AN OVERVIEW OF EUROPEAN SUPPLY OF SOCIAL SCIENCE INFORMATION

On the basis of the EUROSPES Directory of Social Sciences Data Bases, issued in 1990, this paper aims to analyze the information supply throghout the data bases created in the latin language European countries. lt is divided into two parts. 1 n the first part thc study shows thc main traits of the information contained in the data bases: kind of information ofTercd, subjects covered, information technologics used, bibliographic rules. etc. The information diffusion is also studied: vendors, software. helping tools for uscrs, services olTcred for access. Thc second part is dcvoted to thc analysis of the data described in the first p;irt : a comparative study on the availability of information by country is givcn: this comparativc analysis go1.:s ovcr softwares compatibility, exchange formats, quality of access serviccs, prices, training offered, etc. It ftnishes with sorne conclusions ahout the foreseen dcvelopmcnts and also about the possibility of coopcration in order to a more complete and efficient supply of automatic information on Social Sciences.

An overvicw uf European supply of soi;ial sdence information

An overview of European supply of social science information
EUROSPES is an informal grouping of the main institutions dealing with social science information and documentation in a number of countries of Western Europe. These countries are those using mainly romance languages: Belgium, France, ltaly, Portugal and Spain. One of the main aims of EUROSPES is to promote the electronic delivery of social science information.
For this purpose, it compiled and published a Directory of Social Science Data Bases, ( 1) issued in 1990, after a first preliminary edition dated 1987. The Directory describes 173 bases, comprising 24 produced in Belgium, 65 in France, 41 in ltaly, 14 in Portugal, and 24 in Spain. In addition 3 bases produced in Luxembourg are also described (but they are not national, originating within the ECE), and 2 produced in Switzerland (but they are not fully representative of the country). In this report, reference will only be made to the five major producing countries.
The Directory intends to cover ali types of data bases: bibliographical, textual, factual and numerical. In principie only bases available to the general public by electronic means are covered, but it was found convenient to include sorne bases only planning to give this access during the next few months. All data were, when possible, provided and thereafter checked by the producers themselves; many letters were sent and received for this purpose during 1988 and 1989. lt is a matter of course that such a process does not guarantee that ali information is updated at the same moment, as the whole set of data has to be retained until the answers from the slowest correspondents have been received. However, it may be said that the Directory gives a fair representation of the situation prevailing at the end of 1989. Everybody knows that any situation, in this field is rapidly changing, and we do not deal with that of March 1991. E ven for 1989, we accept that the Directory may be incomplete and sometimes inaccurate.
A secondary analysis of the data collected in this Directory allows us to report on social science information supplied at the cnd of 1989 via electronic media by romance-language speaking countries of Western Europe. This report will <leal with three aspects: origin and coverage of the bases, tcchnologies used and linguistic contents.
Figures will be given as percentages, unless otherwise stated. Admittedly, percentages may be mcaningless when referring to so small totals as e.g. 14 bases for Portugal; but it is the only practica! way to make comparable data from severa! countries. However, account has to be taken of the limited number of producers in some countries; when there is somewhere a body producing an important proportion of thc national bases, gencrally using the same technologies, thc global percentages are ipso facto biased. This is for instance the case of Italy, where the Corte Suprema di Cassazione produces 21 bases from a total of 41, or even of France, where the lnstitut de l'lnformation Scientifique et Technique produces 12 bases from a total of 65.

Origin and coverage
In the live countries, the bases are mostly recent, if we consider the dates of their computerization which in many cases followed a sometimes large period of In four of these countries, producers are predominantly in the public sector, the exception being Belgium, where 63 % of the bases are produced by prívate companies. But 77 % of the French bases and 76 % of the ltalian ones, come from the public sector. Only one producer in Portugal against 10 and one in Spain against 16, is private. ll appears that in Latín countries, there is some slowness in establishing a profit-making information industry.
The majority of the bases remain bibliographic: 67 º!ti in Belgium, 58 % in France, 54 % in Portugal, 71 % in Spain. The exception is here ltaly, where 51 % of the bases are textual. This is due to the large number of bases produced by the Corte Suprema di Cassazione and the Camera dei Deputati, making available legal and judicial texts. Textual bases are only 17 % in Belgium, 23 °./4, in France, 12 % in Spain and do not exist in Portugal.
An analysis of the subjects covered cannot · go very far, for two reasons. First, the compilers of the Directory dccided to include some importan! bases which are not devoted to the social sciences, but are more general and contain inter alia a wealth of social science data. This is e.g. the case of POR BASE (Portugal), BIBL (Spain), ORIADOC (France), etc. The second reason is that «social scienccs» are understood with a loose meaning: a number of the bases have no connection with any science, but cover data produced hy business, hy financia! activitics, by daily life ... which could be material for analysis by social scientists (2). Not surprisingly, 12 % of the bases in Belgium, in France as well as in Spain cannot be catcgorizcd otherwise than «multidisciplinary».
The two fields best covered, in thc majority of thc five countries, are: law and jurisprudence, by 49 % of the bases in ltaly (for the specific reasons mentioned above), 25 % in Belgium, 21 % in Spain, 20 (1/o in Francc: and cconomics and society, by 28 º/4> of the hases in Portugal, 25 % in Spain, 25 1 1/., in Belgium, 17 '1/.i in Italy. Other concentrations of interest are more specific to one or two countrics, c.g. enterprises (28 % in Portugal, 1 O % in ltaly), labour and cmploymcnt ( 12 % in Belgium), parliamentary and política! life ( 15 °/ri in France). Only in Francc a number of bases are specifically dcdicatcd to onc or other among the social sciences, such as sociology, ethnology, economics, cducatinnal scicnces ...
Taking an oppositc viewpoint on the whole situation, 011c may conclude that in each of these countries, sorne importan! topics are not yct cowrcd by elcctronic information in an appropriate way: room is left for development (J) (4)

