Introduction to Software Systems Unix C + GNU HTML Perl Python

Transcription

1 Introduction to Software Systems Unix Comp-206 : Introduction to Software Systems Lecture Week 1 Joseph Vybihal Computer Science McGill University

2 Announcements Instructor coordinates Course outline Participation Web CT discussions, assignments, mail, lectures 2

3 Unix 3

4 Why COMP 206? Application programs are no longer created from a single programming language. Database languages, off the shelf libraries, and the internet has changed programming forever. This courses asks you to build a web store, a single application, made from:, CGI, Java, C,,, Bash and Unix! 4

5 Why COMP 206? No longer green screens and simple command-line prompts Windows & Apple Unix and Linux Minis & Mainframes 5 (also robots, POS real-time)?

6 Why COMP 206? How software inter-plays... Libraries Process inter-communication Networks Multiple languages Operating Systems How to use software tools... Team programming Debugging Code optimization Code generators The idea of a software SYSTEM... Single application Multiple languages Multiple tools Multiple operating systems Network connectivity Multiple hardware architectures 6

7 Official Course Description Comprehensive overview of programming in C, use of system calls and libraries, debugging and testing of code; use of developmental tools like make, version control systems.

8 My Course Description Introduction to Software Systems is a course about the various types of techniques and tools used to build software.... serves as your introduction to the C programming language, Unix operating system, and internte programming.... gives you tools to be more productive during your undergraduate studies. gives you the opportunity to develop complex integrated software projects. 8

9 Course Content The course will cover the following topics: Unix operating system Shell Scripting C Programming (and related techniques & tools)? C Libraries Debuggers and Profilers Source Control, CSS and CGI and Web scripting (not exactly in that order)? 9

10 Knowledge of Unix This course assumes you have some basic knowledge of Unix (and the school labs). login into your account, copying a file, editing a file, etc If you don't, I highly suggest you attend one of the SOCS Unix seminars, to be held in the Trottier Building, 3rd floor, Room 3120 (Lab 2). Check the SOCS web site for tutorial times We will also do our own tutorials and review in class 10

11 Part 1 What are Software Systems?

12 What is it? Software Systems Which are the systems? What systems do you think makeup a web site? A software system is The software development environment The software execution environment The application program system 12

13 Software Systems Development Environment System Compiler Library Code generators Version control and optimization software Execution Environment System Interpreters Operating Systems Network communication features Hardware available Number of CPUs Program System Compiled (some in C, some in Pascal)? Scripted (some in Unix, some in or )? Interpreted (some in, some in Java)? Inter-communication (network and OS provided)? How do they interplay? 13

14 A Software System Definition Unix A group of programs created to interact with one another to solve a common application goal. Examples of programs working together: An object instantiated from another program A system call to the OS A web request to a server 14

15 Software System Boundaries The location where one subsystem interacts with another subsystem. Think of this as an internet application Where is it? and How does it inter-communicate? 15

16 What is Software? Software is a collection of instructions (often grouped as functions and libraries) that allow a computer to complete a specific task. Although these instructions can be written in different languages, they are eventually converted to a machine language. 16

17 C C++ Java Fortran Ada N : 1 Assembler 1 : 1 Numbered codes (machine language)? Circuits Everything in the end is the same... 17

18 Computer Programming Computer Programming is the craft (art) of writing the computer instructions that accomplishes a specific task. The difficulty of computer programming depends heavily on the task itself and the tools that are used to help the programmer. One of the keys to being a successful programmer is knowing which tool to use in each different situation. Not all programming language were designed to do the same thing. 18

19 Name that Programming Unix Java C C++ Basic Ada Lisp Scheme Javascript AppleScript Ocaml D and so on... Language State their purpose... 19

20 Definition (strong form): A System A program constructed in such a way that it is complete in itself. It requires no external calls to other software. Internally, the system is composed of modules or objects that, like a system are complete in themselves, but provide public access methods that allow interconnections between modules/objects. Example: the compiling process, os environment, oo-programming 20

21 What is an Operating System? An operating system is a piece of software that allows us to interact with a computer without having to know the inner working of a computer. Its primary function is to manage the computer's resources. An operating system also provides us with libraries to interact with these resources. 21

22 Name that Operating System Dos Windows Solaris Linux FreeBSD BeOS FreeDos HP-UX AIX MacOs X What is speacial about them? 22

23 What is a Library? A library is a piece of software specially packaged to be used by other software. A library provides specific functionalities, thus avoiding the need for the programmer to build the functionalities him or herself. 23

24 Library Examples On Unix, the Standard IO library provides functionality to open files and write to them. Without it, a programmer would need to write code to use the hard disk itself in assembler. On Windows, DirectX allows a developer to write display graphics. Without it, a programmer would need to write code to access the video card directly. 24

25 Libraries and the OS From Compiler Or 3 rd party From OS Software Libraries Operating System Drivers Hardware High-level libraries (in C or assembler)? Low-level libraries (in assembler)? 25

26 Part 2 Run-time Environments To understand software systems, one must also appreciate and exploit the run-time environments applications execute within.

