Chapter 4

DIGITAL SYSTEM ARCHITECTURE

Probably the likeness between the digital paradigm and pharmacology, and the similarity
between the analog paradigm and herbalism appear rather reasonable after the first pages.
However I have not examined as yet what happens in the working environment; I have not
observed carefully what digital experts do when they design a device, when they build up an
infrastructure or optimize a medium. Electronic engineers create a broad assortment of
solutions and the reader perhaps wonders whether they are guided by rational criteria also
inside factories and laboratories; and whether they use theoretical models connected to the
notions of signifier and signified discussed above. The resemblance of digital experts with
chemists should be validated in the professional practice, when customers demand and pay
for an appliance or a service.
In the next pages I aim at an exploration of the living environment and shall discuss the
following topics:

a) The hardware architecture of digital systems – Chapter 4,

b) Computer networks – Chapter 5,
c) Storage – Chapter 6,
d) Protection/security measures – Chapter 7,
e) Software programming – Chapter 9, 10.

Speaking in general, the principles of a discipline are to explain the basic aspects of this
discipline. This is the first task to accomplish, thus I shall comment on the essential features
of the points from a) to e), and shall overlook the specialists’ details.
I shall proceed in the following way: I shall first present an inference derived from the
contents of the preceding pages in the abstract, namely I shall deduce a criterion of work
consistent with the present logical framework. In the second stage we shall check whether the
criterion of work defined on the paper has been adopted by digital experts in professional
practice.
 1. FROM INFORMATION TO UNITS
Digital systems are required to manipulate a broad variety of information: numbers, as
well as texts, pictures, graphics, music, symbols, and sounds. Those systems achieve
mathematical operations and graphical elaboration, they assemble, store and correlate
messages and execute several algorithms. Because of such an ample assortment of services, I
am inclined to conclude that the hardware architecture of a digital system exploits the
properties of information in a systematic manner.
I assumed that the internal of the machine S depends on the qualities of the product w that
S brings forth; thus the following features should determine the operations of a computing
machine:

1. A signifier has a physical basis,

2. A signifier stands for something.

I mean to say that a digital system should transform E and NE respectively; it should fully
exploit properties (1) and (2) in order to process information without restrictions. I proceed by
deductive reasoning and conclude that a system should achieve two families of processes
determined by (1) and (2), and this pair of functions should have the following profiles:

[1] The first process should consist in changing the concrete form of input data. This
operation transmutes the signifiers and keeps the meanings of input and output
unaffected.
[2] The second process should output novel semantic contents with respect to the input.
This function elaborates the subject matter of the input messages and delivers
original pieces of news through an automatic process while the physical appearance
of signifiers steadies.

Are these conclusions true in the real world?

A. Digital Systems

At the present time, electronic manufacturers offer instruments of measurement, control

units, optical devices, medical equipment, calculators and many other original implements in
digital technology. Universal experience shows how all those devices are equipped with
converters that achieve function [1] and with electronic chips that bring forth function [2].
For instance, a cell phone includes integrated circuits that process signals and four converters:
a microphone, a speaker, a keyboard, and a screen. Thus the special method of study which I
introduced in the first pages proves to be correct. Concrete evidence supports the reasoning I
have just developed.
However, I should improve the validation of my theory and the incredible variety of
digital solutions interferes with the accurate analysis of those machines. I am running into
some trouble but happily find assistance in the technical terminology which places all the
digital devices into two boxes. An electronic appliance falls either in the group of special
 People like to Communicate 3

purpose (SP) digital systems or in the group of general purpose (GP) digital systems. The
former is a machine specialized in a definite number of functions, e.g. a digital camera and a
digital control station. The latter is a flexible machine that can fulfill different duties and runs
in various environments e.g. a personal computer and a tablet.
The conversion functions [1] and data-processing functions [2] are not so visible in a SP
system. Things are better in a GP computer. The physical parts [1] and [2] are so large as to
constitute independent units. For example a printer – which converts electrical impulses into
signs of ink – can be separately purchased. Moreover a GP system executes a broad variety of
functions, hence the discussion of conversion and data-processing operations will preferably
refer to computer systems in the next pages.

