CD ROM Days @ ICRISAT - AGRIS, AGRICOLA & CABI Databases on CD-ROM

After INSDOC's AIS (Associateship in Information Science) course, I joined ICRISAT's Library & Information and Documentation Unit as Library Associate in 1989. My work was primarily two fold. Conduct Literature Searches for Icrisat scientists and Research workers from other Institutions. The Literature searches were carried out using AGRIS, AGRICOLA and CABI databases on CD ROMs. Primarily FAO's AGRIS database, which was using a common Agricultural Thesaurus called AGROVOC (Agricultural Vocabulary, a Controlled Vocabulary in the field of Agriculture and allied fields). Before that there was a Printed version called AGRINDEX ..the fat bound book with a Glossy Green Cover and the nostalgic blac and whilte FAO Agris logo. We were referring agrindex only ocassionally because the AGROVOC thesaurus was built in and was a part of the software - Silver Platter. The search was based on Boolean logic. We were using the OR operator to group all the synonyms, especially the Botanical and common names of crops, and all variant forms, including hyphen, space, without space etc. I went to the extent of using common misspellings also! In one case it worked very well. We could retrieve a very useful document which without using the mispellt form of the crop name we would have missed. I was following the popular legal adage and was trying for maximum Recall. Of course the focus on recall and or precision depended on several factors, mainly the Topic of the Literature Search, for whom and for what purpose the search was being done. For projects undertaken for our own scientists, we were doing a very comprehensive search going as far as back in time, taking the help of various printed bibliographies that were the main Secondary Sources prior to the databases on CD-ROM. Example: Busseola fusca project which was a joint project. More about this work in another post. We were using extensive use of parantheses to build the logic through efficient boolean operators. We were taking a printout of the Search Strategy used and we were pasting the strip of the hardcopy paper with the `beautiful' search strategy on sheets/forms in our physical BOX file. we were also entering the topic of the Literature Search in a Register for further reference. We were using various codes and techniques like the use of Yellow highlighter pens to mark certain things and they were actually serving as POINTERS to the individual sheets filed in the Box file. We were trained to follow these simple conventions meticulously by our mentor. We have greatly benefitted by adhering to the carefully thought of practices and a well organized system. Consistency! We were the first to enjoy the fruit of our own labour because when we are working in a well organized system, work is not a stress, but a stress buster. Gratitude to our the then Senior Library Manager Ms. S. Prasannalakshmi! Silver platter software and the various fields in Agris database enabled us to narrow down the search using various labels, codes etc. We were limiting the search to various fields in a record, say Only in Title, or Only in Title or Abstracts, Author search, Names of Coneferences, place of Conference so on and so forth. We could limit our search to a particular period mentioning years. We were using geographic codes. We were using some Category Codes too specified by AGRIS database which served as Broad Subject categories. with the use of all these operators we could effectively increase our precision. After we retrieve the records we go through the entries and further refine the Search Strategy until we arrive at a reasonably good Recall and Precision ratio. We were then taking a print of these entries comprising of Bibliographic Citation and Abstracts and also Keywords. Keywords!! The most important thing in ISAR systems!! In Agris database we had two sets of Keywords - Descriptors and Identifiers. Descriptors were/are the terms that are the acceppted and preferred term listed in AGROVOC database. Primarily CONTROLLED Vocabulary. Identifers field is to give keywords that are not mentioned in Agrovoc but are felt important to input/retrieve a particular document. (At a later stage the terms given in Identifier field were reviewed for possible inclusion in the main Agrovoc) INPUT and RETRIEVAL were the two main activities and the persons doing input to Agris database are supposed to follow a common procedure across the various Agris input centres. There are specific guidelines to enter various fields. When it comes to KEYWORDS the procedure is roughly this, which of course become compiled to some extent during regular practice periods. 1) Read and Understand the content of a document 2) Pay particular attention to title, abstract and conclusion parts of an article 3) Note down all the relevant concepts 4) Look if the concepts exist as a Descriptor in Agrovoc, if yes Use it as a Keyword in Descriptors field 5) If a direct descriptor is not available, choose a Descriptor that is nearest to it 6) See if a particular concept could be represented using a combination of existing Descriptors 7) Finally, where a particular concept is not found in Agrovoc, include that term in the Identifier field and make an authority list of such terms so that some degree of uniformity or consistency is followed. While conducting a Literature search on the databases thus built, i.e. while retrieving the stored information, we adopt a rather reverse approach. We analyse a search request and identify the concepts. We match the concepts with existing AGROVOC descriptors and choose relevant ones and include in our Search Strategy. We also include other words/terms to increase the recall and these words or terms are mostly found in Titles, Abstracts and Identifier fields. We had proper control over the ISAR systems and it used to be a very fulfilling experience! Good Old Days!!