the journal or monograph are listed and the University of Maryland System Libraries owning the item are listed. From the advanced search page, the records can be automatically sorted. Also set size and record format can be selected. Records including full text can be formatted for printing or emailed to a given address. The help pages for MdUSA do not provide a list of stop words but are otherwise helpful. Conclusion Overall the four command languages performed as expected. This is surprising, because some of the expected features were not discovered until the documentation was thoroughly examined. It is also surprising that Google does not offer truncation to improve accuracy. Perhaps the designers should also eliminate OR? After doing this research the author feels comfortable recommending these command languages for any search. Dialog Classic would not be the first choice if Info Trac Web or MdUSA are offered due to system pricing. Also the time the information professional would have to allot for packaging the search would be much greater for Dialog than the other two. Google remains the premier Internet search engine. Many of its features are not equaled in other engines; however, most of the features are rarely used. 4 Appendix: Comparison Chart Feature Indexes (can vary with database) Dialog Author AU= Company CO= Corporate Source CS= Document Type DT= Journal JN= Language LN= Publication Date PD= Publication Year PY= Sales SA= Update UD= Info Trac Abstract ab Author au Keyword ke Source so Subject su Text Word tx Title ti (Can browse indexes) MdUSA Abstract Author Keyword Subject headings Subject words Title Year of publication (Can browse indexes.) Wildcards/ Truncation ? any number of characters ? ? exactly one character ?? number of characters And Or Not ( ) to nest * any number of characters ? exactly one character ! one or no characters And Or Not ( ) to nest + must contain - must not contain Wn within Nn near ? any number of characters Boolean Proximity Range Quotes (nW) within (nN) near (S) subfield (L) linked (F) same field (T) same term : between > greater than < less than And Or Not ( ) to nest Google Allintitle Allinurl Cache Date Domains Format Intitle Inurl Language Link Occurrences Related Safesearch Spell Stocks Site Not offered And (understood) Or + must contain - must not contain Wn within Nn near Note: must have number Since, after, gt, > more recently Ge on or after Before, lt, < earlier than Le on or before To, - between Yes Yes 5 Bibliography The Dialog Corporation. Dialog Pocket Guide. Cary, North Carolina: 2001. Gale Group. “Help on Help Index.” http://web1.infotrac.galegroup.com/itw/infomark/15/111/18362646w1/3!help_Help_Index, accessed November 7, 2001. _____. “Info Trac Web Fact Sheet.” http://www.galegroup.com/pdf/facts/itweb.pdf, accessed November 25, 2001 Google. “Advanced Google Search Operators.” http://www.google.com/help/operators.html, accessed November 7, 2001. Google. “Google Help.” http://www.google.com/help/refinesearch.html, accessed November 7, 2001. MdUSA. “Searching.” http://mdusa.lib.umd.edu/html/help.html, accessed November 7, 2001. 6