The Grace Hopper Celebration of Women in Technology is an annual conference that draws women from around the world. There are lots of students, as well as women from industry, academia, non-profits, educators, and beyond (there are also a few men).
It's named after Grace Hopper, who was a truly amazing woman in technology. Among the things she is known for are: the language COBOL, standardization of programming languages, popularization of the term "debugging" and her visual aid for nanoseconds. She also is credited with the saying that it is easier to beg forgiveness than to ask for permission. Here are some links to video clips of her: part 1, and part 2.
This is long because it contains all my notes (and also because my days were packed), but you can always skip those (I'll put any notes I took during the sessions in blue to make them easy to spot). Any mistakes in them are because it can be hard to take notes and pay attention at the same time (also: hard to take notes on a borrowed iPad that I haven't gotten good at using the keyboard on; also, hard to write coherently on the plane after not enough hours of sleep.). If you're interested in more notes (and seeing what was going on during all those sesions I didn't go to), there is an official wiki. It's not complete yet, but hopefully will be soon. Also soon, the Anita Borg Institute should be putting up videos on its youtube channel. For the perspective of another grad student from this department, Emily Jacobson has notes up as well -- including for some sessions that I didn't attend.
If you have comments/questions/found an error/want to know how to come next year, feel free to email me (at lena at cs.wisc.edu).
This year the conference was in Portland, Oregon. So on Tuesday, I got up bright and early to catch the 7 am plane. The conference itself didn't start until Wednesday morning, but we wanted to be as available as possible for Hopper shifts (more on that later). Because of the time difference, my friend and I arrived early in the afternoon, and went walking around the city to see what we could see. We were particularly impressed by MAX, the public transit. That evening, our other two roommates at the hotel arrived and we went out to dinner. And then, we went to bed so we'd be awake and cheerful for the conference!
5:30 am: Wake up, because our brains are on central time still. But that's actually fine, because our first event is at the conference center at 7:00 am. We catch the light rail there (it's even free!).
7:00 am: Hopper meeting! You might be wondering what a Hopper is. There's this great system where you can sign up to volunteer (be a Hopper) at the conference for 8 hours, and in exchange your conference registration fee is waived. It's also a great way to meet people and actually pretty fun. Plus, you usually get to wear a stylish red vest. The meeting just gave us the basic info we needed to do our shifts.
7:45 am: Registration! When you go to GHC, you get a nice bag of free stuff from sponsoring companies. The contents vary widely from year to year. Some things from this year: a metal water bottle, lots and lots of pens, a notebook, sticky notes, a bunch of microfiber cloths (for cleaning touchscreens or eyeglasses), some mirrors, a purse hook (?), a folding hairbrush, a set of nail clippers and mini scissors, a bottle opener, and probably a bunch of other things I'm forgetting. Why so many mirrors? I don't know. While I was briefly annoyed at the stereotyping (women like mirrors?) I then realized that mirrors are actually useful for a variety of things, like seeing around corners. I guess. (But next year, I wouldn't mind fewer mirrors!)
8:30 am: Breakfast. Sitting on the floor of a hallway and talking with other students (even a Canadian) about what being a professor is probably like and the differences between our schools.
9:00 am: The welcome! Lots of really smart women in a big room. I foolishly didn't take notes from it but I think a lot of it was administrative.
9:30 am: PhD new investigators sessions. There were several choices, but I went to one on hardware and security. The presenters were doing some really cool work!
Samira Khan: dead blocks, 86% are dead in LLC
Virtual victim cache - put victim blocks in dead blocks
Predict which blocks are dead. Behavior is almost uniform across sets, sample sets to predict
Change insertion policy to reduce dead blocks. How?
Should read her papers
Maritza Johnson: access policy management
Facebook privacy settings. Do settings match intentions? It turns out, it often doesn't, even among fairly tech-savvy users.
