Representatives of various areas (research, authorities, and users) in Hungary met on 26 October 2004 to consider the draft of the new ICRP recommendations. Consensus was reached on the following: In general, the changes introduced by the draft are positive. Suggestions for modification and/or further clarification are given below. (1) It is of great theoretical importance that the uncertainty of the risk estimates are emphasised throughout the paper. For example: para. (120): “… current uncertainties … are too great for the development of practical judgements on low dose cancer risk.” para. (112): “… the information available is insufficient to provide a meaningful quantitative judgement … ” para. (A18): “The estimated risk of radiation-related cancer is uncertain and the sources of this uncertainty are many.” Taking the above into consideration, it is clearly stated that the increase of risk with increasing low doses is just a hypothesis introduced as a means of practical radiation protection: para. (38): “For protection purposes it is assumed …” As a further consequence, the risk coefficients given are termed nominal: para. (113): “… nominal risk coefficient …”. Suggestion: In Table 6 (para. (113)): A footnote should be given stating that these coefficients are to be used only below the thresholds of the deterministic effects (i.e. about 200 mSv) and that at low doses (i.e. below about 10 mSv) the values are extremely uncertain. Taking into account all the above, scientifically correct, arguments it is clear that the whole system of radiation protection cannot be based on a very uncertain quantity: the risk. Comparison of the new limits with the natural background, with a quantity that can be measured much more accurately, is welcomed. A system based on risk comparisons cannot be rejected in principle; in the distant future, when much more accurate risk estimates (not only for the effects of radiation but also for other possible quantities for comparison, e.g. chemical carcinogens) become available, we can return to a risk-based system. Suggestion: A broader discussion on the fluctuation of natural background should be given. (2) The argument to change the term “deterministic effect” to “tissue effect” is not convincing. Stochastic effects also result in effects on tissue. Suggestion: Keep the term “deterministic effect”. (3) Tissue weighting factors, due to the risk uncertainties outlined above, are uncertain. It is questionable if their tuning is worth the re-computing of dozens of tables. Nevertheless, the additivity of the effective dose is a clear advantage. (4) The broader understanding of optimisation, the change for qualitative analysis, and the involvement of stakeholders are welcomed. Suggestion: Further guidance is needed for the practical aspects of stakeholder involvement. (5) It is correctly reflected that in the process of justification radiation is only one of the points: many other social, economic, etc. factors have a role. Suggestion: Since many readers have the feeling that the role of justification has drastically been decreased in the draft, some detailed explanation may be needed describing that justification as a step is still very important, but radiation protection shares the position with many other aspects. (6) There are many good arguments for decreasing the role of the collective dose. It masked the right of the individual for equal protection and it was frequently misused in calculations based on multiplication of collective doses with strongly believed risk coefficients. The dose distribution or matrix representation suggested instead - whose practical implementation is not a real problem with the present computational tools - enables a much finer evaluation. (7) The increased role of the source-related dose constraint is understandable. In terms of checking the population’s radiation burden, it was impossible to determine those responsible for exceeding the dose limit; in the new system, the responsibility is clearly imposed on the individual licensee. Suggestion: In Table S1, our view is that the value of 0.01 mSv as the “minimum value of any constraint” is unrealistically low, 0.1 mSv is suggested instead. Suggestion: There should be an explicit indication that the “minimum value of any constraint” is also the lower limit of any optimisation. (8) The new unit Gy-eq. introduced for the RBE weighted dose is welcomed. (9) Bearing in mind that particles interacting with organisms cannot be distinguished by their origin, we cannot see the rationale for the distinction (by a factor of 100) between exclusion levels of natural and artificial sources. (10) We agree with the general aim of simplification of the system, especially since non-professional stakeholders are to be extensively involved in the process. * - * In summary: The draft makes the new system of radiation protection more transparent and clearer and results in relatively few changes in the practice. Its introduction would lead to a better radiation protection system.