Searching CAplus: A Cost Analys

Article excerpt

Tibor Toth (tibor.toth@pliva.hr) and Sonja Katanec Frankovic (sonja.katanec@pliva.hr) are Information Specialists at the Zagreb, Croatia-based Pliva's Research Institute.

Comments? Email letters to the Editor at editor@onlineinc.com.

Online information retrieval should always be rapid, and yield good results. This, of course, is said than done. A hastily-constructed search can mean poor retrieval quality while a well-performed and thought-out search can require more time and money. It's the goal of every searcher, therefore, to optimize the information retrieval process.

Back in the heyday of traditional online database access, the problem was less complicated. The choice of interface, not to mention of source, was minimal. These days, however, the Web has influenced producers to serve up a varied array of product channels and pricing structures. Even those who. prefer a serendipitous approach to searching are sometimes forced to take time to research a cost-effective approach to a project. Thanks to the Web, the selection of an inexpensive, quick, and quality search has become more complex.

Take Chemical Abstracts (CA), for example. Through STN, we have access to the database (CAplus for the purposes of this analysis) by Messenger command language (http://www.cas.org/stn.htm) and via STN Easy's Web interface (http://stneasy.cas.org; http://stneasy.fiz-karlsruhe.de; http://stneasy-japan.cas.org). The same database is also accessible within the SciFinder service directly from the producer, Chemical Abstracts Service (http://www.cas.org/scifinder/scicover2.html).

What are the pricing structures for each? Which is the most cost-effective to use? And under what circumstances? By analyzing each of the three channels, we can gain insight into the pricing strategies behind each interface and, at the same time, discover a savings-friendly route to the CAplus database.

We've focused on retrieval from the CAplus database because it is more up-to-date than CA. Also, access to CA through other hosts--Dialog, DataStar, ESA-IRS, etc.--is not covered due to their incomplete records and absence of abstracts (only bibliographic data and indexes are offered).

COMMAND LANGUAGE SEARCH

Basic Features

* Designed primarily for information professionals; familiarity with the host command language (Messenger) is required.

* Special agreement has to be concluded with the STN host.

* Special front end software (STN Express) is useful.

* The language enables finding solutions to both simple and very complex queries.

* The profile can include a large number of keywords.

* Keywords can be retrieved separately step-by-step and the resulting subsets can then be freely combined into very complex logic expressions. To obtain better results, the classical Boolean operators can be used, as well as all proximity operators (adjacency, sentence, paragraph, link, etc.).

* During the search, multiple results, narrowing, and broadening is possible by introducing additional keywords and conditions into the search process.

* In addition to the CAplus database, other bibliographical databases may be introduced, simultaneously searched, and duplicates removed from results. Search results can be further analyzed using additional suitable tools, since the command language offers a number of other benefits. (Note: The cost of CAplus is the same as the CA database, plus 5% added to the total session.)

Search Costs

The STN host offers three versions of CAplus database search. They differ only in the rate charged for the search.

Standard CAplus: Connect time is charged $0.50 per minute (taxi-meter), making the search a function of time. The unit price is $2.11 for full-record display and printout. Thus, the cost is also a function of the number of printed records. The price paid for each search term is $1. …