GHC 2013: Day Two

Vint Cerf, President of the Association for Computing Machinery (ACM) welcomed us with a video since he couldn’t be here in person. He encouraged us to join ACM-W, ACM’s Women in Computing group and welcomed us to the second day of GHC. He talked about how it was a “waste of talent to not have more women in CS”. He also had a message for his male colleagues: he is “very disappointed that some women have been treated badly in the computing professions” and he hopes that “every single person that has the opportunity to stand up for women will do so”.

The stage in the main ballroom was pretty cool.

The stage in the main ballroom was pretty cool.

In Effort There is Joy

The keynote this morning was Megan Smith, VP at Google[x]. Google[x] is the branch at Google that works on experimental and secret projects, like self-driving cars and Google Glass. She talked a lot about finding your passion and getting to feel the power of what it is to invent and change things.

In her sophomore year, she was working at IBM on a rig where she used tools to discover how deep they were and what was in the soil there. She was the only female on site and the other riggers would stare at her. She decided that she was okay with that. There was some advantage in being the only female: after a while, people would come up and show her new and interesting things. She said that “having the courage to let people watch you is worth going through because on the other side, you get to have cool experiences.”

Megan had a lot of examples of diverse people making a difference. She talked about George Washington Carver, the scientist who brought diverse agriculture to the South in the early 20th Century; Catherine Swifter, the female runner who had her numbers ripped off during a marathon; and the group that cracked the Enigma machine during WW2 was composed of several women and Alan Turing.

I think the thing that I’ll take away the most from her talk is the idea of Moonshot Thinking. Moonshot thinking means that when you look at problem, you think about how to make it 10 times betters instead of just 10% better. I think trying to get yourself to think this way forces you to come up with novel approaches to problems.

Megan’s talk has been posted online.

Looking for Bugs in All the Right Places

Elaine Weyuker is a researcher and visiting scholar at Rutgers University. Her talk was about her research which is a statistical model to determine which files in a software system will contain the most bugs.

The tool that they built has two parts:

  1. Extracting data needed to make the predictions
  2. Analysis on data, making the predictions, and displaying them.

The information they used for the predictions was:

  • Size, larger files = more bugs
  • Recent changes, new code = more buggy. She also mentioned that looking at who changed a file was a poor predictor.
  • Past performance, if the file had more bugs than average in the last release, it will be so for the new release.
  • Programming language, some languages tended to be more buggy

Her team worked with a huge system and looked at their version control and bug history. They were able to predict where the bugs will be in the next release very accurately.

Prediction results for 9 systems.

Prediction results for 9 systems.

I can see how this could be useful but I think there would be some barriers to adoption:

  • Let’s say you know that a certain file will contain a lot of bugs. You still do not know where these bugs are. I suppose that if the file is for a specific feature or functionality then you can re-test that part of your software more thoroughly.
  • The severity of bugs is not provided. There are bugs that we prioritize higher than others and some that we don’t really care about. If there were some way that the severity of bugs could be predicted, this tool would be more useful.
  • We are constantly building new features, which means new code and new files. Because this relies so much on historical data, I doubt it could be very accurate when comparing new files. So it would fail to give you a priority on what needs to get tested the most.

Gaming Lightning Talks

This panel was actually 4 talks that related to gaming in some way. The one that I enjoyed the most was by Shannon I. Steinfadt whose talk was titled “Gaming the System: Gamification for Nuclear and High-Hazard Response Training”. She talked about how she worked with a group that creates immersive 3D virtual training grounds for nuclear and high-hazard facilities.

From left to right: Gail Carmichael, Brittany Arthur, Rosa Thomas, and Shannon Steinfadt.

From left to right: Gail Carmichael, Brittany Arthur, Rosa Thomas, and Shannon Steinfadt.

Her work is especially appreciated in teams that only have access to facilities once every three years for training, and so the virtual environments help them become familiar with the layout of the buildings and where things are without actually being there.

An example of a group that they work with are firefighters. They also incorporate mobility limitations into the game based on what gear a firefighter would be wearing. With these types of “games”, firefighters are better able to retain what a building looks like. Training when on-site goes better and results in cost savings.

Another area where this kind of system is useful is when designing buildings. For security systems, having a 3D environment can help you understand where fences and guard towers will be needed. It can help you look at lighting and where to put security cameras by allowing you to see what the camera would see.

Systers

Overall, day two went well. During lunch, I saw with a group talking about Systers. Systers is an e-mail community composed of women in tech at various levels in their careers.

I did not attend the party on Thursday night but I heard it was a lot of fun and I really wish I had gone. I missed it because I was feeling a cold coming on and thought it would be better to rest. =(

Tagged ,

Leave a Reply

Your email address will not be published. Required fields are marked *