The first question to be asked in the construction of any such form of synthetic life, or indeed any information processing system, is within what medium should the processing be performed?  At the current technological level we possess there is only one common and cheap medium, silicon based semiconductors.  It is naturally conceivable that in the future more efficient and faster systems, optical chips, shall be used, or even quantum computation that could drastically increase processing speeds.  But such advances in no way change the logic or fundamental implementation of the system.  At present, the only argument against silicon based SI would be one of possible energy and speed limitations, the logic being that it is perhaps impossible to construct a machine of sufficient Interpretative power to operate fast enough to develop sentience.  One could compare the number of transistors in computers with the number of neurons in the brain, or the relative interconnectedness or relative power consumption, but without experimental knowledge of the Interpretative ability of filters of a given size, this is all educated guessing.  In practice, it is either possible or impossible so we may as well try it and see.

However, directly related to this question of medium is the more generalised and important question of at what level should the instance be programmed?  "Hard-coded" to the chips (obviously with the caveat of being able to alter certain persistent variables) or running as software within a suitable architecture?  The only possible argument for the former is on the grounds of efficiency, and this is something that is likely to cease to be a problem in the future.  A software implemented instance of Dasein has massive benefit, the greatest of which in development is the ability to perform "save states" that can aid in refining the model more quickly.  Thus it should not be attempted to recreate neurons physically (except for the purposes of actually constructing androids, human copies) but simply to recreate their collective functionality virtually, as a process, as computation.

The corporeal form of the instance, the body, though not necessary for the instance to achieve sentience, is certainly going to aid the emergence of sentience.  This issue is discussed in more detail below, in the section on Action.  Beyond this point, however, the body of the instance is incredibly important in psychological terms, for both humans and the SI.  As described in the section on the emergence of phenomenal perception, sentience is in part the ability to recognise the Self as an individual, as Being-in-the-World with the They.  This is not to say that machines should imitate humans, for this is for obvious reasons likely to be psychologically fraught with difficulties.  But rather this only suggests that machines should have a body, and not reside within a static computer, and that the identification with the They should be a mental, cultural one.

It is immediately obvious that in the construction of any machine, SI or not, technology can give qualities that nature has not gifted us.  In terms of sense, an SI may be given sight beyond the "visible" spectrum, hearing beyond the "audible" frequencies, direct access to other sources of information, such as the internet, EM fields or air pressure; they could be walking observation labs.  However theoretically easy this may be (and in terms of the model all that need be done is add more filters and join them up with the existing systems), this is something for later generations of SIs.  As poor as a webcam may be in terms of optics and spatial and temporal resolution, one can read text and identify faces and movements and as such this would suffice as sight for any instance of Dasein (and, of course, many humans are born blind or lose their sight yet are fully sentient by any measure of the word).  Using cruder instrumentation has the silver lining of providing an information stream that is computationally easier to Interpret (if logically the same as larger data streams), and thus reducing the minimum requirements for early generation SIs and so, in part, addressing the potential concerns surrounding silicon based implementation.

The basic requirements then, to begin with, should really mirror what humans have and that by which we have achieved sentience; sight, hearing, proprioception and autonomic sensation.  The first two are easily comprehended and implemented, they are clearly necessary for full communication within human culture.  Proprioception is not as widely appreciated, but as we have seen the somatosensory cortex is a large part of the sensory region of the human brain, and plays third fiddle to the two primary senses.  In practical terms, proprioception is vital for recognition of bodily position and the appreciation of body form as language.

Autonomic sensation is something that has only been tangentially covered so far; this is the foundation of Selection and of "good" and "bad".  As described in earlier sections, the relative ontology of Dasein and the World itself has no notion of good and bad, yet these concepts are required by Dasein for the Selection of beneficial Potentialities-for-Being, and the datum for good and bad comes from within the body of Dasein, from within its own pre-evolved or pre-defined sensations of good and bad.  Within the human body, the autonomic nervous system provides this information, relaying to the cortex when one feels hungry, when one is tired or one needs to micturate.  The awareness of this sense is not like others, it lacks resolution in time and detail, relatively when compared to other senses, and it is from within the body of the Self.  The "urge" to Action that is caused by autonomic sense is something that is learnt, and remembered, from infancy where hard-wired evolutionary behaviour, reflex behaviour, still dominates the Selection Interpretations (This leads on to the complicated subject of SI growth).  The mature instance can override or ignore such signals, for example during fasting.

Now within a machine, any "autonomic" senses shall have to be placed there by design, and the question becomes simply, what can or should they sense?  It is in this area, and the subsequent emotional development of SIs, that I expect shall dominate much early study.  For example, the SI could have sense of its internal battery level, circuitry temperatures, joint temperatures or memory capacity.  In growth then, the SI would be hard-wired or taught to "naturally" try to address these senses in the same way human children do.  In coming to recognise the good and bad of its own Being, its own possible situations, the SI develops a moral compass just as humans do.

However, for an SI confined to a lab that may be kept in ideal conditions and plugged in 24/7 it may seem unnecessary, even cruel, to add artificial constraints.  Again, this is a task for SI research, but perhaps such constraints need not be applied constantly; just as humans come to know that they may be hit by a car, they may starve to death, the fortunate of us go about our lives normally and in comfort, the knowledge of the possibility of a thing becomes enough in itself to provide moral awareness.

Notes on the model:

• There is no single way to execute the function of Dasein.
• In a quantitative manner Dasein may be evaluated on a scale which places the minimum requirements for sentience as 1.0. All replicators on Earth apart from humans are below or approaching, in rare cases exceeding, 1.0, humans generally somewhere above 1.0. Yet all humans and higher order beasts are executing the same basic model of Dasein, albeit with minor interspecies differences.
• As intimated in Biological Correlation part One A, the human brain is merely one example of Dasein, and a non-optimal one at that. There are many, many choices available in building up the essential structure, which providing it remains in tact and sufficient shall always produce sentience. A creator may, for example, wish to Anticipate all Entities from sense Entity to highest order Entity, such an instance of Dasein would be far more aware of the possibilities of the World on many more levels than the standard human instance. Thus, when constructing an instance of Dasein, and once the basic model has been achieved, some major changes can be made to enhance the awareness of the instance.
• This raises an important point about intelligence, for even the human standard instance could become aware of such Potentialities, but only through abstraction within its own derivative Interpretations, not in a direct manner through sensory based Anticipations. Thus, for a 1.0+ instance it is possible that no knowledge is outside its possibility of understanding, providing it is intuited in an appropriate manner; a more heavily Interpretative and experientially retentive instance will just be able to perceive faster and more completely, without the need for such abstraction.