Over the past few months, work has been accelerating on the ELSNET survey of Language Engineering Laboratories in Eastern Europe--or more precisely, in Central and Eastern Europe and selected Newly Independent States of the Former Soviet Union. Edinburgh sent out over 300 questionnaires by email, fax, ordinary mail and in some cases, by hand courier. The flood of responses has now diminished to a trickle and so far we've clocked up almost 100 returns. This is an excellent response rate and testifies eloquently to the enthusiasm and commitment of the professionals in the countries being surveyed.

The data in the questionnaire will be used in two ways. First, as input to a report that we will submit to the EC about future possible actions involving Eastern Europe (or more precisely, Central and ...). Second, the data will be massaged into a volume of site profiles. analogous to those already produced by ELSNET for its nodes. As far as possible, the massage process is being automated: the electronic versions of the completed questionnaires are are converted into LaTeX format and also into HTML format since we plan to make the site profiles available over WWW as well. A first release of the latter is planned for late June.

Although we still have quite a lot of work left to do in analyzing the data, various patterns are beginning to emerge. The bar graph on page 20, for example, gives a rough profile of the distribution of laboratories across the different countries in the survey. The figures have to be taken with a large grain of salt because of communication difficulties; in some countries, we managed to reach a representative sample only by using hand-delivery of questionnaires. In other cases we were less successful. Overall, we have found the ration of speech groups to NL ones to be about 3:5, though this varies of course from country to country. (Most of the sites classified as "Other" work in both areas.)

Not surprisingly, virtually all groups cite problems whose solution requires more funding. The absence of economic and political stability is perceived as a major factor in the raising funds for research. Salaries, for example, are very low, and bright computer science graduates can usually earn three times more in industry than in academia. Unfortunately, many groups are unable to benefit significantly from EC project funding, which is awarded on a marginal cost basis. This is because the staff that they want to retain are typically employed on long-term teaching contracts and their income cannot be met or supplemented from external research funds.

Although we hope that EC funding in the Fourth Framework will be able to ameliorate the severe infrastructure problems that afflict many labs in Eastern Europe, there is also much that can be done at the grassroots level by individual institutions in the West, especially by encouraging more scientific exchanges of software, ideas and people.

Reprinted with permission from:
European Network in Language and Speech
email: elsnet@ed.ac.uk
mail: Centre for Cognitive Science, Univ. of Edinburgh, 2 Buccleuch Place,
Edinburgh EH8 9LW, Scotland
Tel: +44 31 650 4594
fax: +44 31 650 4587
WWW: http://www.cogsci.ed.ac.uk/elsnet/home.html