Two views on the acquisition of Dialog by the Thomson Corporation

Not another foster home

Since Tunbridge's editorial, of course, important things have happened to the company and maybe now is a good time to focus on the future. While the sale of Dialog to Thomson may seem like yet another step from foster parent to foster parent, this time Dialog may have found a good home. Why do I think so? Partly because, in the past, Thomson managed an elegant merger when it acquired the assets of the Information Access Company (IAC). Within a few months it had beautifully integrated the full-text records of IAC into the Biographical Resource Center (BRC) database family. True, with Dialog we are talking about a much larger scale integration of nearly 500 heterogeneous databases, but it is the know-how and competence that really matter and Thomson appears to have both.

Intra- and inter-database linking

Although Dialog developed its cross-database search facility more than 10 years ago, it has never fully utilised the potential of being the largest aggregator of databases. After an agonisingly long time it did 'webify' its system, but without the genuine traits of the World Wide Web. In its web service, there are some useful gadgets - pull-down menus, check-boxes and buttons on the query and result screens - but the most useful thing about the web is still lacking: there are no hot-linked elements. Instead, users still have to formulate a new query observing the not-so-friendly syntax rules, and carefully type the occasionally tongue-twisting personal and corporate names and the oddly abbreviated journal names when a good record from a subject search gives inspiration for another search on these elements. Intra-database linking from record elements to index entries for launching a new search is a given in most web-borne databases and all the professional web information services that I know, and such intra-database linking could be implemented overnight.

More important would be to develop inter-database linking - and here is where the Gale Group's experience and Thomson's other database assets (like the ISI citation databases) could help. For instance, when you are making a search in Books in Print, Dialog should offer to search for reviews in Book Review Index for an extensive citation list or for the full-text reviews from the Trade & Industry or the A.R.T.S. databases. When you search Social SciSearch on a psychiatric treatment, Dialog should offer you the option to repeat the search in SciSearch. From an EMBASE or MEDLINE search Dialog could link to such Thomson assets as Micromedex, Drugdex or the Physician's Desk Reference for related information about side effects or contra-indications. And so on.

Improving and extending Dialog's own databases

Dialog is known as an aggregator of other content providers' databases, but it has three very useful databases of its own: its Journal Name Finder, Product Name Finder and Company Name Finder. These databases tell you which databases have records and how many, from a journal or about a product or a company, respectively. These are taken from the journal name, company name and product name fields of databases that have such fields. They are incredibly inconsistent and erroneous, both within and across databases, and so you have to search them in a variety of formats.

Nevertheless, the results may be very well worth it because you may easily find out that, for example, INSPEC, Compendex and PASCAL have far better coverage of journals dedicated to artificial intelligence than Information Science Abstracts , which considers the topic its forte , despite the fact that it barely covers or does not cover at all many of the essential artificial intelligence journals.

He felt the Forbes article was unfair to Dialog, which "is still a powerful, viable business, and our industry would be a lot poorer without it". However, I think Lubove's criticism represented the opinion of most veteran Dialog users.

Undoubtedly, it would require significant manual efforts to consolidate these variants of journal names. However, with the help of intelligent humans and intelligent software agents, and through the ISSN and its accompanying key title, the Journal Name Finder database could be converted into a Journal Name Thesaurus. The uniform key title and ISSN would be the collocators listing all the variants found in the databases hosted by Dialog. Once in place, the journal names used by the searchers could be checked against this thesaurus and replaced by the variant(s) used in the database(s) selected by the user or offered by Dialog as suggested above. Finding the databases with the most comprehensive coverage of specific journals would also be a cinch, and much appreciated by users. Similar consolidation could be done with company and personal names.

This is not a pie-in-the-sky idea, as the Getty Foundation has done an awesome job - admittedly on a smaller scale - with its vocabulary standardisation programme for geographic names and artists' names. This exercise could serve as a model of how standardisation and consolidation can be done.

Of course, the first step should be the diagnosis of the problems, and the finding of the niche areas, followed by the development of the solutions.

Indeed, Dialog still has a lot of mileage in it, and in competent hands it can thrive.

Péter Jacsó is an associate professor at the Information and Computer Science Department of the University of Hawaii.