Abstract: A method of automatically selecting an item (e.g. s3) from a set (S) of items involves using item properties to determine weighing factors. The weighing factors are then used to produce a preference distribution having sections, the relative sizes of which correspond with the respective weighing factor. A random number may be used to point to a section of the preference distribution and thus select the associated item. The method may be used to select multimedia items, such as songs, stored in a mass storage device.

Abstract: The present invention relates to Assistant Software Modules (ASMs) providing services to users of e.g. a mobile phone. The ASMs are becoming learned (or skilled) by the user modifying the behaviour of the ASMs. Adding rules to a configuration file of the ASM preferably performs the modifications, but the modification may also be implemented as a neural network being trained. The invention relates in particular to management of the ASM, which management comprising uploading an ASM from a user who has modified the functionality of the ASM and downloading the ASM to a new user The uploading, downloading and terms for acquiring an ASM is regulated by rules prescribed by a manager of the ASMs.

Abstract: The multimedia purchasing apparatus (100) for purchasing multimedia content (161) over a network (140), interacting with a user profiling unit (122) containing information on user preferences of multimedia, has a user negotiation unit (120) comprised, which is arranged to autonomously purchase the multimedia content (161) guided by a budget specification (114). Different negotiation strategies may be used to purchase content as desired by the user.