semester_7

         ****************************************************
       ***                                                  ***
      **               SEMESTER 7 - COURSES                   **
       ***                                                  ***
         ****************************************************

In this semester, we are going to have totally new & interesting
stuff. According to Omar Yousry, this semester is technically important
as it contains topics that are considered as the core of our field,
like operating systems, microprocessors and algorithm analysis.

The courses:
-------------

    1. CS 311 - Analysis of Algorithms    - 5 hours - Group 1 (MATH AND TCS)
    2. CS 331 - Microprocessor Systems    - 6 hours - Group 3 (HARDWARE)
    3. CS 332 - Digital Signal Processing - 5 hours - Group 3 (HARDWARE)
    4. CS 333 - Operating Systems         - 5 hours - Group 3 (HARDWARE)
    5. CS 341 - Discrete Control Systems  - 6 hours - Group 4 (CONTROL)
    6. HS 373 - Man-Machine Interface I   - 3 hours - Group 7 (COMPUTER)

As you can see, there is no any single course from CS.group2 :')

-------------------------------------------------------------------------------
                         1. Analysis of Algorithms
-------------------------------------------------------------------------------

This course is actually two parts:

    Part 1 (usually Dr. Amr Elmasry):
    ----------------------------------
    This part is mainly concerned with "theory". It is well known
    that theory of computation is one of the fundamental topics
    in computer science. Theory of computation is divided into three
    major branches:
        a. Automata Theory: The study of abstract machines (like FSM & Turing)
        b. Computability Theory: To what extent this problem is solvable
                                 on a computer?
        c. Computational Complexity Theory: I think you know it well.
    This course is mainly about (c). We will study proof techniques,
    fundamental analysis of algorithms, complexity classes, and
    some other hot stuff that I have no idea what they are.

    Part 2 (usually Dr. Mohamed Elsayed):
    --------------------------------------
    Problem Solving Techniques: Direct methods, greedy methods,
                                backtracking, etc..
    There will also be some lectures about graphs (again, but am
    not sure really), geometric algorithms, and asymptotic algorithms.

-------------------------------------------------------------------------------
                         2. Microprocessor Systems
-------------------------------------------------------------------------------

In my point of view, this course is about "how to use microprocessors
to build microcomputer systems", not "how to build microprocessors".
However, we will have a lecture on control unit design.

A microprocessor is a processor that is integrated in one chip.
This allowed people to design single-board computers (SBC). And
because single-board computers and microprocessors are everywhere nowadays,
it is really worth it to study this course.

The course consists of two parts:

    Part 1 (usually Dr. Leila):
    ----------------------------
    8086: architecture, programming, and interfacing (8086 bus,
          I/O interfaces, memory interfaces, computer design
          and organization).
    Control unit design: hardwired control unit - microprogramming.

    Part 2 (usually Dr. Salah):
    ----------------------------
    Some various topics about computers, like the PCI bus, USB bus,
    and optical storage.

If there is time and interest, we might also study more advanced topics
like 80386 protected mode, paging, and virtual memory.

-------------------------------------------------------------------------------
           3. Digital Signal Processing and Transmission Algorithms
-------------------------------------------------------------------------------

Digital signal processing (DSP) is one of the substantial topics of electrical
engineering and computer science. DSP is the manipulation of signals
in discrete domain rather than continuous domain. So, what does that mean?

High pass filters are used to pass high-frequency components of the
signal and block low-frequency components. This has a lot of applications
in audio and image manipulation software. This filter can be constructed
using a simple RC circuit:

                           C
                          | |
                    O-----| |---------*-------O
                          | |         |
                                     ---
                                     | |
                   Vin             R | |     Vout
                                     | |
                                     ---
                                      |
                    O-----------------*-------O

If we take the output as the voltage across the resistor, we will
have the same input signal but with low-frequency components removed.
Why it works?  Let's see what happens to sinusoidal signals:
Simply, the reactance of C is much greater than R at low
frequencies (Xc = 1/wc). Because Vout = R/(R+jXc) * Vin,
we find that Vout = <a very small imaginary number> * Vin. Thus we
can deduce that lower frequencies disappear in its way. We can
also deduce that higher frequencies pass through the circuit smoothly.

This circuit is actually an *analog* band-pass filter; it works on analog
signals. Believe it or not, we can implement this thing using
a digital microprocessor!