A. Converters

Computers handle physical parameters such as temperature, hours, and pressure that are
to be displayed in readable forms. Facts show how a GP system operates with a variety of
signifiers required by users and environmental entities, and converters cross certain types of
boundaries from continuous to discrete, from a physical form to another physical form and so
forth.
The technical literature places the conversion devices into the ensuing classes:

1) Sensors transmute a physical signal into another physical signal for example an
optical signal into electrical;
2) Actuators are used to transform a signal into a motion. They are typical components
of robots;
3) Analog-to-digital converters (ADC) change the form of an electric signal from
digital to analog (and vice versa);
4) Modulators change square waves into continuous waves and vice versa;
5) Transcoders change codewords from one coding system to another for example from
ASCII code to EBCIDIC code.

I mean to overlook appliances 3, 4 and 5 since the most significant converters transmute
the physical nature of signifiers and belong to classes 1 and 2 which have been inventoried in
the previous chapter. LED and LASCR have been quoted as examples of optical converters; a
key in the keyboard offers a third straightforward case. When you press the key A on the
keyboard, you enter a piece of information which is mechanical. The key is a switch that
receives a mechanical signal and emits a sequence of electric impulses, namely it converts the
material nature of A from mechanical to electrical.
Only an analog appliance can handle natural signifiers such as the finger press, colored
light, the rotation of a body etc. and the technical literature officially classifies sensors and
actuators as analog equipment (see previous chapter). A device 1 or 2 can include digital
circuits, but those circuits work for the analog core which transmutes the signals. Conversion
operations are under the charge of analog components.
Converters come in many dissimilar types due to the variety of features of the phenomena
they measure: how they work, what they are made of, their function, their cost and their
accuracy; it turns out that two devices belonging to the same class can have far differing
shapes, volumes and prices. We find a converter as small as a tiny chip and a converter as
huge as a unit equipped with sophisticated electronic circuits. Printers, display screens,
loudspeakers, microphones, keyboards, and faxes are popular conversion units that exhibit
different appearances and dimensions (Doyle 1999).

Figure 4.1. Mechanical-electrical converter.

Companies commit substantial investments in research and development of new

transducers for printers that change electric impulses into ink letters on the sheet; for display
screens that convert electric bits into bright points; for loudspeakers that transmute electric
waves into sounds; for disk-drivers that switches magnetic bits into electric impulses. Charge
coupled devices (CCD), LED, magnets, electric motors and other parts provide examples of
analog components employed for conversion purposes and demonstrate how statement [1]
comes true in the world.

B. Data-Processing

Let us conduct an accurate examination of what data-processing really consists.

1. Numerical Data-Processing
We begin by scrutinizing numerical data-processing which probably is the most familiar
form of data-processing. Take the following subtraction where the input numbers signify the
price of a product, and the payment by cash

Figure 4.2. Numeric processing of data.

The output number of this subtraction stands for the change that the cashier has to give to
the client. Note how the outcome symbolizes something new with respect to 10 and 3 in input
and for this reason 7 is useful to the cashier. The result of data-processing conveys meanings
quite different from the input.
This case appears rather elementary as an electrical device creates novel pieces of news
through several operations. A satellite which flies around the Earth suggests a more complex
case to examine. The speed of the satellite is obtained by the time derivative of the space
function s(t) and the acceleration is given by the second derivative of s(t)
 People like to Communicate 5

ds (t ) d 2 s(t )
v (t ) = a (t ) =
dt dt 2

The control station calculates the derivatives, and minute by minute exhibits the results
on the control panel to flight controllers. Numerical data-processing proves to be precious
because in advance of computation the speed v(t) and the acceleration a(t) are unknown
pieces of information for the flight controllers.
Numerical data-processing matches with definition [2] instead the abstract study of
Computing hints at the idea that the systems handle exclusively abstract numbers – see
Chapter 1, Paragraph 5 – and therefore prevents people from grasping definition [2]. I am not
sure that one can realize in what processing of data consists, in reality as long as he/she
discards the significance of numbers.

