While the consideration of aesthetic
needs is delicate, the consideration of cognitive needs relative to the
transcendent personal device is relatively straightforward. The primary
concerns are parallel computing capability, language capability and memory. In
considering the derivation of human needs in Chapter 5, we suggested that a
precursor to the development of cognitive needs was the establishment of
context. This facility within the human brain would seem to occur as a
characteristic of the multiple processing pathways enabled by neural networks
connecting sensory input with motor control output. Sensory input such as
signals from the myriad of sensors in the eyes feed into multiple, parallel
neural network structures. Similar signal processing activities are performed
on the raw input and a refined image is transferred downstream within each
network. Ultimately, it would appear that these images arrive at a variety of
neuron constructs where some particular contextual situation is recognized as
preeminent among all the various choices. Various, perhaps subtle, indicators
are identified through cognitive pre-processors so as to establish that an
image of another person drawing back an arm is, for example, a pitcher in a
baseball game rather than an attacker with a knife. Once constrained within the
appropriate context, cognitive functions can metaphorically interpret the
sensory input and then more efficiently guide appropriate motor responses. In
order for the transcendent personal device to facilitate support for the
cognitive needs of the human mind of its bearer, it would be most useful if it
could better establish appropriate context for the situations that it becomes
aware of through its sensory input from the outside world.
The human mind is capable of multiple,
simultaneous cognitive activities. Some are reflexive, some are the products of
unconscious or subconscious processing, and some are functions of conscious
activity. A single threaded computer, if clocked at a high enough rate and with
significant interrupt facilities can approximate such multiple streams of
processing. However, true replication of this facility will be greatly enhanced
by actual multi-processor architectures.
We suggest that multi-processor configurations can be effectively utilized both
within the device body as well as in its trusted core agent.
In concert with memory enhancements, the
evolved trusted core agents can certainly use the benefit of these multi-processor
configurations. Current secure cores that provide cryptographic services
already include a peripheral processor in addition to the main processor. This
peripheral processor is capable of performing long integer multiply operations
at a very high speed, a computation that is essential for public key
cryptography. A true multi-processor facility would allow the token to further
enhance its security characteristics as well as be better able to involve
itself in multiple, simultaneous transactions. As we noted in the first
chapter, the ability to deal with the programming and provisioning of parallel processing
threads is something of an art form, related in style to the composition of
music in chords. As a consequence, a derivative requirement from cognitive
needs is that of programming language support. Moreover, language capabilities
to support logical reasoning are mandated by the self-actualization need
related to social system ontologies.
Memory enhancements derive from the
discussions in the last chapter in which context was presented as an integral
aspect of memory. In this case, we suggest that the device memory should
encompass knowledge base facilities that allow a tight integration between
information and its relevant processing.
|