27 The Run-time Environment Software You program this You take advantage of this Libraries Operating System Drivers Hardware The run-time environment 27

28 The Development Environment The development environment Comparison Language Tools Compiler Software Libraries Operating System Drivers Hardware The run-time environment These may also come from the compiler or a third party. 28

29 Different run-time environments

30 Still important to know Unix Single CPU Single Process Examples: MSDOS, CPM, ProDOS, MiniUnix Properties: command-line prompt only one program runs at any time Architecture: Advantages: Direct hardware control Low OS security TSR Ability Real-time systems OS video drivers kernel interface User Space Zero Heap Run-time Stack Code Data C:>dir 30

31 Single CPU Multi-Process Examples: Windows, Unix, Linux, Mac OS Properties: command-line prompt or windowed interface more than one program can run at same time Architecture: Advantages: Spawning, Killing, status OS Hardware Services Inter-process communication OS ptr drivers kernel time slice To graphics card interface C:>dir User Space Heap Run-time Stack Code Data Heap Run-time Stack Code Data C:>dir Zero P1 P2 31

32 Multi-CPU Multi-Process Examples: Windows NT Family, Unix Pro Family Properties: command-line prompt or windowed interface more than one program can run at same time Schedules and divides processes between CPUs Architecture: Advantages: True Divide & Conquer Large programs / data Master / Worker processes OS ptr User Space Zero drivers kernel time slice interface CPU management Method 1: give a Pi to a CPUj for entire execution life Method 2: divide Pi into parts and share parts with CPUs Method 3: pool of CPUs time-sliced with pool of Pi s 32

33 Batch Processing Examples: AS400 (simulated in DOS, Windows and Unix)? Properties: Program execution is managed by a simple OS script language User activates script, then script takes over Advantages: Scripts are text files interpreted by the OS Pseudo-new OS commands Automate manual processes (backup)? Execution environment auto-setup Architecture: OS ptr User Space Zero drivers kernel time slice interface CPU management interpreter Setup OS environment: map drives, check hardware Setup program s execution environment: validate, copy files Run program multiple times for each input data set Clean up: cancel maps, delete temporary files, do backups 33

34 Distributed Processing Examples: Java,, and Properties: Uses a network to share resources and execute processes on different computers Architecture: PC / Server interface PC / Server run-time clients Advantages: Not CPU dependent Spawn process across network Access data across network Server / Client architecture Master / Worker architecture 34

35 Code Migration Processing Examples: Experimental Properties: Process Migration When you execute a program on your PC and your PC is too slow, or too busy or does not have enough memory the OS will automatically use the network to download your program to another computer to execute on (invisible to you) recursively implemented. Object Migration Advantages: The pseudo-infinite computer Same idea but with each object or module of your program. 35

36 Questions How do we code to take advantage of these run-time environments? Single process Multi-process Multi-CPU Distributed processing What features do we have access to in one environment and not the other? 36

37 On your own At Home

38 Try the following Using your two most favorite operating system environments, identify the libraries and tools provided by each one. Then answer the following: Which one gives you better libraries and tools? Is the development environment optimized for a particular type of application programming? After compiling a program, which development environment generates big and heavy code (measured in bytes after compiling)? Does this effect your judgement on which is a better development environment or effect your view on its usefulness? Is GUI or command-line better? Why? 38

39 Part 3 Software Development Techniques

40 Integrated Programming Unix Microsoft Visual Programming Borland Builder and Studio Environments (Borland Demo)? Free non-integrated versions on the web (just google it)? Eclipse (this is installed in the Trottier labs)? Windowed environment that permits you to manage and compile many source files at once. Handles access control and version control. Handles modification management as well. demo 40

41 Command-line Programming Unix Command-line prompt Environments Individual programs: Editor program Compiler program Debugger program Project builder program Our course focuses on this. demo 41

42 Single Source File Programming The entire program exists within a single source file There is only a single user File is stored in a single directory Manipulated using (command-line programming): vi filename gcc filename > error_messages a.out more error_messages 42

43 Multiple Source File Programming Each source file is organized around an object or a module Assumes multiple programmers Issues: Access Control (who can edit the file)? Version Control (access to older/differnt version of code) Manipulated using: vi filename_i make project > error_messages a.out more error_messages 43

44 Source File Management Techniques

45 Manual Directory Organization Single Programmer Project - Directory / Folder Structure Projects Project1 Archive Documents Backup Build1 Zip old builds increment Description of what needs to be programmed : Project2 Archive - long-term backup (history)? Backup - short-term temporary files 45

46 Manual Directory Organization Team Programming Project - Directory / Folder Structure Projects Project1 Project2 : Development Programmer A Archive Documents Backup Build1 Programmer B : Integration Archive Documents Logs Build Zip old builds increment Merge, one programmer at a time, only changed files. After the merge, must compile and make sure the entire project still runs correctly. This is either the responsibility of the programmer who did the merge or a special integration programmer. 46

47 Modifications Management (Writing code so that it is easier to determine what went wrong in a program and how we might fix it)?

