During the workshop on Thermal Aspects of Radio Frequency Exposure on 11--12 January 2010 in Gaithersburg, Maryland, USA, the question was raised whether there would be a practical advantage in shifting from expressing the exposure limits in SAR to expressing them in terms of a maximum allowable temperature increase. This would mean defining adverse time--temperature thresholds. In this paper, the HCN discusses the need for this, considering six points: consistency, applicability, quantification, causality, comprehensibility and acceptability. The HCN concludes that it seems unlikely that a change of dosimetric quantity will help us forward in the discussion on the scientific controversies regarding the existence or non-existence of non-thermal effects in humans following long duration, low intensity exposure to electromagnetic fields. Therefore, the HCN favours maintaining the current approach of basic restrictions and reference levels being expressed as SAR and in V/m or mu A mu T, respectively.