Skip to content

Latest commit

 

History

History
253 lines (213 loc) · 15.8 KB

07-SoftwareDesign.md

File metadata and controls

253 lines (213 loc) · 15.8 KB
Raw

Software Design

  • BIG IDEA — How do we talk about these things with other people efficiently and effectively?

Algorithms - How are we gonna solve this?

algorithms

Time Complexity - How long is this gonna take?

time-complexity

  • BiG IDEA — Let's have a standard way to communicate about how long an algorithm or program will take to execute.

  • The "time complexity" of a program is a measure of how the "running time" of the program grows as the "input size" of the program grows.

  • Also known as "Algorithmic Complexity" or "Big O notation"

  • The "time complexity" is usually expressed as a "big O" notation, which is a way to express the "upper bound" of the "running time" of the program, ie: the worst case scenario.

  • The "big O" notation is a way to express the "upper bound" or maximum expected of the worst-case running time of a program.

  • 6 Common Types of Time Complexity:

    • O(1) - "Constant Time" - The running time of the program does not change as the input size of the program grows.

    • O(n) - "Linear Time" - The running time of the program grows as the input size of the program.

    • O(n^2) - "Quadratic Time" - The running time of the program grows as the square of the input size of the program.

    • O(n^3) - "Cubic Time" - The running time of the program grows as the cube of the input size of the program.

    • O(log n) - "Logarithmic Time" - The running time of the program grows as the logarithm of the input size of the program.

    • O(n log n) - "Linearithmic Time" - The running time of the program grows as the product of the input size of the program and the logarithm of the input size of the program.

  • Shorthand Guide & Sample Code

    • time-complexity-example

    • If you are doing a simple lookup in an array, it's O(1) because the time it takes to find the value in the array does not change as the size of the array grows.

    • If you are looping over items, it's O(n) because the time it takes to loop over the items grows as the size of the array grows.

    • If you are looping over items that are also looping over items, it's O(n^2) because the time it takes to loop over the items grows as the square of the size of the array grows.

    • Each time you add another loop, the time complexity grows as the square of the size of the array grows.

    • If you are doing a binary search, and each time you search you are cutting the size of the array in half, it's O(log n)

      // Program in Kotlin to illustrate the time complexity of various operations
// Note: The `//` symbols means the rest of the line is comment, it is not part of the program, it's just for explanation and is ignored by the compiler.

// Algorithm: Binary Search
// Perform binary search a sorted array - O(log n)
// Returns the index of the value in the array, or -1 if the value is not found.
fun Array<Int>.binarySearch(value: Int): Int {
    var low = 0
    var high = size - 1
    while (low <= high) {
        val mid = (low + high) / 2  // search 1/2 of the array each loop
        val midVal = this[mid]
        when {
            midVal < value -> low = mid + 1
            midVal > value -> high = mid - 1
            else -> return mid // ends the looping when the value is found
        }
    }
    
    return -1 // -1 = value not found
}

fun randomInt(max: Int): Int = (0..max).random() // <-- O(1) - Simple Lookup

// Algorithm: Quick Sort
// Perform recursive quicksort - O(n log n)
// Returns a new sorted array.
fun Array<Int>.quickSort(): Array<Int> {
    if (size < 2) return this
    val pivot = this[randomInt(size - 1)]
    
    val less = filter { it < pivot }.toTypedArray().quickSort() // recursive call
    val equal = filter { it == pivot }.toTypedArray()
    val greater = filter { it > pivot }.toTypedArray().quickSort() // recursive call
    
    return less + equal + greater
}


// Start of the program
fun main() {
   val x: Array<Int> = Array(100) { randomInt(100) } // <-- O(n) - Fill Array `x` with 100 random integers
   
   val a: Int = x[50] // <-- O(1) - Simple Lookup
   
   for (i in 0 until x.size) { // <-- O(n) - Single Loop over items 1 time
       println(x[i])
   }
   
   for (i in 0 until x.size) { // <-- O(n^2) - Double Loop over items (squared)
       for (j in 0 until x.size) { // <-- O(n) 
           println(x[i] + x[j])
       }
   }
   
   // Notice this one will take MUCH, MUCH longer than the previous ones.
   for (i in 0 until x.size) { // <-- O(n^3)- Triple Loop over items (cubed)
       for (j in 0 until x.size) { // <-- O(n^2)
           for (k in 0 until x.size) { // <-- O(n) 
               println(x[i] + x[j] + x[k])
           }
       }
   }
   
   val sorted: Array<Int> = x.quickSort() // <-- O(n log n) - Fastest known sorting algorithm for general case.
   val y = sorted.binarySearch(50) // <-- O(log n) - Searches a sorted array for a value, halving the search space each loop.
}

// << NO OUTPUT - Just code illustrating the time complexity of the operations >> 
// << - Run the live code example to generate output >>

      Live Code Example: Time Complexity Example

Communicating Software Designs Visually

communicating-software-designs

Some Thoughts on Common Software Production Methodologies

some-thoughts-on-common-software-production-methodologies

  • BIG IDEA — Are there ways to organize the work of discovering solutions and designing software that are better than others?

