Overview

Open Source software is found everywhere
- It is world changing technology
- Backbone of Internet servers
- Found in nearly every consumer device
Open source software is software still
- Failure reasons are similar to proprietary software
  - Vague requirements
  - Shifting scope
  - Ignoring user feedback
    - DO NOT do this!!!
- Fails also for other reasons
  - Onboarding new devs
    - ... have lots of questions
    - ... doesn't provide immediate value
    - ... have to train devs on your culture
  - Presentation and packaging
    - Dev doco must be awesome
    - Your web page cannot suck
    - Easy to get, build and run
  - Marketing woes
    - Coders may not be good UI folks
    - Webmasters and presentation needs
  - Management issues
    - Same tactics don't always work in open source as they do internally
    - An already independent group (developers) needs cohesion with strong project management
      - Should be behind the scenes
  - Cultural
    - Geographic and simple cultural differences
    - Code of Conduct for proj is ESSENTIAL
Emphasize again
- It is a lot of work to open up
- It is hard to open closed software
- Benefits take time to show up
- These costs are up front
Benefits are plentiful
- "Free labor"
- Peer recognition
- Resume booster
- Exposure to outside ideas (this is huge)

History

Originally, hardware was the differentiator for tech businesses
- Software was extended by users
- Changes shared with mfg and other users
Major differences from today: *Architecture was not generic
- What you learn and do on platform A was not portable to platform B
- Your skills were tied to the platform
- Sharing software meant more lock-in
- No Internet!
  - Hard to distribute data
- Software wasn't a selling point
Big industry shifts
- A few architectures won out
- "High level languages"
  - Write once, compile the same code for different architectures
  - If written for machine A, it could now work on machine B
- HW performance levelled out (less of a differentiator)
  - Margins dropped
  - Required mfgs to find new sources of revenue
Results:
- SW became a value add for the platform
  - Sharing SW cuts the value
  - Competitors could also get the SW, too
  - Tgis must be stopped!
- Access to code was restricted
  - Legal agreements (NDA) if code was available
  - Sometimes also required for binaries

Several separate people/groups had a similar idea
- But with different motivations
Richard Stallman
- AI lab at MIT
  - Culture of SW sharing at MIT
    - Common to academia
    - Similar to sharing research
  - Co-workers hired away by a proprietary SW company
  - New HW w/proprietary OS went to the lab
- Resigned and started FSF
  - GPL license
    - Code can be copied and modified
    - Derivatives must be GPL
      - (Viral)
    - Must always be free
    - Uses legal protections opposite of traditional use
      - Instead of enforcing restrictions, it enforces freedoms
  - GNU project started
    - Free OS
    - Free utilities
    - Utilities were so good they started being used on non-free OSs
    - 1990 most of OS done, but kernel still far away
  - Linus Torvalds and collaborators...
    - Linux - the missing kernel
  - FSF Rhetoric
    - Tied politics to code
    - Made it a cause, not convenience
    - Idealism motivation
GNU Not the only free OS project
- BSD
  - UCB
  - Already underway in 1990
  - Reimplemented UNIX (versus being a new system altogether)
  - Non idealism motivation
- X Windows
  - MIT - partnered with vendors
  - 1980s
  - Unique license
    - Free core
    - Proprietary stuff can be add on
    - Level playing field
      - Not idealogic
      - More for standards
      - Business motivation
- TeX
  - Donald Knuth
  - Code can be modified, but can't be called Tex unless it passed a compatibility test
  - Trademark protection motivation

Ideology
Technical excitement
Collaboration with others
The code is generally good
- Comparable to proprietary
- Sometimes better
- Cheaper, at least