Ting Fang Yen: detecting malware using behavioral features in network traffic
centralized, decentralized nets
Previous work relied on signatures or specific behaviors
Infected hosts show similar behavior, which is different than for uninfected.
Clustering. Using past traffic as a baseline.
Hosts getting a lot of similar communication content
Look at platforms, bots tend to be platform dependent.
Problems: encrypt communication, application dependent rather than host, p2p network
Tell apart file sharers and bots. By traffic volume, peer churn, and human vs. machine driven (based on time between events)
10:30 am: Merchandise! Last years Hoppers got an item free. Not this year, though, so I bought a stylish shirt myself. Also, I took the chance to grab some lunch, since I knew I wouldn't get a chance later.
We also took a photo! I may put up more later once I get permission from all people in the pictures, but for now, here is one with me and Emily Jacobson. (Full disclosure -- this was probably on Wednesday, not Tuesday. It's staying here for now anyway.)
11:00 am: More PhD new investigators! I accidentally came in late this time (took too long acquiring and eating lunch). The session I chose was on software. I missed most of the first talk, but it was on collaboration platforms. Here are my notes for the second two, including one by our very own Cindy Rubio Gonzalez:
Gabrielle Anderson: dynamic software update
Formal analyses lets you model things and maybe prove something? But it is simplified
When to do an update without corrupting any data
Bugs: dropped unhandled errors
Defective error/pointer interactions
Undocumented error codes
Used in NASA/jpl mars rover code to find a bug!
12:00: Hopper shift number 1. I showed up wondering what I'd be doing for the next five hours. Turned out it was a seminar on leadership, which is not something I would have thought of attending (if I hadn't been Hoppering, I would have gone to the CRA-W grad track. They have sessions for grads, undergrads, and faculty every year at both GHC and other conference such as Grad Cohort. On the other hand, I'd already been to the CRA-W sessions twice so it was nice seeing something new!).
At the session, I helped check RSVPs at the door. We had 800 women registered, but some didn't show up; there's so much going on that sometimes people don't make it and other people can attend instead. At the session we learned from Jo Miller about building a personal brand. Even though we were on duty, we still got to sit and listen, which was pretty nice. I got a chance to talk to some wonderful women, including some undergrads and grad students, and briefly to Carol Eidt from Microsoft, who was very interesting. Then we had speed networking practice. I was herding people rather than participating, but I enjoyed the energy and I did talk to the other Hoppers.
5:00 pm: Became overwhelmed (I am definitely an introvert) and went to hide in a hallway and write emails.
6:00 pm: I was supposed to go to a meetup, but it was in another building and I wasn't sure about my ability not to get lost. Continued writing emails.
6:30 pm: Wandered around the career fair. There were a number of companies, many with free stuff. I probably should have done interviews with them (and gotten said free stuff), but it was crowded and I felt a little weird since most of them were not really places I wanted to intern -- they were for the most part looking for programmers and software engineers. I could have used the practice, though. If you're at all looking for internships or employment, by the way, GHC is a great place to do it. Lots of opportunities for interviews and finding out who's hiring. I did eat dinner there (boxed dinner but still tasty) and talk to various people.
9:00 pm or so: Headed back to the hotel to sleep and prepare for the next long day.
6:30 am: Time to get up!
7:30 am: Yahoo breakfast. We got t-shirts, which made one of my friends very happy (she'd been wanting one for a year now. "We code like girls"!). I met some nice ladies from the other UW, and chatted with them. I also talked a little with the recruiters, and while they were doing some cool stuff with cloud computing, they were not so much into computer architecture. But they were still interesting to talk to. They had a drawing for an iPad2 and some sweatshirts. I didn't win, especially not the iPad.
8:30 am: Welcome and keynote. The speaker was Sheryl Sandberg, COO of Facebook and according to Forbes, 5th most powerful woman in the world. Found out there were 2908 people attending the conference, most of them in the same room for the talk, and most of them women. Wow! And I'm sure there were many others watching the live stream online.
