The paper is an analytical overview of the main results presented at the 3rd Dosimetry Workshop in Hiroshima(9-11 of March 2005), where different aspects of the dose reconstruction around the Semipalatinsk nuclear test site(SNTS) were discussed and summarized. The results of the international intercomparison of the retrospective luminescence dosimetry(RLD) method for Dolon' village(Kazakhstan) were presented at the Workshop and good concurrence between dose estimations by different laboratories from 6 countries (Japan, Russia, USA, Germany, Finland and UK) was pointed out. The accumulated dose values in brick for a common depth of 10mm depth obtained independently by all participating laboratories were in good agreement for all four brick samples from Dolon' village, Kazakhstan, with the average value of the local gamma dose due to fallout (near the sampling locations) being about 220 mGy(background dose has been subtracted).Furthermore, using a conversion factor of about 2 to obtain the free-in-air dose, a value of local dose approximately 440 mGy is obtained, which supports the results of external dose calculations for Dolon': recently published soil contamination data, archive information and new models were used for refining dose calculations and the external dose in air for Dolon village was estimated to be about 500 mGy. The results of electron spin resonance(ESR) dosimetry with tooth enamel have demonstrated the notable progress in application of ESR dosimetry to the problems of dose reconstruction around the Semipalatinsk nuclear test site. At the present moment, dose estimates by the ESR method have become more consistent with calculated values and with retrospective luminescence dosimetry data, but differences between ESR dose estimates and RLD/calculation data were noted. For example mean ESR dose for eligible tooth samples from Dolon' village was estimated to be about 140 mGy(above background dose), which is less than dose values obtained by RLD and calculations. A possible explanation of the differences between ESR and RLD/calculations doses is the following: for interpretation of ESR data the "shielding and behaviour" factors for investigated persons should be taken into account. The "upper level" of the combination of "shielding and behaviour" factors of dose reduction for inhabitants of Dolon' village of about 0.28 was obtained by comparing the individual ESR tooth enamel dose estimates with the calculated mean dose for this settlement. The biological dosimetry data related to the settlements near SNTS were presented at the Workshop. A higher incidence of unstable chromosome aberrations, micronucleus in lymphocytes, nuclear abnormalities of thyroid follicular cells, T-cell receptor mutations in peripheral blood were found for exposed areas (Dolon', Sarjal) in comparison with unexposed ones(Kokpekty). The significant greater frequency of stable translocations (results of analyses of chromosome aberrations in lymphocytes by the FISH technique) was demonstrated for Dolon' village in comparison with Chekoman(unexposed village). The elevated level of stable translocations in Dolon' corresponds to a dose of about 180 mSv, which is close to the results of ESR dosimetry for this village. The importance of investigating specific morphological types of thyroid nodules for thyroid dosimetry studies was pointed out. In general the 3rd Dosimetry Workshop has demonstrated remarkable progress in developing an international level of common approaches for retrospective dose estimations around the SNTS and in understanding the tasks for the future joint work in this direction. In the framework of a special session the problems of developing a database and registry in order to support epidemiological studies around SNTS were discussed. The results of investigation of psychological consequences of nuclear tests, which are expressed in the form of verbal behaviour, were presented at this session as well.