Language
Processing

Figure 4.3. Linguistic processing of data.

2. Verbal Data-Processing
Probably the reader has read a sentence of this kind: “Data-processing consists of
automatic mathematical calculations”. This definition is to a certain extent true but does not
fit with the variety of operations executed by modern systems. Besides numerical data-
processing more forms of data-processing are widely in use.
All the day long modern computers process linguistic information. Suppose you press the
key-word CELLULAR in a search engine, and in a few seconds you obtain several news and
offerings from the Web.
You can now purchase mobile phones at low cost; in fact the system has provided you
with messages whose contents absolutely differ from the input key-word.

3. Visual Data-Processing
Suppose that a satellite takes a snap and a special coloring program highlights the area
rich in mineral resources using false colors. The announcement of a bonanza is the final
message ignored in advance of visual data-processing.
 Visual
Processing

Figure 4.4. Visual processing of data.

4. Operational Data-Processing
The responses of data-processing just seen are symbols, images etc.; the output may even
be an operational signal that guides a device. For example the subtraction in Figure 4.2 can
control a mechanical cashier that gives the change to a customer through an automatic drive.
The digital system S that has the control of the device C offers another example. Suppose
C breaks down and emits the bit string “10011” to S, promptly the system response “01101”
switches off the device. The digital system sends a signal whose meaning differs from the
input meaning, that is to say, S executes an operational data-processing.

Figure 4.5. Data-processing for control purposes.

The meanings NEinput of “10011” and NEoutput of “01101” are independent of human
feeling. It is unnecessary to translate NEinput into English words such as “failure” and NEoutput
into “switch off”. The verbal descriptions are superfluous since the strings “10011” and
“01101” determine two specific actions. It may be said that “10011” and “01101” have proper
significance in accordance with the remarks placed in the last part of Chapter 1.

5. Static and Dynamical Responses

Data-processing provides single data items or otherwise provides dynamic results and
portrays a real event, or a state of affairs. When the system displays realistic images and/or
allows the operator to use realistic devices, then we have a simulation. For example, a flight-
simulator immerses the operator in a virtual world and the operator has the impression of
flying anywhere in the world by means of a joystick.
 People like to Communicate 7

Figure 4.6. Flight simulator.

Simulation is used in many contexts, including Physics, Chemistry and Biology as well
as engineering and even Economics and Social Science, for the purpose of gaining
information about phenomena which will occur or have occurred in a mysterious way.
Practical evidence shows how data-processing transforms the input subject matters and
prepares novel pieces of news; it delivers original messages; it makes unexpected
representations of the reality and conforms to definition [2]. The present interpretation of
data-processing based on semiotic notions is consistent with common experience.
In current literature some authors talk about “transformation of content” and mean to
describe the translation of a text into a more useful format or into a more meaningful
structure. This operation occurs especially in the Web where different platforms and systems
communicate (Stojanovic 2009) but this special transformation of contents has nothing to do
with the present argument.