Her talk was interesting, although there were things I didn't entirely agree with. Things I liked:
I didn't like as much the comments about the ambition gap in general or that the glass ceiling is gone. Just because a few people (including her) have broken through doesn't mean it no longer exists. I also get grumpy at generalizations, like that you have to train your husband to support you or that all men act in some certain ways.
All in all though I thought it was a pretty good speech. You can watch it online.
10:00 am: Sessions! As with most of GHC, this is pretty much choose your own adventure. There are 5 or 10 interesting things going on all at the same time, and you get to choose which you'd be saddest to miss. Since recently I've become quite interested in outreach, I attended "What if every public school student learned computer science?" I managed to delete most of my notes on it by accident, but here's what I do have:
There are 4 arguments why people should know CS.
Democratic: it's necessary to be an active participant in society
Economic: so that they can do jobs
Instrumental: allows us to do other things, like apply it to other fields
Cultural: the knowledge is part of our culture that defines us
The session was pretty convincing, though it probably helped that I already thought CS should be taught in the public schools. They talked about both benefits and challenges. I particularly liked that the potential curricula focus on more than just programming, but instead introduce students to all sorts of computer science concepts.
11:30: "The CS10K Project"
More education! The goal of the project is to get a new CS curriculum taught in ten thousand schools by the end of 2015. Ambitious! They're running pilots of classes in college classes and having lots of success. I wonder if we can adapt some of what they're doing to outreach?
AP principles class
CS AP test has biggest gender gap of all ap tests, at 19% women
Passion beauty joy awe of CS
Lightbot to teach programming
12:30: Lunch! I attended the LGBT lunch. It was quite an experience! I met women from Norway and Kentucky and Portland and Ohio and I forget where else at my table. I found out that there are rankings of tech companies by friendliness, so I will definitely have to take a look. It also sounded like a lot are friendly though, which is very encouraging. I wore the LGBT ribbon on my badge after that, because there were some good points about visibility. Also, I got an umbrella and a hat. Both people next to me at my table won the book from the It Gets Better project, but I didn't win anything. Not even an iPad. Again.
2:30: Ran into the undergrad from UW and talked about interviews. Also, data structures. Remembered that I like binary trees.
3:45: What if... there were more women in technology? The
business case for diversity
I took lots of notes in this session. It was a popular one! (Hopefully I didn't mangle anyone's name too badly, or attribute other people's comments to them).
Diverse teams are better for innovation, while routine tasks maybe not
Only 9% of patents from women or teams with women. I don't know if this is just for tech patents
50/50 team is good gender composition for a team
Women are especially good at making sure everyone's ideas are heard
Facebook (Katie?) privacy settings. Women feeling physically in danger on street, etc. need to have privacy settings.
Intuit (Tayloe) women control the finances of the world. Mint is geared towards women.
Google (Alan) Products for blind users. People couldn't use the apps, etc, so it was a problem when schools were switching to google apps. The reason they could fix things so quickly was that they had blind engineers already working for them.
Cisco(Kathy) workplace is better for women when it is better for all kinds of people. Telepresence came out with good quality, sound. Deaf interns fixed problem where it didn't switch to the person speaking without audible speech.
Symantec (Mark) At IBM, Frequent crises in hardware team (men), ended up pulling all nighters, would get awards for being heroes. Software team (which had some women) worked nine to five and dependably got things done. Women didn't worry so much about being heroes, just wanted to get things done.
Unconscious bias despite good intentions. How to measure success of organization in creating a diverse workplace?
Statistics are easy to come up with, but results are the real test. Don't just track surface changes.
Interviews are an unreliable predictor of how well people will do. Internships are better.
Referrals. Most referrals were from male engineers and were men, lots got hired.
Added policy to interview any woman in the stack of resumes, and hiring went up.
Maybe women came across better in interview than on paper.
Patents with diverse teams give better business results.
Supporting other diversity. Example is networks with company level support.
