This paper consists of abstract:Usize descriptions of a number of current Soviet research projects connected with machine translation. Three systems are described: operational (the AMPAR English:URussian system, and the NERPA German:URussian system, or experimental (the FRAP French:URussian system). All three were developed at the USSR Center for Translation of Scientific and Technical Documentation. From the very general descriptions one can conclude that these systems feature a variant of the transfer approach to machine translation and rely on the services of posteditors to produce a final result. Special attention is paid to the modularity of the system and the development of tools that enable linguists to write and modify gammars and dictionaries without the necessity of becoming actual programmers. Many of the machine translation projects in the West are concerned with similar problems. Also mentioned are the Englis:URussian first:Ugeneration (word:Ufor:Uword) translation project at the Chimkent Pedagogical Institute and a microcomputer-translator project at the Minsk Institute of Foreigh Languages.:P Among the theoretical matters mentioned are the study of sublanguages and the development of a formalized model of a human translator for eventual use in a machine translation system. Both are topics on which research is being also pursued in the West. However, the research on automating simultaneous interpretation is not.:P In general, this paper can probably be used only as a news article, since only one out of its 22 references (and none on the actual systems) is in English.:L