6. Impact of Data-Processing in the World

It may be said that most economic processes are information-based in the modern world.
Managers and leaders make influential decisions with success as long as they have trusty
elements in hand, but the manual acquisition of information normally requires exhausting
efforts and substantial investments. The possibility of obtaining pieces of news via swift
mechanical processes turns out to be an extraordinary opportunity. The automatic production
of information items has a vital value and a great effect in the civilized world and in emerging
economies.
The posited concept of data-processing [2], which is grounded on Semantics, offers an
aid to explaining why computers support people everywhere, how computers assist
individuals and communities and achieve dominant success. Conversely one is inclined to
underestimate the relevance of systems to modern economies when he/she assumes that
digital systems treat ethereal values. Computers have a considerable influence on our lives
and strongly affect the course of events and human relations because the strings of bits
represent dollars, euro, gallons, products etc. If by absurd, computing machines ran
exclusively with abstract data, they would have a negligible impact on real life. Very few
people would pay attention to data that has generic reference to physical reality. The
extraordinary contribution of data-processing evaporates and the importance of computer
systems vanishes in the mind of those who have some sort of ethereal view of Computing.
The abstract approach to Computing prevails in some scientific groups which overlooks
the semantic aspects of data and believe that data-processing coincides with calculus. This
idea was not completely wrong in the pioneering age; most software programs were
mathematical programs in those years. Things no longer go in the same way.
Nowadays a lot of programs handle sounds, texts, letters, and pictures which have
nothing to do with numbers. Surveys conducted on the use of programming languages reveal
that mathematical computations constitute a minority group in comparison with the other
forms of data-processing. Fortran ranks at the last place of popularity for PC programmers
(Norton et al 2002). Mathematics-oriented languages such as Fortran and APL hover around
5-10% of programmers’ work-time in mainframe environment. Other recent statistical results
imply that currently numerical data-processing is not a large portion of computer applications.
Innovative data treatments – linguistic, visual, multimedia etc. – are overwhelming and the
percentage of mathematical programs is decreasing day by day.
All this should help the reader to gain and to consolidate a practical view of data-
processing.

7. Creative
Data-processing – the most astonishing activity of computers – brings into existence
original pieces of news through mechanical rules. It is not an exaggeration to claim that data-
processing simulates human thought, in fact thinking is the improvisation of something which
has never existed before. Creative reasoning is the process which people use when they come
up with a new idea. And the information processes illustrated from Subsection 1 to 5 in this
paragraph exhibit something like creative thinking.
Since the seminal article by Turing who introduced the ‘test for consciousness’, scholars
argue about the possibility of reasoning for computers and conduct endless debates on
whether a machine can discover something new (Koch et al 2008).
The controversy centers upon two elements: the computing machines and the human
mind. The former appears rather complex and the latter really constitutes a knotty
conundrum. The two obscure terms of the problems – the computer and the brain – need to be
clarified, and the present logical framework offers a small contribution toward that goal. The
proposed definition of data-processing casts light on the creativity/ inventiveness of machines.
From the present perspective, it is evident that data-processing proves to be a creative
process as long as the computing machine generates original pieces of news. Though data-
processing does not constitute an inventive process in that digital systems apply the
procedures assigned by programmers who establish the plan of work through software
instructions. A processor does not concoct an answer because it cannot operate on the basis of
an original idea as man does. Intuition does not trigger an appliance that instead rigidly
executes the scheduled operations even in the most advanced applications in Artificial
Intelligence.
 People like to Communicate 9

8. How Data-Processing Comes to Be True

Perhaps the creativeness of data-processing puzzles the reader who wonders how data-
processing works in reality. The doubt may be expressed in the following terms:
How can a computer system convey novel information if that machine is unconscious, if
circuits are unaware of data meanings?
Usually authors view this argument through philosophical glasses. Thinkers debate the
autonomy and the consciousness of machines; instead I like better to dissect the material
elements of digital systems that carry on data-processing.
Mechanical data-processing brings forth novel messages due to complex sets of
operations coordinated through cells of memory called data-fields (and also areas, elements,
spaces etc.). Computer operations handle strings of bits in various manners with the support
of data-fields that are portions of space which keep pieces of information for a short while
(Kraft 1979) (Scott 2008). In substance, a specialist – either hardware or software – designs
and implements automatic data-processing by means of correlated instructions and data-
fields. Thanks to this couple of tools, a specialist commands a machine to bring forth novel
contents through an automatic process.

Figure 4.7. The program Speedy.

