Summer of Code/Application/2014
Introduction & overview
This is the application of the ScummVM project for the Google Summer of Code. Let's start with a quick overview of the facts before turning to the more elaborate parts of this document.
| Project: | ScummVM |
| Participation in prior SoCs: | Yes |
| Applications for prior SoCs: | 2013, 2012, 2011, 2009, 2008, 2007 |
| Organization administrator: | strangerke |
| Organization co-administrator: | |
| Backup administrator: | wjpalenstijn |
| Project License: | GPLv2 |
| Ideas page: | GSoC Ideas |
| IRC channel: | #scummvm on irc.freenode.net |
| Development mailing list: | https://lists.sourceforge.net/lists/listinfo/scummvm-devel |
| Mentors: |
|
| Backup Mentors: |
|
Please address all items in RED ;-)
Application organized according to program FAQ
Description of the Organization
ScummVM is a collection of Virtual Machines for playing classic graphical point-and-click adventure games on modern hardware. Supported games include favorites such as Monkey Island, Simon the Sorcerer, Space Quest, and many more. To this end, the Virtual Machines (called Engines) are complete reimplementations in C++ of the engines used in the original games. The development team works either by reverse engineering game executables (usually with the permission of creators of the game), or by using the original source code of the games provided by the creators. The number of engines is constantly growing thanks to a very agile and diversified development team and ScummVM is able to run more than 200 games. The VM approach followed by ScummVM results in efficient code, which has been ported to numerous Operating Systems. Besides running on all mainstream desktop environments, namely Windows, Mac OS X and most Unix variants (Linux, *BSD, Solaris), ScummVM also runs on popular game consoles (Wii, Nintendo DS, PlayStation 2, PlayStation Portable, Dingoo and more), smart phones and PDAs (Android, WinCE, iPhone or Symbian based), and even on many not-so-mainstream systems (like BeOS, AmigaOS or OS/2). ScummVM has a highly productive team of about 51 currently active developers (out of an all-time pool of over 110), working together on a codebase of 2,100,000 lines of code. In addition ScummVM has many non-developer contributors, and a huge and highly active community. ScummVM is among the top ranking projects hosted on sourceforge.net with over 100,000 monthly downloads and ~10 million project web hits per month.
What Open Source Initiative approved license(s) does your project use?
GNU General Public License version 2.0 (GPL-2.0)
What is the URL for your Ideas list? **This is the most important part of your proposal. Please make sure we can access it and it is complete when you submit this proposal. “Placeholder” or inaccessible ideas pages will be grounds for an automatic rejection for participation in Google Summer of Code 2014.**
What is the main development mailing list for your organization?
- scummvm-devel AT lists.sourceforge.net
What is the main IRC channel for your organization?
- #scummvm on irc.freenode.net
Who will be your backup organization administrator?
- wjpalenstijn
Why is your organization applying to participate in Google Summer of Code 2014? What do you hope to gain by participating?
Each year since 2007, the program gave us the opportunity to have talented and motivated students working with us. We also love the GSoC mentor summit where we have the opportunity to share our knowledge with mentors from other projects and to learn a lot from them. That's why we wait impatiently, each year, for the announcement of a new GSoC program!
As usual, what we hope to gain is valuable code contributions. In previous years, we've had students take up and complete tasks which had been marked as "to-do", but the main developers had no time to implement them. The students took these up as self-contained GSoC projects and realized them. But also, some students have come up with new functionality, proposing and finally implementing it. These have been our favorite.
In addition, we hope to gain new developers for the project. We hope that after their projects, students will stick around and improve them or work on other interesting tasks. We hope that GSoC brings the students in touch with open source and, in our case, brings them in touch with game development. We hope these students will add their piece of code to this project, but will also keep on contributing afterwards.
We've been successful in the past seven years, and we're really looking forward to great results from the program this year too.
How many potential mentors do you have for this year's program? What criteria did you use to select them?