48 Comments as Documentation Why write a programmer s manual when you can do that through comments? Per method, function, object, source file // Method Name: Version: // Programmer : // Created : // Purpose : // Inputs : // Outputs : // Notes : // Modification Log // ============= // DATE NAME REASON/DESCRIPTION In-line comments (descriptive and flow)? 48

49 // Step 1: Get input from the user for(int I = 0; I<10; I++) scanf( %d, ages[i]); // Step 2: Sort values using bubble sort In-line flow comments Blank lines, to put concepts into block and help readability for(int I=0; I<10; I++)? { Good variable name for (int j=0; j<10; j++)? { if (ages[i]>ages[j]) swap(&ages[i], &ages[j]); // swap trades the values } } // Step 3: Print out the results to the screen Constant indentation for(int I=0. I<10; I++) printf( %d\n,ages[i]); In-line explanations 49

50 Errors to fix Log Files Filename, description, line number Integration history Date, by who, problems/fixes/comments Validation testing Filename/object/form, what to do, expected result 50

51 Automated Project Management (later in course)? Project / Make files RCS Revision Control Systems Modifications Management Techniques 51

52 Part 4 Operating Systems (the OS)?

53 Question What is an OS? What does it do? Does is have responsibilities? How does it do it (how is it programmed)? 53

54 The Operating System 54

55 Services provided by the OS Process management Allows applications to run simultaneously Memory management Manages the allocation of memory Disk and file systems Manages the writing to disk of files. Networking Allows inter-computer communication. Security Provides authentication, privacy and protection. Device Drivers Allows the use of hardware in a generic fashion. Graphic User Interface (and command-line interface)? Provides a visual interface to interact with the computer. 55

56 Process management The first generation of operating systems only allowed one process to run at a time. A multi-tasking operating system allows multiple task to run at a time. Most computers only have 1 CPU, the multiple task must share this CPU. Most often, the CPU will time-slice the task, so they get a fair amount of CPU time. Computers with multiple CPU can run multiple task at a time. This introduces numerous concurrency and coherence challenges. Any Concurrency examples? 56

57 Memory Management At the hardware level, computer memory is a collection of 0 and 1, stored in volatile silicon chips. At the operating system, computer memory is an array of bytes, where data can be store and later retrieved. The operating system is responsible for allocating blocks of memory to the different processes. When the OS runs out of memory, it can use the hard disk to temporarily store blocks of memory that it does not often use. Do you know what the standard memory organization in the computer is? 57

58 Disk and file systems At the lowest level, a computer disk is a collection of 0 or 1 stored on a magnetic or optical disk. This hardware has no understanding of files and directory. The operating system provides file systems, which describe how files and directory should be stored on the disk. Each operating system feature different file system: Windows : Fat32, NTFS Linux: EXT2, EXT3, XFS, Reiser, etc MacOS X: HFS+ Describe the standard diskette organization, what about CDs? 58

59 Networking At the hardware level, networking is the exchange of 0 and 1 over a communication channel. A network protocol describe how data should be transformed (and encapsulated) before being sent over the network. The OS provides the necessary tool for a computer to join a network and transmit(receive) data over this network. The most common network protocol is TCP/IP. What should the computer do if the channel is noisy? 59

60 Security Operating Systems feature authentication systems which control who can use a computer. They also offer privacy features on many levels: A user/administrator can define which files can be used by which users. A process should not be able to access the memory block of another process. A good operating system will also monitor itself to prevent action which could compromise it s stability: A process should not be able to delete a file which is currently in use. Certain spaces in memory should never be overwritten What should be secured in a public computer? 60

61 Device Drivers Each device connected to the computer has their own different way of communicating with the computer. Device drivers are small pieces of software which allow an external device to communicate with the computer in a generic fashion. Without device drivers, an application would need to deal separately with all the different external hardware it supports. How does one communicate with: the screen, keyboard, or printer? 61

62 Abstraction Unix Graphic User Interface A graphic user interface (also known as a GUI) is a piece of software which allow interaction with the computer using visual components (icons, buttons, scrollbars, etc). These visual components are often referred to as widgets. A GUI also provides libraries which allow the development of new software with these visual components. Before GUIs were distributed with operating system, interaction with the OS was often achieved through a CLI (command-line interface). 62

63 Open vs. Closed Operating Unix Systems An OS is closed (or proprietary) when it owned by a single company. It is often designed to work on a single kind of hardware. An OS it open (or non proprietary) if no single company is responsible for its development. Since many people can work on the code, it is often available on different hardware platforms. 63

64 Open FreeDos Linux FreeBSD OpenBSD Solaris Open or Closed Closed MS Dos MS Windows HP-UX AIX 64

65 On Your Own At Home

66 Try the following Get into a team of 3 to 4 people and think of a way to develop code with multiple programmers that would minimize the need for programmers to modify each other s code. What disciplines would your team of programmers need to follow to insure that this works? What files or logs would you need to keep? Is your technique hard to do? If so, how can you make it easier to work with? 66