A very short software program should elucidate how things go. Suppose a computer
operator enters the distance covered by a car and the time spent by this car in covering it. The
program Speedy displays the velocity of the intended car. Figure 4.7 exhibits the list of the
software instructions on the left side, and the data flows through the data-fields on the right
side. Let us examine the algorithm step by step.
It is evident how the software program provides a new item of information – the velocity
of the car – due to the sequel of statements and the set of data-fields prepared by the software
programmer. Ahead of time the programmer is unaware of the numbers that the computer
operator will enter, but this ignorance does not debar the programmer from planning the
overall scheme of work. He is capable of organizing the items necessary for the machine to
obtain the intended results. The correlated arrangements of operations and fields constitute
the factual answer to the serious problems that proved to be intractable from the philosophical
standpoint so far.
The algorithm Speedy turns out to be a straightforward case, deterministic and static. One
could suspect that advanced applications deviate from the aforementioned scheme. Many
problems in Artificial Intelligence such as problems in reasoning, planning, learning and
robotics require an agent to operate with incomplete or uncertain information. Genetic
programming creates evolutionary applications. Advanced software applications are
incomparably more complex than Speedy, nonetheless the previous conclusions remain true in
the sense that instructions and fields are the irreplaceable elements of any software
construction. The methods of work and the algorithms developed in the living environment
cannot be compared with the scholastic example related above. AI solutions are intricate in a
superlative manner but the substantial elements of data-processing – instructions and fields –
are always the same, and these elements enable us to infer further conclusions.

9. What Software Programmers Can and Cannot do

Thinkers have explored the limits of the computer tasks. For long, they have inquired
what a computer can and cannot do.
In formal terms a mathematical problem is said to be ‘computable’ if it can be solved in
principle by a computing device. Common synonyms for ‘computable’ are ‘solvable’ and
‘decidable’ (Cooper 2003). Hilbert believed that all mathematical problems were solvable,
but in the 1930's Kurt Gödel, Alan Turing, and Alonzo Church showed that this is not the
case. There is an extensive study and classification of which mathematical problems are
computable and which are not. Later theorists investigated the computational complexity
classes that classify computational problems according to their inherent difficulty (Epstein et
al 1990).
Theoretical researches are still in progress. A number of intricate questions are waiting a
complete answer but instead of arguing from the abstract stance I prefer to discuss practical
issues on the basis of the elements just introduced.
Software designers obtain the project specifications of what a computer system is
required to do for the customers and investigate whether a set of instructions and fields meets
with those requirements. There is no cookbook for software designers: sometimes the
algorithm is immediate; other times it is hard to devise the algorithm; lastly, a software
solution may be simply unachievable in practice. Normally it is not a question of abstract
‘decidability’ in companies, businesses and institutions; it is a question of several pragmatic
details which turn out to be fuzzy, uncertain or even unknowable by professionals. One could
hold that a computer can treat all the problems that software experts are capable of analyzing
and converting into instructions and data fields. Instructions along with fields are the ‘silver
bullets’ for software practitioners but there is no simple and infallible rule to implement an
effective software application.

2. TWO MODELS ARE NEEDED FOR A DIGITAL SYSTEM

By definition, a digital system manipulates signs and facts make clear that a computer
processes E and NE using distinct units. A conversion process modifies the forms of
information while meanings are kept untouched. I put forward the following formalism to
depict the basic property of converters:

EIN ≠ EOUT
NEIN = NEOUT
 People like to Communicate 11

Data-processing carries on novel meanings while signifiers are uniform from the physical
viewpoint:

EIN = EOUT
NEIN ≠ NEOUT

Practical experience shows how definitions [1] and [2] are true in the working
environment but the analysis is incomplete and one may wonder:

What is the overall model of a digital system?

How can converters and data-processors work together?

Inasmuch as digital engineers follow a pattern of rational behavior, I shall deduce the
answers on the logical plane, later on we shall validate the inferences.

A. Star Model

Converters adapt the pieces of information in aid of the processor that works with
physically uniform signals. Therefore a digital system should be equipped with one data-
processing unit placed at the center and a number of conversion units placed all around it. The
radial model (or star model) should exhibit the logical displacement of parts in a computer
system.
Let us see whether computing machines comply with the inferential reasoning that I have
derived from the definitions of operations [1] and [2].
Computer designers coined the term ‘central unit’ for the data-processing unit because of
its position, while the conversion units were called peripherals or input/output units. The
latter devices lie around the former and prove that the radial configuration is correct.