Mentors
- djwillis
- lordhoto
- sev
- somaen
- strangerke
Backup Mentors
- wjp
- fuzzie
- aquadran (TBC)
Description
We want our mentors to have the following qualities:
- Be able to commit to participating for the entire duration of the program. They first and foremost have to be available to their students and the mentor team.
- Have a considerable track record hacking on ScummVM. They can help the students more effectively and in an immediate fashion this way.
- Have the patience and skills to explain to their respective students on how to tackle their tasks. Also, to be able to help the students out in sticky situations.
- Have a clear vision on how a task should proceed, both in broad strokes as well as in the technical details level. Allowing, of course, some freedom of movement to the students, where this is applicable.
- Be regularly present on our #scummvm-gsoc channel, where we continuously inform each others of the progresses and issues of the students
For this year, and this holds for our previous participations too, our mentors have volunteered to work with GSoC. This means that they primarily want to be involved in the program and that they are not dragged in to participate. Moreover, they have all been contributors to ScummVM for several years. They feel comfortable around the ScummVM code and can guide students to perform their tasks. The majority of the mentors have also participated in past ScummVM GSoCs so they know their way around the procedures and have also refined their mentoring style. Some of them have been GSoC students one of the previous years. Some of them are/have been part of academia, guiding real students. They have seen the student mentalité in-action and have experience helping people along. We are drawing the best available from our pool of developers to mentor GSoC students this year again.
What is your plan for dealing with disappearing students?
We know that the students can do that. We learned this the hard way the first time 'round the GSoC ride :-) The measures we set in place during the previous years almost eradicated this problem.
As in the last few years, we will again be sure to enforce certain rules at the program start; we will clearly explain that this work should be considered as full-time, and if that is in any doubt, we will not let the student enter the program.
We have also introduced the policy that students have to provide status to their mentors on a bi-daily basis at most. If a student disappears for more than 3 days without notifying his/her mentor, the student will fail the project. The students will be made immediately aware of this during their applications. We have positively identified that frequent and meaningful communication goes a long way in keeping the students engaged and interested.
Of course, comprehensive timelines will be required as usual, and we will accept only those of the students which will set realistic goals, thus minimizing the risk of getting intimidated. This factor too frequently leads to the student disappearing.
During the program, we will make sure that the students will feel comfortable with their tasks. Our mentors already have experience with that. Moreover, in difficult situations in the past there was not just a single person feeling responsible for a particular student, but rather the whole project was trying to help when needed. These reflexes on behalf of the mentor group but also on behalf of the active team members of ScummVM have proven to be a pretty good "tool" and we are going to do our best this year too.
Based on our more negative experience two years ago, we will also make sure that internal project tensions stay internal and under control, since they were an identified cause of demotivation of our disappearing student in 2011. The current project management structure should make sure that it will not as easily happen in the future.
What is your plan for dealing with disappearing mentors?
For the mentors, the risk is relatively low; our past experiences absolutely verify this claim. In order to be even more efficient this year, three of our mentors are project leaders, core team members or project administrators and are reachable virtually 24/7 (in case of emergencies).
- We all have exchanged sufficient contact information (including cell phone numbers etc.) to be able to discover our whereabouts. Should something really bad happen which precludes a mentor from fulfilling his duties (including personal reasons), we will attempt to shift students to new mentors among the existing mentors, or drawn from our backup pool of mentors. This will depend on the number of students we have to mentor.
- We also defined for the last three years a co-mentoring system so the students have a primary and a secondary mentor, which comfort us in the idea that, in any case, the students will not be left hanging for any reason at all, no matter what happens.
- On top of that, we have a specific #scummvm-gsoc channel on IRC where mentors (and only them) are connected all the time. We use this channel to keep ourselves informed constantly of the situation of each task, each student and eventually each mentor. A mentor wouldn't disappear without being noticed very quickly by this mean too.
What steps will you take to encourage students to interact with your project's community before and during the program?
The development team actively uses the forums, IRC, Wiki and the development mailing list during the entire project development. We consider out students to be special, but developers nonetheless. Each developer including our students is encouraged to take part in discussions, whatever the means these discussions occur. As a pragmatic fact, on IRC any student will be able to get support literally 24/7, as our developers are scattered all over the globe.
