CSC 321.01, Class 11: Modeling basics

Overview

• Preliminaries
  • Notes and news
  • Upcoming work
  • Questions

News / Etc.

• Sample departmental schedule should be distributed later today.

Upcoming work

• Reading Journal due Friday.

Good things to do

Nope, no extra credit.

• Thursday extras, Thursday, 16 Feb 2017, 4:15 p.m., Science 3821: 4-1 Program with UIowa.
• CS Table, Tuesday, 21 Feb 2017, noon: Net neutrality.
• Senior day for basketball on Saturday (women’s and men’s).
• Swimming and diving this weekend.
• Orchestra Saturday.
• Slavic coffee house, Saturday, 5:30-7:00 p.m. Bucksbaum Rotunda.

Questions

Modeling Data

• Almost every non-trivial application will have different kinds of data with which you will work.
• As a designer, you have a responsibility to think about the different kinds of data you have, what values are associated with each kind of data, and how the values relate.
• Although agile may seem to suggest that you can clean up from your choices, some early choices can have very negative implications.
  • Example: Head start phone notification
  • Models: Students, Parents, each student has exactly one parent; each parent has exactly one student.
    • Data duplication!
    • Not possible to do some things
• Getting models somewhat right early on is important
• Model:
  • List of information to store for each “thing” of a particular type
    • Post:
      • ID: identifier
      • Date/Time posted: datetime
      • Title: string
      • Contents: string
      • Tags: ???
      • Writer: ???
    • Writer
      • ID: identifier
      • Name:
        • Option 1: string
        • Option 2: first name string, last name string
          • May have duplication
          • Cultural expectations
          • Family name and individual name
          • Not everyone has two names
          • Pre-marriage name and post-marriage name
        • Option 3: name datatype
      • Posts: ???
      • Fake biography: String
  • In some cases, we have a clear basic type (or complex type), in others, we have relationships to other objects
    • Each post has one writer; each writer may have many posts (one to many relationship)
    • Good modeling systems let you “rip out” those relationships, and express them separately, so that they can deal with them behind the scenes
    • Three general kinds of relationships
      • 1:1 relationships
        • ideally, we wouldn’t replicate names
        • each writer has a unique id, each unique id refers to one writer
      • 1:many relationships
        • One writer has many chirps
      • many:many relationships
        • each child may have multiple guardians, each guardian may have multiple children
        • each chirp can have many tags, each tag can be associated with many chirps

Doing modeling

• List the kinds of values you will deal with (e.g., students and guardians or chirps and beaks)
• For each kind of value, list the key fields (ignoring id and relationship)
• List relationships

Alumni.grinnell.edu

• Values
  • Users (alumni, current students, …) [0,5]
    • re’sume’s, type: data
  • Donations [1]
  • News stories [2]
  • Photographs [3]
  • Events [4]
• Relationships
  • E.g., Each user has many (or zero) donations

Task: Figure out the data you would associate with each person

Models

User

• name: name
• cv: data /* can also hold résum&ecute;s */
• home-phone-number
• work-phone-number
• mobile-phone-number
• facsimile-phone-number
• other-phone-number
• info-public?
• contact-me?
• home-email
• work-email
• location: address
• biography: string
• linkedin: URL
• Facelessbook: URL
• relationships
  • user:major is a many-many relationship
  • user:job is a one-many relationship
  • user:event is a many-many relationship, with annotations

Donation

• id: (implicit)
• amount: long;
• currency: string;
• anonymous?: boolean - a signal that we are not supposed to reveal the donor’s name
• recipient: string
• additional-notes: string

Detour: Currency

• name: string
• conversion rate to US dollars: ratio of two unsigned integers
  • doubles round

News story

Photograph

• photographer: (either with a relationship to an author model or name)
• datetime: datetime
• caption: string
• description: string
• image size
• image type
• image resolution
• relationships
  • 1 event, many photographs
  • many-many: photograph tag

Event

Relationships