DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS OF THE INVENTION

[0057] Subjects wear devices equipped with computing, communication, sensors and actuators. A typical device senses the proximity of others by means of its communication system, and exchanges simple “applets” and data objects with them. An operating system in the devices executes certain applets according to a rule that gives greater computing resources to applets whose decisions have been found successful in the past, including giving successful applets access to actuators so that they can attempt to modify the behaviour of their subjects. A device would generally have an initial collection of applets known to be broadly successful, but new applets could be introduced to the community at any time and would have the power to reproduce if successful.

[0058] The success of the system depends on ensuring that applets with desirable behaviour survive and propagate. Their propagation from one device to another is essentially viral, with the individual devices modelling the state of mind of their bearers and favouring the infections that can most cheaply predict their desires; and by allowing the viruses to touch their bodies through a variety of actuators—sounds, smells, electrical shock, micro-injections—devices arrange that these viruses of desire will succeed best if they can manipulate estimated desire to make their predictions more accurate.

[0059] Two particular types of data object are envisioned as useful in an initial embodiment: a table of (name, value) pairs and a labelled graph of interactions. It is envisioned that data objects generally permit any applet to read them and to add new entries, but that overwriting of existing entries is controlled through a cryptographic signature mechanism. Objects should in general also implement specialized garbage collection methods that allow the operating system to save storage by deleting data saved by unsuccessful applets.

[0060] An interface for the (name, value) pairs, for example, can have a single instance of a Table, with methods including:

[0061] Table.Set(Name, Value, Signature)

[0062] Value=Table.Get(Name)

[0063] size=Table.SpaceOf(AppletID)

[0064] void Table.Delete(Name)

[0065] An applet “linger” can use this mechanism to estimate the attractiveness of the owner of a device, for example, by estimating how long other owners usually stay nearby (polling the communications interface to determine this at the current time, and averaging the result with an earlier estimates by some formula such as

[0066] “LingerTime=0.99*LingerTime+0.01*LatestTime”), then setting

[0067] Table.Set(“Linger”, LingerTime, LingerSignature)

[0068] This applet could then compare the LingerTime values for two subjects as they approached each other, perhaps before they have had an opportunity to see each other, and make recommendations using an earpiece transducer; for example telling a moderately attractive male to suck in his gut on the approach of an attractive female.

[0069] This applet uses a small amount of memory resource (the LingerTimetable entry), which is therefore not likely to be tidied away, and occasionally requests the use of the relatively expensive earpiece resource. If its advice is found to be good, then the operating system will allocate it a higher priority, and its access to the earpiece will become more probable.

[0070] Note that the subject does not explicitly set LingerTime, and therefore cannot overtly claim to be more attractive (or “lingery”) than he or she actually is. In general the author of an applet will not explain or publicize its algorithms lest users then modify their behaviour to mislead the algorithm: for example sleeping with a collection of devices so as to appear to cause lingering.

[0071] The operating system uses a nonlinear optimization algorithm and a defined criterion to select how many resources to give each applet. An example of such a criterion is that an applet that correctly predicts that a particular subject will spend the night with the owner of the given device would score highly, and in general accurate prediction of duration of contacts would score positively. An example of a suitable optimization algorithm is one in which resources are initially allocated equally, and then the resources available to each applet are increased by a small percentage for each applet scoring a more accurate estimate than the median and decreased slightly for each applet whose estimates are less accurate than the median.

[0072] According to FIG. 1 a community 22 of subjects 23 , such as concertgoers, wearing instances of device 31 , such as will be described below, are led to interact with others likely to be objects of desire by the actions on each said subject 23 of actuators 63 , such as those described below, in each device 31 wherein the plural device 31 communicate by wireless links 41 , such as by BlueTooth, IEEE802.11b or infrared links or by wired means such as 10-base-T Ethernet.