Changing culture to be more functional aka not all nighters. First have to stand up for yourself such as for "mom time". Be willing to say no. If you are in a position of responsibility, do things like cancel meetings to go to parent-teacher conferences and tell everyone. Show what the priorities are.
Not graduating enough cs students of any sort. Increase the pipeline. Google programs for high school students and teachers. Don't require programming experience before college (how CMU increased their percentage of women).
Make it cool to be a computer scientist. Post pictures of beautiful computer scientists. Most girls have never met or talked to or asked a question to a single female computer scientist.
(about many women) Every time you're talking and someone cuts in, you just stop talking.
5:15: Birds of a feather session. Betting on yourself: Regaining and maintaining confidence as a woman in technology.
I needed this talk! After I left it everything seemed worlds better about my research, my classes, my life, and myself.
Surround self with peer support, environments that build confidence
Establish reachable goals. Start a spreadsheet and measure it.
Dress for success. As in, if you start not taking care of yourself, it's probably because you are losing confidence, and maybe you can fake it until you make it. Also it might make people treat you more like an adult than as an undergrad, especially if you're the youngest woman in the room.
Talk to yourself in the mirror in a boastful way with a weird accent. Like Arnold Schwarzenegger.
Faking it till you make it
Even Robin Jeffries spent a couple years panic stricken every day over her job.
Tooting your own horn, find a way to talk about yourself.
The art of moderation: know when to say no. Take care of your mind, body, and spirit.
Defining and living your values. What brings you joy?
Other people also wonder why they're in grad school, and find grad school painful.
Women are damned if they do and damned if they don't . If you aren't pushy, you don't get what you want. If you are, people call you a bitch, and you maybe still don't get what you want.
Put things in your office that show how awesome you are to remind yourself. Unfortunately that depends on having the awesome things to start with.
Lots of undergrad women doing really well come to office hours asking to drop because they think the men are doing better.
You can feel scared, angry, sad and still make progress.
People who started over at their fourth year and still are successes.
Men are sometimes more socialized to take risks. But we're in a world of innovation and we should figure out where we at comfortable taking risks, starting with small ones.
Don't be so scared of answering wrong that you just don't answer. Take a risk and correct yourself later.
Other: the number of women in the room makes me suspect that I'm not that only one that feels like a failure or not smart enough.
6:30: Dinner and poster session
The nice thing about the poster session is that you get to meet some really awesome people! And hear about all kinds of research. I talked to someone about binary translation for automatic parallelization, and someone else about flash, and intel about their parallelization tools. And some various other people as well. I finished right at 9:00, which was also when the session got over, and headed back to the hotel.
The day I got to sleep in (until 7 am)
8:00 am: Hopper shift! At the general session and keynote. The highlights of Hoppering were directing people towards the bathrooms, as well as standing next to the microphone in a room with a couple thousand people and flailing my arms wildly.
The keynote was by Shirley Ann Jackson. Since I was Hoppering during it I didn't have a chance to take notes, unfortunately. I will post a link to it if I see one.
10:00 am: My Dance with Research session
11:30 am: Addressing heterogeneous programming challenges using PEPPHER
Portability: take software that runs well and port it to another system
APU: means fusion architecture
Aim: application runs on any hardware
Hot spots: performance critical sections of program
Metadata for predicting execution time of a component invocation
Can choose whether to try for execution time or energy consumption, for example
Portable compilation techniques. Source to source transformation
Adaptable algorithms and data structures
Pre-defined structures and methods. Such as stack, queue, linked list
Also SkePU skeleton c++ template library
Flexible runtime system. Takes tasks and schedules them across multiple architectures. Open scheduling platform, to use different scheduling algorithms
Hardware support mechanisms
What segments to peppherize?
Ones that are a large fraction of execution time
Ones that have a variant on the other architecture or that follow patterns (like sorts)
Can be used for pipeline models like bzip2.