Figure 4.8. Radial Model.

 The radial model identifies an external unit even if integrated and rather hidden. For
instance, the star model pinpoints the hard disk as a distinct input/output unit whereas the
central box in a personal computer encapsulates the disk drive and conceals this unit from
human sight.
A SP digital system usually has a few peripherals, whereas a GP system offers a variety
of services and is equipped with several units which operators can change at their will.
Devices lately sold as regular stand-alone devices, are becoming peripherals of modern
computer systems. This modern visible tendency in the computer sector is currently known as
digital convergence. The USB (Universal Serial Bus), a plug-and-play interface between a
computer and add-on devices such as audio players, joysticks, keyboards, telephones,
scanners, and digital cameras, reinforces the connectivity of a portable or fixed system. The
Internet facilitates digital convergence as hypermedia information can be instantly transferred
from the Web into a personal computer (Sullivan et al 2004).

Probably the world in the future will become a place without books, letters, post cards,
billboards, telephones, photographs, movies, televisions, stereo systems, and fax machines. In
lieu of the media that we now take for granted there will be the one digital medium that will
take full advantage of the star model.

B. Hierarchical Model

The central unit needs homogeneous signals and can run provided the input and output
flows are homogeneous. Process [1] adapts the physical forms of data in function of process
[2] and it may be said that the peripherals are servants of the central unit. Peripherals appear
to play a subordinate role with respect to the central unit according to the present reasoning.
A principle of operational mastery which may be associated with the arguments treated in
(Pattee 1973) should govern the computer system from the current viewpoint and the
hierarchical tree (Salthe 2001) should complete the description of the computer hardware.
Let us check whether the previous reasoning deduced on the basis of definitions [1] and
[2] is correct.
Common practice shows how it is not sufficient for a conversion unit to be connected;
this unit can only run under the central unit grant (Brookshea 2004). There is no other way to
avoid conflicts amongst the peripherals; an input/output unit is idle unless the central unit
empower this unit. In other words all the peripherals lie under the operational supervision of
the central unit.

Figure 4.9. Hierarchical Model.

 People like to Communicate 13

Computer designers strengthen this natural hierarchy and place the control unit (CU) – a
circuit specialized for the operational control of the overall system – into the central unit.
Normally CU integrates into a chip named central processing unit (CPU) and the acronyms
CU and CPU are frequently used as synonymous (Green 1988).
A computing machine is equipped with hybrid devices, and CPU orchestrates the varied
components in order to avoid conflicts amongst the parts and to optimize the overall
performance. In this way the CPU reinforces the supervision role played by the central device
and one can conclude that the tree model turns out to be even truer in this case.
Several hardware procedures relate minutely how CPU executes its operational
supervision. I cannot look closely at those rules, and confine myself to three examples that
should clarify the style of the central unit and should furnish further evidence of the correct
application of the tree model.

• A peripheral device cannot receive or emit data unless the CPU enables this
peripheral to work. Any external operations cannot commence without the central
authorization. The periphery does not run autonomously but relies on the center.
• As soon as a peripheral ends a task, the CPU checks the results of the input/output
operations that have just come to an end. In other words, all the outcomes are tested
and assessed with care by the central device that undertakes the appropriate
initiatives to manage the situations.
• When a special event occurs – e.g. a circuit breaks down – every running operation
comes to a halt at that moment and the Central Processing Unit takes full control of
the machine and handles this event. This proves that the CPU leads all the units
during ordinary jobs and special occurrences alike.