  • Waterfall

    • Was never a thing, it was a straw-man argument. Winston Royce, the creator the term "Waterfall," made it clear on page 2 of his famous 1970 paper "Managing the Development of Large Software Systems" that he was not advocating for the "Waterfall" process, but was trying to make a point about how software development was being managed like a manufacturing process, and it is a counter-productive way to manage the knowledge work.
      • It's Time For Waterfall To DIE
      • Managers never read the second page, and took it as a "best practice". Poor Winston Royce.
        • Note to self: Always put the most important information on the first page.

  • Scrum

  • Agile

  • SAFe

    • Another overcomplicated process that tries to "productionize" the software development process.
    • It's a "one-size-fits-all" process that was supposed to work for all kinds of software development, and it doesn't.

  • I call all of the above (and supporting hyped-books and hyped-conferences) the "Age of Conference©®™ Driven Development"

    • The problems around planning and organizing ever-larger teams, and increasingly complex problems became so onerous that many people looked far and wide for ANY solution to gain a foothold on the process and make it more predictable and manageable. Unfotunately, knowledge work is FUNDAMENTALLY risky and not possible to plan because its a process of DISCOVERY, not a process of manufacturing.

    • The biggest risk to manage in software development is that WE MAY NEVER FIND a workable solution in the time allotted. This is not changeable. It's a fundamental limitation of all knowledge work.

    • Dealing with the principal risks involved with knowledge work is the core problem in all software development.

  • The Problem with Applying "Production Management" Methodologies to Software Development Knowledge Work

    • The main issues with all these approaches is that they assume that software development is just another manufacturing process, and that the people who are doing the work are just another kind of "resource" that can be managed like a replaceable manufactuing worker on an assembly line.

      • Software development is fundamentally knowledge work and a process of discovery. The people who are doing the work are not interchangeable cogs but "knowledge workers" who are doing creative work that requires a lot of thinking and discussion and failed attempts before success, and does not have predictable outcomes.

      • The "Agile Manifesto" was written to address this problem, and to help people understand that software development is a fundamentally different kind of work than manufacturing, and that the people who are doing the work are fundamentally different kinds of workers than the people who work in manufacturing. This was a good thing.

      • The problem is that the Agile movement was soon co-opted into the "Agile® Certification Business" and turned into a "one-size-fits-all" process that was supposed to work for all kinds of software development, and it doesn't. It's a cargo cult that does not work for most teams, as most teams implement the "Rules™®©" directly without understanding the spirit of the agile rules, which there are no rules... just some values:

      • The 4 Values of the Agile Manifesto

        • The Agile Manifesto was written to address the problem of trying to manage software development like a manufacturing process, and to help people understand that software development is fundamentally different than manufacturing.

        • The 4 Values of the Agile Manifesto are:

          1. Individuals and interactions over processes and tools.
          2. Working software over comprehensive documentation.
          3. Customer collaboration over contract negotiation.
          4. Responding to change over following a plan.

        • While we value the items on the right, we value the items on the left more.

      • We Live Inside Our Experiences and Beliefs, Which Build Our Reality

      • The fundamental idea of Agile is "RESPONDING to CHANGE."

        • How can we make software development more flexible and faster to respond to changes and refinements in the requirements of software?

        • Cumbersome processes allow managers to "monitor" using charts and graphs is a standard way for businesses to manage risk in a manufacturing environment, but it's not a good way to manage risk for knowledge work.

        • Agile was a response to the progression towards more and more cumbersome processes that were driving projects to a standstill. It was a way to get back to the "heart" of software development, which is "respond to change."

        • It's primary result is working in much smaller steps, and being able to change direction quickly and easily based on what is discovered taking those small steps.

        • Each small step creates some kind of "valuable working software" that can be shown to the customer, and the customer can then give feedback on the working software, and the software can be changed based on that feedback.

          Key finding of the Agile Manifesto

          • Humans rarely know what they want UNTIL they experience what they don't want, the solution is to work iteratively and get feedback often, integrate that feedback to make sure we are building the best known solution at the time.

  • Continue Reading - Class Oriented Programming

  • Back to Index