Take sequential c++ with (TBB?) annotations, outputs pipelined code with source to source transformation
Almost as good performance on bzip2 as in pbzip2 ( which took 2 years) or tbb (which took a month), but took only four hours
Static and dynamic performance targeting
Does not parallelize code, but rather lets maintain performance across different architectures
12:30: Systers lunch
Lunch with a bunch of Systers. For those of you who don't know, Systers is a mailing list for women in computer science. It is a great resource and I encourage people to check it out! It was great to put a face to some of the names.
2:00 pm: Plenary panel
I was a bit overwhelmed/exhausted by 2:00pm (and annoyed that it wasn't an imposter syndrome plenary, possibly), so I kind of didn't pay as much attention as I should have. I apparently took some notes though.
Mentors and sponsors
Amplifying one another's ideas
Being clear and explicit about goals, what you want to achieve
Plan your goals, then talk to people about them
Everyone makes mistakes
3:45 pm: Following the non-academic track as a PhD
Benefits of PhD in entering the field:
Umit: 2 kinds of jobs you can get. Those where a PhD is a must, such as those in research labs. Applied research in an industry setting. Or jobs in project development or consulting. For those, in many companies a masters is fine.
getting a PhD demonstrates a level of perserverence.
A title which carries respect in some circles
Demonstrates an ability to think critically
Identify problems in well-established and emerging fields
Critical eye, perserverence, ways of thinking
Going deep into a problem, fixing it, but also being able to step back and see context and the significance
Being able to communicate in both the high and low level
Thinking in tees: very broad, very deep. Maybe also a circle around it.
Take very complex ideas and simplify them
Be a good communicator, good simplifier, a big picture person, enthusiastic and persistent
Not only getting the content right, but also the details. Expressing yourself well, being a good communicator.
Being able to accept critiques and criticisms
Whether you can write real code and not just toy problems
Hired despite having a PhD and not because of it
PhDs live in weird worlds not the real world, don't know about deadlines or regular working hours
Academics "assume away" problems, but you can't do that in the real world
Book educated but don't understand applicability to the real world
Gain credibility by doing grunt work like other engineers
No decision is ever made in stone. You can always change your mind.
Learning agility: how fast can you take an idea and put it into practice
5:15 pm: Computational systems panel
Department of Energy
Carol Woodward, LLNL
Aerodynamic drag reduction for heavy vehicles
Class 8 tractor-trailers responsible for 12-13% of total petrol consumption.
They lose a lot to drag.
Core collapse supernova modeling
Massive problems that are really really hard
Need to have lots of really tiny steps
Have to be very precise because 99% of energy is carried off as neutrinos and only 1% observed
Uncertainty-quantified climate system models
Hydrologic cycle usually simulated with disconnected processes each run separately
Trying to couple the models more tightly
Accurate Coupled simulations critical to wind power prediction and use as well as water resources management
The moisture in the soil actually has a big effect on low-level winds, like the ones that turn the wind turbines
1 acre of corn transpires 4000 gallons of water a day!
New Moore's law: number of cores per chip doubles every 18 months instead of the clock frequency
System power is a first class constraint
Programming models, OS design, memory, reliability, resiliency
Janine Bennett, Sandia
combustion research. Combustion is a big source of energy
In situ data analyses, because there are terabytes of data
Analysis will need to be in lockstep with the simulation
Structural simulation toolkit, to identify causal relationships and test out new ideas
Graphs are everywhere! Social networks, power distribution, etc.
6:30 pm: Award Ceremony
I didn't take notes for the award ceremony (because I was eating m&m's...). But they were very inspiring, even if I can't remember the details well (it was a long day). Anne Ikiara was particularly inspiring. She founded Nairobits, an organization for teaching youth and women in Nairobi about web development, allowing them to improve their lives and their wider community. The idea is spreading and there are now organizations in multiple countries in Africa.