[0073] In the embodiment of FIG. 1 subject 23 a Eros caresses subject 23 b Aphrodite including direct mechanical activation of a simple pressure switch in nipple device 36 , which communicates over wireless link 41 a to embedded device 38 in subject 23 c Jealousy. The effect is amplified by a further interaction in which wireless link 41 b from device 31 a held by subject 23 d Folly to peach device 31 b held by said subject 23 b Aphrodite causes it to secrete a sweet-smelling liquid, distracting said subject 23 a Eros from the operation of arrow device 31 c held by said subject 23 b Aphrodite which can be violently but partially embedded in the body of either said subject 23 a Eros or said subject 23 c Jealousy in order that its actuator 63 a may inject an aphrodisiac such as Viagra, hallucinogen such as lysergic acid, muscle relaxant such as curare or other chemical or plural chemicals such as Rohypnol known in the art of seduction. The choice of chemical may be modified by interactions with external parties, such as subject 23 e Deceit who carries a handheld device 31 d connected by wired link 42 to headband device 31 e which is more conveniently placed to make wireless link 41 c operable to said arrow device 31 c . By means of sensor 62 a operable to detect pheremones said handheld device 31 d selects which type or combination of chemicals can best be injected to make the behaviour of injected subject 23 a Eros or injected subject 23 c Jealousy more desirable to both subject 23 e Deceit and subject 23 b Aphrodite. Other nearby subjects, such as subject 23 f Time may have only indirect interactions with the principal emotional drama engulfing subjects 23 a through e.

[0074] According to FIG. 2 each subject 23 may wear plural device 31 , all communicating by wireless links 41 or by means of wired link 42 . These plural device 31 combine their actions in such a way as to act towards outsiders as a single more capable distributed device 40 having plural sensors 62 and actuators 63 .

[0075] In the embodiment of FIG. 2 headband device 31 e contains sensor 62 b operable to detect local electromagnetic radiation and sensor 62 c operable to measure electroencepholographic potentials together with actuator 63 b operable to cause a tingling or tightening sensation in the scalp, such as by myoelectrically activating scalp muscles. It is connected to ferret device 31 g having whisker sensors 62 d operable to detect air currents from movement of nearby bodies together with claw actuators 63 b operable to inflict pain or alternately or in combination claw actuators 63 c operable to inject venom, aphrodisiacs, pheromones, muscle relaxants, Mickey Finns, sodium pentathol or other chemicals known in the art of seduction. Ferret device 31 g may also have tooth actuators 63 d operable for the same purposes as claw actuators 63 b or actuators 63 c , or because of the possibility of scissor action also capable of severing small body parts in a subject 23 . Necklace device 31 h connected to headband device 31 e may simply offer more computing resources, such as memory or battery power.

[0076] According to FIG. 3 each device 31 contains a CPU 51 such as a StrongArm processor maufactured by Intel, a radio module 52 such as a Bluetooth module manufactured by Conexant, antenna 53 such as a whip, an address and data bus 55 such as that standard with said StrongArm, a flash memory 56 such as a 28F640W18 chip manufactured by Intel, a random access memory 57 such as the IBMN612404GT3B manufactured by IBM, a battery 58 such as a wristwatch battery. It may also contain one or plural actuator 63 or sensor 62 , each connected to the address and data bus 55 by means of a signal conditioning circuit 61 .

[0077] According to FIG. 4 plural device 31 may be connected by means of wireless links 41 or wired links 42 to form a more capable distributed device 40 .

[0078] According to FIG. 5 a device 31 or plural devices 31 may be programmed by a board 71 of judges 72 , such as retired medical professors, who evaluate the subject 23 intended to bear the said devices 31 , in terms of characteristics deemed to be likely to substantially affect his or her attractiveness to various groups, by means of programming devices 73 such as laptop computers. The programming devices 73 and love device 31 may all communicate by means of wireless links 41 . It will be apparent to one versed in the art that this communication may also take place by wired means.

[0079] According to FIG. 6 the software running in a device 31 may contain an applet management system 85 which includes an applet database 90 , a score keeper 94 that tracks the accuracy of predictions registered by applets in a prediction registry 95 and gives applets access to a collection of standard objects including labelled graph and table 91 . The applet management system 85 is operable to exchange applets with other instances of a device 31 and to give applets greater or lesser proportions of resources offered by resource allocation manager 88 as a function of their cost/performance ratio as calculated from their scores and their resource usage; this algorithm being chosen to maximize likelihood of correct predictions from the subject desire and pleasure modellers 89 with minimum use of resources.

