Abstract
Art repeatedly becomes the challenge for artificial intelligence. Its embedding into a computer programme gets its sense as well as in other AI solution partly in searching new autonomous art production, partly in uncovering general principals of creating, or methods of existing authors' styles.
Non-representational art from the beginning of 20th century came with convenient reduction, which seems to be in the computerized ways, in the meaning of symbolic attitudes, manageable and principally achievable by unifying the functions creating visual structures. Successfulness of programming non-representational art works from the beginning of the last century can be tested by the alternative of Turing test which would compare adequate visual outputs.