Machine translatability of Russian scientific texts

The problem of the identification of text parameters having negative impact on the quality of machine translation is considered in the article. Definition and classification of translatability indicators is given. Linguistic phenomena in the scientific texts in Russian are considered at the graphic, lexical and syntagmatic levels. The special attention is paid to textual features causing problems for machine translation at the syntagmatic level, including both universal features and ones specific to Russian. A classification of formal machine translatability indicators for Russian scientific texts is compiled on the basis of language features analysis. The distinguished classes are matched against text translation problems associated with the corresponding indicator. The obtained results can be applied in translation practice and when developing tools for computational support of translation.