[0080] The software running in device 31 may also contain a resource allocation manager 88 which includes software operable as a usage monitor and cop 96 that estimates and controls usage of such resources as CPU 51 , random access memory 57 and sensors 62 with algorithms well known in the operating system art. This resource allocation manager 88 may for example operate by having targets for average utilization of each resource and offering increases or demanding reductions in such use from the applet management system 85 according to a message interface of a type well known in the art of object-oriented programming systems.

[0081] The software running in device 31 may also contain a subject desire and pleasure modeller 89 which uses estimators of standard metrics 92 , which process raw data from sensor drivers 93 to measure such things as heart rate and pupil dilation.

[0082] It will be apparent to one skilled in art that there are many metrics either for pleasure or desire. At a commercial level it may be expected that devices 31 implementing metrics that correspond to popular concepts of pleasure will be preferentially selected by the market, according to the dictum “I may not not much about love, but I know what I want”. In this case, however, the covert nature of the algorithms means that subjects 23 will not be aware of how they are being manipulated; this would, for example, allow for commercial success of a device 31 so designed as to make mood swings extreme while taking credit for the positive swings and hiding its participation in the negative swings. From an abstract point of view this is simply a definition of the criterion “pleasure” and for the purpose of this disclosure the term shall be taken to comprise all manipulations of mood that tend to enhance dependence on devices 31 .

[0083] According to FIG. 7 the random access memory 57 of each of the devices 31 described above contains a resource allocation manager 88 operable to increase or decrease the resources available to plural Java applets 84 depending on the success of their predictions. It is composed of specialist resource allocation managers 88 a through c each operable to drive usage of a particular resource towards a target by passing requests on to the applet manager described below.

[0084] FIG. 8 shows an applet management system 85 operable to distribute resources among plural Java applets 84 according to their consumption and success at prediction. It contains specialized component applet management systems 85 a through c with algorithms that rank applets and distribute resources or cuts in resources as required by the resource allocation manager 88 . It would implement other methods, including ones that download applets from new acquaintances so that successful ones can propagate. It would tend to prefer to download applets with high scores and low resource usage.

[0085] It will be apparent to one skilled in the art that many algorithms, broadly called evolutionary algorithms, exist that have the general properties required for an applet management system 85 , and that these will have different success in propagating desired viruses. At a commercial level, devices 31 with good behaviour will be selected by the market.

[0086] According to FIG. 9 an exemplary Java applet 84 “linger” would contain methods responsive to offers of CPU and earpiece resources from applet management system 85 . The CPU offer would be exploited by checking for neighbouring devices with “lingerTime” greater than that of the current device. The earpiece offer would be exploited by playing a prepared message such as “sweet!” if a recent high-lin-gerTime neighbour had been found, and in this same case would predict a momentary surge of desire within the next ten seconds. If this algorithm thus succeeds in alerting the subject 23 to the presence of an attractive person within radio range, then for example a male subject 23 may suck in his gut and look around, perhaps seeing an attractive woman and fulfilling the prediction of desire. This in turn will improve the score of this applet, and hence its probability of survival.

[0087] It will be apparent to one skilled in the art that many similar algorithms can be developed, each having several or all of the characteristics that it attempts to predict desire, that it may use access to actuators 63 to improve the likelihood of its prediction being accurate, that it uses sensor data, and that it communicates with applets or standard objects in nearby devices 31 to improve its accuracy. These applets will be written to be parsimonious in their usage of resources in order to enhance their chances of survival in an individual and propagation to contacts, thus using viral propagation to the advantage of the system. These applets may be developed by anyone competent to develop Java code, thus enabling the development of a large supply of viruses.

[0088] According to FIG. 10 an earring device 32 , which is a particular type of device 31 , may be constructed as shown in FIG. 3 sup wherein the sensor 62 is operable to detect heartrate and the actuator 63 is operable to make a variety of sounds such as whispered messages or songs. Also according to FIG. 10 a nose stud device 35 , which is a particular type of device 31 , may be constructed as shown in FIG. 3 sup wherein the sensor 62 is operable to detect pheromones and carbon dioxide and the actuator 63 is operable to dispense a variety of pheromones and scents

[0089] According to FIG. 11 a breast device 34 , which is a particular type of device 31 , may be constructed as shown in FIG. 3 sup wherein the sensor 62 is operable to detect heartrate and breathing rate and depth, and the actuator 63 is operable to constrict the chest in patterns that simulate rapid breathing or that temporarily restrict breathing altogether.

