The Hardest Problems in Computer Science

AlterConf Berlin 2017 (https://alterconf.com/speakers/anjana-vakil)

They say there are two hard problems in computer science: cache invalidation, naming, and off-by-one errors.

In this talk, we'll consider these problems in the context of diversity and inclusion in the CS/tech space, where they seem to get even harder:

- Cache invalidation: The assumptions we make can be extremely damaging, and need to be carefully monitored and easily rejected
- Naming: Names hold great power for individuals and communities, which means we have to take great responsibility for how we use and respond to them
- Off-by-one errors: Individuals from marginalized groups often find themselves the odd-ones-out in their tech team/community, which undermines alleged diversity goals

We'll explore these problems, and possible solutions, via recent studies, surveys, talks, and writings from the tech community. I'll share my perspective, as a female engineer with a humanities background, and you'll have time and space to share yours. Together, we'll get motivated and prepared to tackle these problems, individually and collectively, to make the tech world an easier place for everyone.

Transcript

1. > the hardest problems
   in computer science
   > @AnjanaVakil
   > AlterConf Berlin 2017
   

   View Slide

2. there are two hard things in
   computer science:
   

   View Slide

3. there are two hard things in
   computer science:
   [1] cache invalidation
   

   View Slide

4. there are two hard things in
   computer science:
   [1] cache invalidation
   [2] naming things
   

   View Slide

5. there are two hard things in
   computer science:
   [1] cache invalidation
   [2] naming things
   [0] off-by-one errors
   

   View Slide

6. there are two hard things in
   computer science:
   [1] cache invalidation
   [2] naming things
   [0] off-by-one errors
   - Phil Karlton via Tim Bray via Jeff Atwood
   https://twitter.com/codinghorror/status/506010907021828096
   

   View Slide

7. [1] cache invalidation
   > the problem
   

   View Slide

8. caching data improves performance
   but stale data must be removed
   from the cache (invalidated)
   

   View Slide

9. amazon.com/Thinking-Fast-Slow-Daniel-Kahneman/dp/0374533555
   we think with 2 systems:
   1: fast, intuitive,
   emotional
   2: slow, deliberative,
   logical
   - D. Kahneman, Thinking Fast and Slow
   

   View Slide

10. quick!
    picture a software engineer
    

    View Slide

11. - Ola Gasidlo, engineer at Mozilla
    https://twitter.com/defunkt/status/426104782894284800
    

    View Slide

12. [1] cache invalidation
    > solutions
    

    View Slide

13. “Instead of assuming what someone’s
    position or career is, ask, ask
    politely, and don’t act surprised
    at the answer.”
    - Sailor Mercury, “Coding like a girl”
    https://medium.com/@sailorhg/coding-like-a-girl-595b90791cce
    

    View Slide

14. “No feigning surprise”, from the Recurse Center social rules
    https://www.recurse.com/manual#sub-sec-social-rules
    

    View Slide

15. [1] cache invalidation
    > your thoughts?
    

    View Slide

16. [2] naming
    > the problem
    

    View Slide

17. 70%
    “of the source code of a software
    system consists of identifiers”
    - F. Deißenbock & M. Pizka, “Concise and Consistent Naming”
    via P. Hilton & F. Hermans, “What science says about naming”
    http://hilton.org.uk/blog/science-on-naming
    

    View Slide

18. “the names chosen as identifiers
    are of paramount importance”
    - F. Deißenbock & M. Pizka, “Concise and Consistent Naming”
    via P. Hilton & F. Hermans, “What science says about naming”
    http://hilton.org.uk/blog/science-on-naming
    

    View Slide

19. # what do these names assume?
    def foo():
    bar = “hello, world!”
    print(bar)
    

    View Slide

20. J. Terrell et al., “Gender differences and bias in open source: Pull
    request acceptance of women versus men”, via The Guardian
    https://www.theguardian.com/technology/2016/feb/12/women-considered-better-coders-hide-gender-github
    

    View Slide

21. [2] naming
    > solutions
    

    View Slide

22. # self-explanatory == inclusive
    def greet_user():
    greeting = “hello, world!”
    print(greeting)
    

    View Slide

23. “continuing to publicly present
    as a woman is important for
    solidarity”
    - developer/GitHub contributor Heather Booker,
    “I applied to Outreachy, and I didn’t use my name”
    heatherbooker.github.io/blog/code/life/2017/03/14/i-applied-to-outreachy-and-didnt-use-my-name.html
    

    View Slide

24. [2] naming
    > your thoughts?
    

    View Slide

25. [0] off-by-one errors
    > the problem
    

    View Slide

26. you want a fence 30 meters long
    with posts spaced 3 meters apart
    how many posts do you need?
    - the “fencepost error” via Wikipedia
    https://en.wikipedia.org/wiki/Off-by-one_error#Fencepost_error
    

    View Slide

27. you want an engineering team
    with “diverse” members
    how many do you need?
    - the “token hire error” via me, just now
    

    View Slide

28. - Kelly Ellis, engineering lead at Medium
    https://twitter.com/justkelly_ok/status/854030175410864128
    

    View Slide

29. Photo of Gloria Kimbwala via TechCrunch
    tctechcrunch2011.files.wordpress.com/2016/01/gloria_kimbwala.jpg
    “My path through technology and
    through computer science —
    I was always very aware I was
    the only minority and the only
    woman in all of my classes.”
    - Gloria Kimbwala, engineer at
    Square, via TechCrunch
    https://techcrunch.com/2016/01/12/meet-squares-gloria-kimbwala-an-engineer-
    dedicated-to-advancing-diversity-in-tech/
    

    View Slide

30. - Khalid, “Barriers to entry in the JavaScript
    community”, JSConf EU 2017
    http://2017.jsconf.eu/speakers/khalid-barriers-to-entry-in-javascript-communities.html
    

    View Slide

31. [0] off-by-one errors
    > solutions
    

    View Slide

32. - Amy Nguyen, engineer at Pinterest
    https://twitter.com/amyngyn/status/785532779106754560
    

    View Slide

33. “Hire more damn women” - Christina Sass, CEO of Andela, via TechCrunch
    techcrunch.com/2017/03/08/theres-a-simple-solution-to-techs-gender-imbalance-hire-more-damn-women/
    

    View Slide

34. [0] off-by-one errors
    > your thoughts?
    

    View Slide

35. > thank you!
    > @AnjanaVakil
    > AlterConf Berlin 2017
    

    View Slide