As we can see:

    Vout = IR
    I = Cd(Vin-Vout)/dt
    thus: Vout/RC = dVin/dt - dVout/dt

    If we take samples of the input signals (Vin0, Vin1, Vin2, ...):
    Vout(i)*(dt)/RC = Vin(i)-Vin(i-1)-Vout(i)+Vout(i-1)
    Vout(i)(1+dt/RC) = Vin(i)-Vin(i-1)+Vout(i-1)

    Then:
       -------------------------------------------------
       - Vout = RC*(Vin(i)-Vin(i-1)+Vout(i-1))/(RC+dt) -
       -------------------------------------------------

dt is the sampling interval. We can now write a program that
takes Vin(0), Vin(1), .... Vin(n) as an input, and outputs
Vout(0), Vout(1), ....

This is what we call DSP. It is how you can process signals in digital
rather than analog. Usually, a digital signal processor consists
of the following:

                                +----------+
                                |          |
                  *-----<-------| Sampling |
                  |             |   Clock  |
                  |             |          |
                  |             +----------+
                  |                   |
                  Y                   Y
                  |                   |
           +------*-----+          +--*--+         +------------+
           |            |          |     |         |            |
    Analog |  Analog    | Digital  |     | Digital |  Digital   | Analog
(in)------>|  Digital   |--------->| CPU |-------->|  Analog    |------->(out)
    Signal | Controller | Signal   |     | Signal  | Controller | Signal
           |            |          |     |         |            |
           +------------+          +--*--+         +------------+
                                      |
                                      |
                                      |
                                 +---------+
                                 |         |
                                 | Program |
                                 |  Memory |
                                 |         |
                                 +---------+

Our course consists of two parts, the first part is mainly concerned
with discrete-time signals, discrete-time systems and transforms
(Fourier's transform and z-transform). No, we won't study sampling theory
or digital filters. Math is enough, applications are not important.

               IT IS FULL OF MATH, SO TAKE CARE.

The other part is "information theory". Information theory is mainly
about quantification of information. Applications to information
theory (which we are going to study isA) are:
    -> LOSSLESS DATA COMPRESSION   (know .gz and .zip? :D)
    -> LOSSY DATA COMPRESSION      (know mp3 and jpg?  :D)
    -> SIGNAL AND CHANNEL DECODING (know ADSL?         :D)
The study of error detection and correction codes will also be covered
in this course. As you can see, this course is really important and
contains very tough topics which are 100% related to our study. I don't know
why many people hate it.

-------------------------------------------------------------------------------
                          4. Operating Systems
-------------------------------------------------------------------------------

Operating systems do two main jobs:
    - controls the hardware.
    - controls the software.

This course shows (in abstract sense) how operating systems do the mentioned
jobs, with case studies (like UNIX System V and Linux kernel). We
will also learn how to write an operating system in JAVA, Because f**k C,
f**k Rust, f**k C++, f**k everything but java. Accordingly,
we won't see how the OS interacts with the hardware *IN REAL*.

Sorry but don't make high hopes on this course. Too many sheets and too much
theory. Labs are too little and they are mostly concerned with how you program
concurrent threads on Linux. This is useful for application programmers but has
nothing to do with the design of the operating system. This shows you what they
actually want us to be.

The lectures are interesting though:
-------------------------------------
    - Introduction to operating systems
    - Basics & Principles
    - Processes and threads
    - Scheduling algorithms
    - Synchronization and mutual exclusion
    - Memory management (partitioning, segmentation, etc...)
    - Paging and Virtual memory.
    - Hardware and I/O.
    - Disk management
    - File systems

-------------------------------------------------------------------------------
                5. Discrete and Nonlinear Control Systems
-------------------------------------------------------------------------------

This time we are working with nonlinear systems instead of linear systems.
Laplace transform is replaced with z-transform, and RLC circuits
are replaced with *COMPUTERS* :D Just like DSP, ADCs and DACs are
used. It is pretty interesting to see how computers can process
this stuff.

Digital controllers (like PLC) are usually used to control industrial
processes. However, microcontrollers (like PIC and AVR) are sometimes used for
digital control because they are cheap.

The course will include these topics:
    - Introduction to the *DESIGN* of systems (instead of *ANALYSIS*)
    - Discrete-time systems (isn't this related to DSP?)
    - Sampling and reconstruction (isn't this related to DSP?)
    - Z transform
    - Analysis of discrete systems
    - Stability of discrete-time systems
    - Nonlinear control systems
    - Lyapunov stability

-------------------------------------------------------------------------------
                         6. Man-Machine Interface
-------------------------------------------------------------------------------

This is a typical HCI course and is usually taught by Dr. Salah. HCI
involves the study and design of the interaction between users and computers.
Mouse is one of the most famous results of HCI.