[0090] According to FIG. 12 a nipple device 36 , which is a particular type of device 31 , may be constructed as shown in FIG. 3 sup wherein the sensor 62 is operable to detect changes in state of erectile tissue and the actuator 63 is operable to apply mild electrical shocks.

[0091] It will be apparent to one skilled in the art that a similar device could be developed to detect tumescence in the penis and to stimulate it with shocks, warmth, vibration or topical injections of nitric oxide.

[0092] According to FIG. 13 an embedded device 38 , which is a particular type of device 31 , may be constructed as shown in FIG. 3 sup wherein the sensor 62 a is operable to detect activity of a single neuron and the actuator 63 a is operable to stimulate a single neuron. Alternative interfaces for an object of this type could include a sensor 62 b operable to detect sounds characteristic of opera or Gregorian chant and an actuator 63 b operable to dispense psychoactive chemicals or hormones.

[0093] A second important data object is the labelled graph, which is capable of representing complex behaviours without direct reference to their semantics, such as sexual preference. In the preferred embodiment a graph that represents a subset of the transitive closure of those subjects with which the given subject comes into contact is labelled with (name, value) pairs, such as the LingerTime data described above. The data now refers to Lingering with specific individuals, rather than with people in general, so that more subtle types of attraction may be represented.

[0094] According to FIG. 14 a portion of labelled graph 83 may represent the sexual activities of the members represented by nodes 123 a through j of a suburban swingers club, where only edges labelled with data suggestive of sexual activity are shown. Note that there is extensive contact, but that the nodes (representing subjects) may be divided into two distinct groups with contact occurring from one group to another but not within the groups. This bipartite graph is characteristic of promiscuous heterosexuality. An applet able to recognize this pattern would be able to introduce members of the Dallas and Orange County swingers clubs to each other even with no common contacts, because on the first encounter between members of the two groups (at a Jester's conference, for example) the applet would recognize that both subjects shared this activity.

[0095] This graph is semantically ambiguous, in that it could equally represent the violent physical contacts among members of two five-a-side basketball teams. Note that it would still function to introduce basketball players to each other. In the case where the sensor device or recording applet is unable to distinguish between sexual and game contacts it would also tend to introduce basketball players to suburban swingers: depending on the reactions of the parties so introduced the given applet would be more or less apt to survive. This is an example of applet behaviour that is likely not to have been expected by the original applet author, and hence of the importance of the management structure herein disclosed.

[0096] According to FIG. 15 a portion of labelled graph 83 may represent the sexual activities of the regulars represented by 123 k through n at a particular public washroom. Note the presence of a highly connected subgraph, which could suggest homosexual or pansexual behaviours within the group. Note also that the individual represented by 123 m does not participate. Again, an applet able to observe the existence of this pattern in the database of a newly acquired neighbour would be able to suggest a potential relationship; and if that relationship were successful it would then be able to strongly recommend members of the new partner's group.

[0097] Again the semantics of the graph of FIG. 15 are ambiguous. If the sensor or applet reporting this structure is unable to distinguish between sexual activity and the exertion and proximity at a gymnasium or religious revival meeting, it will introduce workout enthusiasts to each other and also to promiscuous homosexuals and fundamentalist Christians.

[0098] According to FIG. 16 a portion of labelled graph 83 may represent the activities of a small congregation 123 p through u confessing privately to a minister 123 v.

[0099] Equivalently the graph could represent the sexual activity of a prostitute represented by 123 v : again a large number of nodes all connect to a single “central” one but not directly to one another. This would be ambiguous if the only sensor were “proximity” as judged by the presence and duration of radio contact between the devices of the various subjects. A more discriminating sensor would be required make it possible to distinguish the shepherd from the black sheep.

[0100] Various alternatives and enhancements to the embodiments described above will be apparent to those skilled in the art and are not intended to be excluded from the scope of the present invention which is defined solely by the claims appended hereto.