- Not only the mentors, but also the entire development team, are encouraged to communicate with the students. The students are marked with a special flag on our IRC channel, so everyone knows who they are. And we require the students to write introductory letters to our development list, so everyone will have an impression about them, their skills and their assigned task (of course).
- In order to help the students familiarize themselves with the project, we also have created several pieces of documentation for them. In particular, we have an exhaustive developer central where we describe the all-important internals of ScummVM. This is valuable as a quick reference as well as during the initial explorations of the codebase.
What will you do to encourage that your accepted students stick with the project after Google Summer of Code concludes?
Last year we decided to require the GSoC student code be merged into our Master tree much earlier in the process, if possible. Our past experience was telling it should be very motivating for students to directly interact with our main repository and could potentially make some of them stay after the end of GSoC: it seems we were right as 2 students out of 4 are still actively contributing, while a 3rd one is still present from times to times. This is obviously for us a very positive sign and we plan to proceed the same way this year again.
If you chose "veteran" in the organization profile dropdown, please summarize your involvement and the successes and challenges of your participation. Please also list your pass/fail rate for each year.
We have participated in the GSoC program for seven years running, in 2007-2013.
In 2013, we were granted 4 students mentored by 6 team members. Each mentor was also backup mentor of another task. All the students passed, and 2 students are still actively contributing the project and one student is still regularly on our channel too. One of the active students would also like to participate again in GSoC this year as a student. Merging code earlier gave excellent results and we'll do the same if possible this year in order to confirm the excellent results of this year
In 2012, 5 team members mentored 4 students. Each mentor was also the co-mentor of another task, so the backup mentoring was still working well. 3 students passed, and one failed at mid-term. One student is still contributing to the project and volunteered to be a mentor for GSoC 2013 and GSoC 2014, which is really awesome. Based on discussion with other projects and on our experience, we are considering merging student code earlier in the GSoC process and have modified our processes accordingly for the next year.
In 2011, 4 team members mentored 2 students. One student succeeded in objectifying the CruisE engine, which really needed it. The other student unfortunately gave up fairly quickly after starting work (although the work done was eventually merged into our main repository, after being worked on further by a team member). After stepping back to review our processes, we feel we can still consider them to be mature. Part of the problems which caused the student to quickly give up came from internal tensions, that we have since addressed by redefining the project management structure.
In 2010 we were granted 4 slots and had 6 mentors, thus we had nice backup mentoring for every student. All four passed the finals this time, and we merged in their code. One student still continues to contribute to the project. We addressed several long standing project needs and it was just excellent.
In 2009 we had 5 students and 6 mentors. 4 of our students passed, and one failed the finals. That year the success was so big that all the students' code was merged within three months into the main development line. We were considered to be mature in our processes by that time with excellent outcome.
In 2008 we had 6 students and 7 mentors. 5 of our students were so successful that their code is included in the mainline of ScummVM, and we consider it a great achievement the fact that 4 of the students continued to contribute to the project.
In 2007 we had 7 students and 4 mentors in total. Two of our students continued to become active, regular developers in the team after having their respective code contributions integrated in the codebase. All but two of the other students succeeded in their projects.
All in all, we maintain that we are refining our method of student selection the past years and this refinement leads to better results each year. The discussions, testimonials and proposed actions which the mentor summit has brought up -and which we have participated in these six years- have helped us a great deal, during this refinement process as well. Our new management organization has also proven its efficiency during these last 2 years, and we expect much of it in the future.
Summary pass/fail: 2013: 4/0 2012: 3/1 2011: 1/1 2010: 4/0 2009: 4/1 2008: 5/1 2007: 5/2
Is there anything else we should know or you'd like to tell us that doesn't fit anywhere else on the application?
If we define a cross-project task with ResidualVM, I guess we should carefully explain this task here as well as the virtual umbrella structure detailed by googlers at fosdem. Having mentors such as somaen and aquadran on both side is obviously making the difference for Google