Technical and functional aspects
There is a limitcd amount of standardization, not only at thc global h.:vel of thc five countries, but also within each country. This would makc nrnrc diffícult !he organization of cooperation and the sctting up of exchange nctworks bctwcen diverse bases. Within the subset of hibliographical data bases, even the bibliographical description of the documents is poorly standardized . The number of bases purporting to follow the prescriptions of ISBD, or of national cataloguing standards (which themselves conform to ISBD) is limited: 19 º,/4, in Belgium, 53 % in France, 11 % in ltaly, 67 % in Portugal, 29 % in Spain. Sorne bases refer to UNISIST guidelines, without more precision. Ali the others use interna) rules.
Diversity is even greater in computer exchange formats . Most of the bases are unable or unwilling to specify what formal they rely on. When they mention a standardized formal, one finds 2 different formats in Belgium, 10 in France, 3 in ltaly, 4 in Potugal, 7 in Spain. Arnong them, only STAI RS is broadly represented, with the exception of France: 9 times in Belgium, 4 in ltaly, 2 in Portugal, 2 in Spain. Sorne bases mcntion MA RC. hut apparently they rcfer to different members of the MARC family; only UNIMARC is prcsent in two countries (1 base in France, 3 in Portugal). Sorne others mention ISO 2709, but this is rather a general frame hosting severa! formats. MISTRAL is used by 4 bases in France, 2 in Spain.
Sorne formats are used in only one country: BASIS by 12 bases in Spain, DSO by 1 O bases hosted in France by L 'Européenne de données, TEXTO by 5 bases in France. In this same country, 12 bases produced by l'lnstitut de l'Information scientifique et technique use the same formal, called 6-74, which has not been exported.
Due to the diversity of the data concerned, it is difficult to present general comments on the documentary processing they receivc. This difficulty had also to be met by the compilers of the Directory who could not offer cornpletely homogeneous information on the topic (maybe becausc therc is no general agreement on the use of technical terms as «keyword», «index», «classification» ... ) Very few of the bases are satisfied with only cataloguing and classifying the data. As a matter of fact they are most oftcn indexed in one way or another; more often than not, indexing relies on a controlled vocabulary, which is frequently a thesaurus. A majority of the bases add summarics or ahstracts to the keywords or descriptors, but the Directory does not inform on the types of summaries.
Access to the data is madc available either directly by the producer or by a computing center which hosts the base and sells the data. A limited number of bases use both methods. Sorne are hosted by two or even three differcnt vendors.
For this reason, thc sum of calculated percentages may exceed I OO.
There are significant differences hetween countries under this respect. 1 n Belgium, Portugal and Spain, most producers directly give access to thcir data. In Belgium there is only one vendor (exccption bcing rnade for two hosts of the European Communities, ECHO and EUROBASES): BELINDIS, set up by the Ministry of Economic Affairs, which hosts onc hase produced by the sarne Ministry plus five from other origins. In Portugal, thcrc are 10 cornputing centers for 14 bases; in Spain, 13 for 24 bases (7 of thcm being distributed by the same Center for data proccssing of the Ministry of Education and Science).
In ltaly, thcre are only 6 vcndors, 5 of thcm bcing at thc same time the most important producers: only one hase is hosted by an independcnt vendor. This explains, in that country, an unusual uniformity of the means of access to the data (sce below). Rev  In France, 23 % of the bases directly give access to their data, orare hosted by a specialized computing center. But most frequently use is made of the services of a commercial vendor. 52 % of all the bases are hosted by L'Européenne de données; 26% by Télésystemes-Questel; 11 % by SUNIST (a vendor set up by the Ministry of Education). In addition, two bases aiming at an international market are also hosted by ESA-IRS, and one by DIALOG. Here ·also we may expect sorne degree of uniformity in the means of access.
The softwares used to retrieve data from the bases depend upon the abovementioned c<;mditions. 8 % of the bases, in Belgium as in France, rely on hometailored softwares; the figure is 17 % in ltaly as in Spain; it is so high as 50 % in Portugal. Sorne bases do not clarify what software they use. In ali other cases retrieval is made by a standard software. But only one, STAIRS, is used in ali Videotex is still very unfrequent: three bases use it in Belgium, one in Portugal, one in Spain. Spain alone resorts to CD-ROM to disseminate data: 29 % from its bases do so. Additionaly, two producers in Spain and one in Belgium sell magnetic tapes.
On the contrary, the great majority of the bases in France (92 %) resort to videotex (Minitel), jointly with one or severa} other means. The data network TRANSPAC gives access to 91 % of the bases, and the telex to 54 %. In addition, the public telephone network is used by 5 %, and rented lines by another 5 %. 15 % also sell magnetic tapes; 5 % only (i.e. 3 producers) publish CD-ROMs.
Most of the producers and/or vendors provide assistance to users and train them to an appropriate use of their products; there is no significant difference in this respect among the five countries. More than half of them issue for this purpose handbooks or guides. Surprisingly enough, few among them pul forward the availability of the thesauri or lexicons they use for índexing the data: 22 % from the French bases, but only one Belgian, 2 Italian, 2 Portuguese and 2 Spanish.
There are also significant differences in the charges paid by the users. In the four countries where online access is as yet less developed, a number of bases are