sims' Blog to be rich is to remember

GSoC 2017 Postmortem Part 2: Org Application

In the beginning, there was an organization application. And with it, the need for someone responsible. That one was me. I always wanted to help with GSoC but felt that my coding skills and codebase knowledge were not good enough to be a mentor as good as I’d like to be. So I followed the call for main admin and on the story went.

Google Application Form

The most obvious piece of the application process is the application form every organization has to fill on the GSoC page. The problems I encountered while drafting the first version of our org profile and “Ten Reasons Why You Should Take coala”-text were essentially the same as for every application I had ever written. How does one sell themselves without making things up.

Although coala had two years of GSoC experience, those were as a sub org under the PSF umbrella. This year should be the first under our own flag, so everything had to be perfect. Again and again I asked for feedback on my drafts and input on things where I didn’t know enough about. Both Lasse and John were very helpful, both with plenty more GSoC experience than I have. We felt that we were in a pretty good position. Our community has a strong connection with the GSoC program, many of our long time members were students once upon a time. We had a total of 10 students over the two years with PSF, and from what I have heard, they were more than pleased with our performance as an organization, both the students and the PSF. The only thing that we felt went against us was us being a novice independent org, and us not having failed any of our students so far. Yup, this is a thing. From what I understood, Google expects around 20% failure rate across all organizations. If an org deviates from that too much for too long, then they are either too strict with their students, not strict enough, or there is another problem with their process. We actually were quite worried about this point but from what I have learned, there is no real reason to worry about this. Document what you and your students are doing and as long as you have good reasons for the decisions you are making, everything should be fine in the end. About the novice part, there wasn’t really anything we could do besides pointing to our past GSoC experience and hoping that Terri Oda, a PSF GSoC admin, would vouch for us.

Projects Page

Another, big, part of the application is the projects page. It is a list of project ideas that students can apply for. Although it is encouraged that students come up with their own project, most applications will come from the project list. And a nice project list is a must have, from what we were told, to get into GSoC. We felt that the project page should be the most well fleshed out part of our application, as it is actually useful during the following parts of the program and offers a lot of value to the students and us when the student application phase comes around. This lead to our web frontend hero Hemang building our new and awesome Project Page. Before this, we just used a GitHub wiki page with some project drafts but the new page was much easier to use, gave a better overview about the projects and it gives us the possibility to add features to it in the future. During this early phase, we included the old projects from the wiki and asked mentors and community members for other projects ideas they might be interested in mentoring. For those who haven’t read any of the project ideas on our page, the structure always is something like this: Introduction and motivation, sometimes followed by some additional information about how to solve the problem if we already had some idea about it. Then we show a time line following the GSoC 4 phase model, that holds some milestones and gives a rough idea about when to deliver the single parts. Finally, we link all additional information we have about the project, like issues or cEPs, add mentors as contact persons, and some more meta information. In addition to the project ideas the page also holds a faq section, where we started to collect questions we received again and again or just felt were important. We actually tried to let the student’s add the questions and answers to the faq themselves to include them into our process and also to make sure they actually understood our answer and thus to double check if our answer was good enough or if maybe we need a better explanation. We had close to twenty project idea drafts at this point and felt that we were more than ready to hand everything in to Google and be accepted. Before we move on to the next part, I want to point out that we also built the projects page so other orgs could use it and we have received some good feedback on it.

FOSSDEM and Feedback from Stephanie

In February, a few weeks before the application deadline, we went to FOSSDEM with coala. We had a stand where we promoted what we did but most important, our stand was right next to the GSoC stand. A great opportunity to get feedback and ask some questions we figured, and so I walked over in a calmer minute to show Stephanie what we had so far. Starting with the projects page, she liked it a lot. She liked the idea of it and how it was structured and could help to introduce students even easier. The only improvement she asked for with it were the project descriptions. Flesh out the motivation for most projects and just write a little more to make it easier for students was what I took away for the projects age. But overall this part gave me a very good feeling!