Another of the award winners was Judith Owigar, who founded Akirachix, an organization to empower African women in technology. I was also impressed with Marita Cheng, who founded Robogals, a student organization that puts on workshops for middle and high schoolers. It's spread from Australia to several other countries.
7:30 pm: Sponsor night!
On sponsor night, the big sponsors of the conference (Google and Microsoft) throw a big party. A DANCE party. This year, we all left the ballroom where the awards ceremony was held to wander over to the party ballroom. I was on the far side of the room and we looked over out the door, and my neighbor said "is that a marching band?" It was. We followed it, in a huge throng of people. They were The Beat Goes On Marching Band (Thanks to their music director (and many-time GHC attendee, with Intel) Steve Tolopka, for emailing to tell me the name!) They played YMCA and several songs by Lady Gaga, among other. They also had a color guard with flags and everything. People were definitely grinning.
Once we got to the ballroom, we headed straight for the shirt tables to pick up our stylish microsoft and google tshirts. Google was also giving out flashing wristbands. I actually got one this year, unlike last year when they ran out early.
Standing next to the google table with Amanda Hittson, Wei Wen, and Emily Jacobson:
We had dinner with some women from Oregon and at least one from Colby College. Then we stood around watching the dancing for a while before deciding to join in. While dancing, I ran into my roommate from CRA-W Grad Cohort 2010, which was fun.
If you don't know me, I'm incredibly awkward. I've been dragged out on a dance floor, briefly, about 5 times in my life. This time, I went willingly and (awkwardly) danced the night away. It was great fun dancing with all sorts of brilliant women (and a couple men), including the award winners and speakers and ABI people. It was also pretty exhausting after a while! But I'm going to remember for the rest of the year, whenever I'm feeling like a failure or like I'm too self conscious, that I danced (terribly but enthusiastically) in front of a bunch of potential future employers and coworkers. (That said, I might be hoping that ABI does not use any images/video of me dancing in their promotions for next year, or people might be scared away -- I am not even kidding about my lack of dancing ability).
Sadly, we eventually had to head home. While waiting for MAX, I overheard someone say how wonderful it was to see so many introverts dancing and just having a great time.
Time to head home! But even heading through the security line at the airport, things went more smoothly for me than usual when the employees noticed my "Geek Girl" Microsoft t-shirt -- one that a significant number of our fellow travellers were wearing. And it was great seeing just how many people were taking out the water bottles.
I can't wait to go back next year!
If you've gotten this far, you've probably been able to tell that I had a great time and came away feeling refreshed and excited. But there are other reasons to go as well: networking, the career fair, learning about interesting research, getting ideas for outreach, panels on being successful in undergrad, grad school, and the rest of our lives.
If you're worried about affordability, there are several things to keep in mind.
First, you should apply for the scholarship in May. If you have never attended before, you have a better chance of being awarded one, and it's definitely worth a shot (and it looks good, besides).
If you don't get the scholarship, the next step is to apply to be a Hopper. Assignment is done based on availability to work and also a lottery, so you probably want to check off as many times as available to work as possible. So far, we've had 100% success rates (or close to that, at lesat) at getting people Hopper positions. A Hopper position means you don't have to pay registration, so that eliminates a big chunk of the cost.
Finally, if you're a student here at UW-Madison, WACM has obtained funding from the department for everyone who attended for the past two years and hopes to continue at least partially funding students. In fact, there have been so few people attending GHC that WACM has been able to allocate more funding than promised to each student. Obviously, this might not be the case in the future, but it is likely that there will be some funding. If you are from another college/university and just stumbled upon my page, you might want to talk to your women's group (if you have one) or to your department about funding you -- many places fund students for GHC.
WACM has plans to have several information sessions about GHC, to encourage people to attend next year. Keep an eye out for those. And as previously mentioned, feel free to email me (at lena at cs.wisc.edu) for more information -- I can also put you in touch with other students who have attended this year or in the past. We'd particularly like to see more undergrads attending.