CONCLUSION
This chapter proves how the digital systems share a universal structure. A logical chain
leads us from the semiotic properties of signs to the system units, and in turn to the radial and
the hierarchical models which depict the hardware structure of any digital system.
Facts demonstrate that systems conform to the logical conclusions developed in this
chapter independently of whether a system is stand alone or connected to a net, whether a
system is embedded in a machine or a big supercomputer. One can verify the star model and
the tree model in cellular phones, hi-fi pods, satellite navigators, flight control equipment and
so forth.

In the second half of the twentieth century we observed the astonishing proliferation of
digital machines which invaded the market and replaced analog devices. After millennia, the
beneficial support of the analog paradigm to human progress gave the impression of ceasing
all at once. The analog appliances which boosted the advance of humankind seemed to be
definitively destined to vanish despite their outstanding support to civilization. As dinosaurs
suddenly became extinct million years ago, so the analog devices appear to be destined to
certain death. This impression however is misleading; we have seen how analog devices – the
converters – occupy prominent positions within the digital system structure. The accurate
analysis just conducted leads to a rather paradoxical conclusion: no digital computer can run
without the support of analog technologies.

A. Annotations on Current Literature

Commentators usually refer to the “First Draft of a Report on the EDVAC” written by
von Neumann (1993) and to other earlier contributions (Booth 1960). John von Neumann
presents the overall organization of a system equipped with a stored program and the ‘von
Neumann model’ constitutes a cornerstone in Computer Science.

Figure 4.10. Input-Process-Output Model.

Unfortunately theorists do not have a strong inclination to plunge into the general
features of digital systems. Several papers and books describe the von Neumann model using
a linear graph that includes three blocks (Forouzan 2003). The Input-Process-Output (IPO)
model sometimes shows the memory, the CPU and other details; these variants are inessential
for the moment. One can remark:

• The IPO model is consistent with the radial model but is linear and conceals the
versatility of the general purpose computers. IPO hides the physical variety of
messages held by systems. IPO simplifies the hardware structure to the extent that
IPO prevents a student from understanding the digital convergence and its enormous
factual impact. At least, it may be said that IPO distorts the computer hardware.
• The linear chain implies – in conformity with Markov’s chain theory – that the
central block depends on the input block; instead the contrary is true. The process
unit factually governs all the external operations and an input unit cannot run until
the central unit enables the operation in advance. Thus IPO gives misleading
information.
• Surhone and others (2010) hold that IPO is an interdisciplinary tool. In fact, a lot of
systems in various sectors of production – e.g. agriculture, mining, hydraulics, etc. –
comply with the IPO graph in the sense that several automatic systems are served by
units that introduce raw materials and bring out finished goods respectively. IPO is
flexible however the labels ‘input’, ‘process’ and ‘output’ sound rather generic in the
computer sector until a commentator specifies in what the peripheral processes and
the central processes consist. If theorists do not specify the special treatments
undergone by information crossing the computer system, the meaning of the IPO
model appears very approximate and the overall purpose of the system is ephemeral.
• Scholars do not accurately examine the real origin of data-processing and are
inclined to conclude that data-processing and calculation come to mean the same
thing; conversely nowadays digital systems treat various forms of data in the real
world and do not restrict their intervention to numbers.
 People like to Communicate 15

The lack of an accurate description of the system hardware entails that technical manuals
seem to be written as cookbooks. As a cookbook provides a guide for expert cookers, so
modern, up-to-date textbooks are able to assist learners in the use and installation of
computers. Theoretical constructs are perhaps unnecessary for those who become wizards at
the operative level but prove to be essential for those who manage or improve an ICT
infrastructure. The sciolism in Informatics is myopic and does not provide thorough
knowledge. We shall see next how a number of computer disasters are caused by poor
education (Dvorak 2004); in fact, the behavior of digital designers appears intelligent, but the
benefits deriving from their conduct sometimes do not fully benefit the people involved due
to scarce explanations. The consequences of the approximate theoretical support on
computing are not of negligible significance and value.
Von Neumann adds this comment to the paper quoted above:

“The three specific parts [that make the central unit n.d.r.] correspond to the associative
neurons in human nervous system. It remains to discuss the equivalents of the sensory
organs or afferent and the motor or efferent neurons. (…) These are the input and output
organs of the device, and we shall now consider them briefly. (…) There remains (…) the
necessity of getting the original definitory information from outside into the device, and
also of getting the final information, the results, from the device into the outside.“
(Neumann 1993)

Von Neumann compares input/output devices to the human sensory organs and associates
data-processing to the neural nets theorized by McCulloch and Pitts in those years. This
analogy with the human nervous system raises vivid suggestions (Boden 2008) but appears
rather questionable on the intellectual plane. Neumann aims at explaining a rather mysterious
topic – say the peripherals and the data-processing unit – by using a comparison term that is
even more enigmatic. Usually scientists follow the opposite direction. Doctors illustrate a
biological organ using an artificial device, for instance, the bones of the arm and the forearm
are compared to mechanical levers; the heart looks like a pump, the blood vessels are similar
to pipes. In fact a machine, designed and built by humans lies normally under the full control
of the machine’s authors; by contrast living beings are ready in Nature and scientists have to
decipher their complex functions. Biologists and physicians help themselves by means of
technical concepts which make it easier to understand biological parts. Neumann seems to
adopt the opposite method which sounds rather strange; he attempts to elucidate the functions
of a computing system by means of the parallel with the human brain which is still a
challenge for researchers world-wide.
All this should prompt accurate studies of the structure of the digital systems. The basic
properties of the computer hardware should be brought into the open and its merits debated,
but this project has not progressed far. Theorists have not proceeded to investigate the
structure of the computing machine and this passive orientation is not a trifling matter, since a
number of fallacies and misconceptions, which are caused by the negligent behavior of
theorists, may be found in the current literature. The magnitude of the negative effects in the
living environment appears even more manifest.
 BIBLIOGRAPHY
Boden M. (2008) - Mind as Machine: A History of Cognitive Science - Oxford University
Press.
Brookshea Glenn J. (2004) - Computer Science: An Overview - Addison Wesley.
Booth A.D. (1960) - The Future of Automatic Digital Computers - Communications of the
ACM, 3(6), pp. 339-341.
Cooper S.B. (2003) - Computability Theory - CRC Press.
Doyle L.F. (1999) - Computer Peripherals - Prentice Hall.
Dvorak J.C. (2004) - The Bottom 10: Worst Software Disasters - PC World Magazine, 8.
Epstein R.L., Carnielli W.A. (1990) - Computability: Computable Functions, Logic and the
Foundations of Mathematics - CRC Press.
Forouzan B.A. (2003) - Foundations of Computer Science, From Data Manipulation to
Theory of Computation - Thomson.
Green D. (1988) - Digital Electronic Technology - John Wiley & Sons Inc.
Koch C., Tononi G. (2008) - Can Machines Be Conscious? - IEEE Spectrum, 45(6), pp. 55-
59.
Kraft G.D., Toy W.N. (1979) - Mini/microcomputer Hardware Design - Prentice-Hall.
Neumann von J. (1993) - First Draft of a Report on the EDVAC - IEEE Annals of the History
of Computing, 15(4), pp. 27-75.
Norton P., Clark S. (2002) - Peter Norton's New Inside the PC - Sams Publishing.
O' Sullivan D., Igoe T. (2004) - Physical Computing - Muska & Lipman.
Pattee H.H. (1973) - The Organization of Complex Systems - in Hierarchy Theory: the
Challenge of Complex Systems, Pattee H.H. (ed), Braziller, pp.1-27.
Salthe S. (2001) - Summary of the Principles of Hierarchy Theory - MIT Press.
Scott M.L. (2008) - Programming Language Pragmatics - Morgan Kaufmann.
Surhone L.M., Tennoe M.T., Henssonow S.F. (2010) - IPO Model - Betascript Publishing Co.
Stojanovic D. (2009) - Context-Aware Mobile and Ubiquitous Computing for Enhanced
Usability - IGI Global Publ.