Then we talked about our chances to get into the program with the experience we had under our belt and our expectations and wishes for slots we would get. As I said before, we had 10 students over the last two seasons. First 2, then 8. We felt, that we could probably grow some more with the mentor power we had amassed but also didn’t expect too much as we had some challenges coming up, that we didn’t have the last two times. What I asked was probably something along the lines of “how high are our chances to get around 12 students this year with what I just showed you + improvements I’ll implement?”. And here is where the not-so-fun part started. She essentially looked at me as if I asked for a million dollars out of nowhere. Turns out, that first time organizations usually get 1 or 2 students. And even with our past experience as a sub org, we kind of counted as a first time org. I pointed out that we had 8 students under PSF last year and received great feedback from their admins for our work and Stephanie couldn’t believe that we did get such an enormous number of students as such a young org. I came out of that part of our conversation with the feeling that we’d probably not get more than 4 slots no matter what we did. Not a great takeaway.

Implementing Feedback

I rewrote a lot of our application texts to focus more on the things Stephanie pointed out as important factors after we returned but the main part of work I did was essentially rewriting all twenty something project proposals we had. Although I wrote some of them myself, most of them were actually written by students or mentors and I did all the review work. The reason or this was that even though I probably would have been faster just writing everything myself, I wouldn’t be able to get enough reviews to merge everything in time. So the solution was to poke students that were interested in the projects and mentors to rewrite it and me reviewing it.

During this time I also struggled with the expected slots we would get. While yes, we would finally be our own organization, with all the pros that would bring with it long term, it also felt like a step backwards to loose students slots compared to last year. But the fact that we couldn’t grow further under the PSF and had plans to act as an umbrella for smaller linter projects made the investment in our long term growth worth it in our eyes. I should probably add that we had more than 5 newcomers per week during this time that were interested in GSoC with us. This meant that the newcomers from one week might be more than we would get. And there were quite some students that came to us end of last year and had contributed a lot. Hard time deciding whom to take on the horizon already.

After completing the projects rewrite we called it done and handed everything in for the final time. Now all that was left was to wait and get to know the possible students.

Result Day, Slots and Thoughts

We were accepted. And although it doesn’t quite fit here chronologically, we received 10 slots. We were happy when we were accepted and a huge weight left my shoulders. But the real joy came when we received those 10 slots. Apparently Terri gave us one hell of a recommendation and all the mentors we had collected and project ideas we had prepared were worth something after all. John, one of the other mentors, also met Stephanie at FOSSASIA and kind of implied that getting only two slots wouldn’t work for us so maybe that also helped. Regardless of how we did it, yeay for the slots!

So what do I think about how everything went down? I feel like the projects page was a great idea and will serve us (and maybe other orgs) well in the future. There are some downsides to it in my opinion that I will cover in the next part more in depth but the main thing is that all those super defined projects left the students with nothing to think about in some cases. In the future, we will focus more on presenting the problem and why we’d like to fix it and less about the how to fix it in detail. This way we give students a chance to shine during their application phase by working on the specifics with the respective mentors and take some of the work off of us.

The part where we included students in writing the projects and faq entries on the other hand was great. It is definitely more work than just doing it yourself but it works as a first identifier for very good students. They will be happy to help defining projects, already starting to do research on it and discuss the details with possible mentors. Doing this in a more structured way will probably something we will try to do in future years.

About the meeting with Stephanie at FOSSDEM, I would recommend that to everyone! Stephanie is a lovely person and helped us a lot with her feedback I feel. She is also just nice to talk to :).

Overall I feel like this part of GSoC went pretty well and without real problems beside our panic of getting too few slots. Things won’t stay this rosy in the next phase sadly so if you want some drama and more interesting “things we could have done better” things, make sure to come back. I’m not sure when I will find the time to finish the next part as I have some university to catch up on and it takes me a good 2 hours to write one of these, at least so far, but I’ll try to keep at least a weekly schedule for this series until I am done.

Thanks again to Terri Oda from the PSF and Stephanie Taylor from Google for your support during this phase that definitely helped us to get into GSoC and learn a lot as an org.