The main problems of radiotherapy planning are discussed with respect to use of computers now being available for a greater number of centers. One of the most essential premises is apart from clear ideas on a modern target volume concept - a sufficiently high speed in producing realistic summarized isodose contours for any radiation therapy arrangement in any individual patient's cross section outline. This problem being solved, those individual summarized isodose figures have to be critically evaluated and therefore the importance of applicable meaningful optimization criteria come into account. The properties of such criteria, which must be quantifiable, generally applicable and really relevant for judgement on quality of a plan, had to be accepted, before automatic optimization procedures could be developed; principles involved are presented. By means of a short series of examples, namely 60Co fixed field combinations with and without use of wedges, combinations of arc therapy for a number of clinical tasks. It has been pointed out, that most experienced estimate by the eye would have been by far insufficient when compared to the automized computer optimization when using such simple criteria as 1. homogeneity of absorbed dose within target volume, 2. numerically limited absorbed dose within areas of risk, 3. as low as possible radiation effects to all "outside areas". It seems to be a real danger, that so called isodose libraries, how high their merits might be estimated, may change into an "isodose bcemetery", unless we'll be successful for each individual clinical case by use of computers, which are now prepared to supply the best possible variant of the standard plan or the primary radiotherapy idea. Regular use of computers in such a way will furthermore give an incomparable documentation material.