# WEH 2025 - Involving students and researchers in web engine implementation

* GitHub issue: https://github.com/Igalia/webengineshackfest/issues/47
* URL: https://meet.jit.si/WEH2025-Students

**Outline of session:**
- Overview of collaboration between University of Bergen (Norway) and TC39.
- Presentation by students from Universidad de La Laguna (Spain).
- Discussion.

**Introduction presentations:**
- Slides: [University of Bergen](<https://docs.google.com/presentation/d/1rdNz_7IMK0z7t4hbc1mp1wsKTu9c_H8glXZQa3K0Wsw/edit?usp=sharing>)
- Slides: [Universidad de La Laguna](<https://docs.google.com/presentation/d/1kYAo2xSDTbYNn3JDCzNO_aZ5gfeQ7EjiXZOHzLS0cuE/edit?usp=sharing>)
- ["Procesadores de Lenguajes" course notes](https://apuntes-pl.vercel.app/). Use your GitHub account to authenticate.
- [Lab "extending JS with Babel"](https://apuntes-pl.vercel.app/labs/babel)

**Discussion:**
Q: Should we aim for real contributions everywhere, or is it okay for some projects to be mostly educational?

- Sideshowbarker: My vote is for optimizing for real contributions. A way to find out what could be a real contribution is coming to places like this. We have real need for assistance in our test suite, like test262. On the web platform API side, we have "web platform tests". Test case coverage against the spec: look at the spec, find out if there is test coverage, how sufficiently is it covered? Is the test suite aligned with the spec? Too much incentive to get patches landed.
- Casiano Rodríguez León: In this case, I believe there is room for master thesis on these tests. Depends on the students and their capacities. One way in which students can help and contribute.
- Nicolo: Very cool, learn how to implement and maintain. I'm a maintainer of Babel, we make it hard on purpose to make new syntax in babel. Surprised you did something good from scratch.
- Adrian (Student): When doing that, I wondered why it was so hard.
- Pablo (Student): It took long to build. It was mostly experimental, there are something in the course in a toy language we were making. How do we fit it into javascript. How does it fit within the design of javascript.
- Nicolo: Other than the feature to implementation, from the feature try writing spec for it -- a "javascript proposal"
- Casiano Rodríguez León: We hope that future students will add formal specification and address real (small) proposals. I'm now 70, it is my last year and I'm retiring this year but hopefully, the next teacher will continue this line of work.
- Presentor (Jonas): I want to go back to what sideshowbarker -- if you find bugs in implementation, would implementors be more open to being mentors to students?
- Side show barker: It would be hard to find in a company, people are busy. People will be open to some degree to help.
- Presentor (Mikhail): From the University of Bergen, we participated in 5 proposals of the API. The recent proposals of TC39, they were all API proposals, introducing new syntax is a very high bar. If you have no proposals introducing new syntax, how do we make sure we will have students who are able to implement syntactic proposals. With that in mind, it might still be beneficial to have educational results.
- Shane: the one of the things that could make the problem of mentors be more successful, the faculty member is a mentor, the person in the open source process is just a subject matter expertise. So the open source mentor gets to look at code that is already polished, the faculty member helps the student first.
- Morgan: Flip this back to the academics, like any open source project, we have full time maintainers. We would like to attract the next generation and invest time in new developers. From your persective, as academcis, as students, what are you looking for in a project as new contributors?
- Student: I'm looking for a way to get started on myself. It's a high bar to ask questions to maintainers, I don't want to be a bother.
- snek: This got me thinking about a project called [engine262](<https://github.com/engine262/engine262>), a JavaScript implementation written in JavaScript, intended to validate the spec, tests, implementation. Maybe this is a great place for students, more educationally directed playground, while taking on real world things, like validating real proposals, test coverage.
- Casiano (Professor): I think KAIST [worked](<https://github.com/es-meta/esmeta>) on something like this.
- snek: There a project that tries to validate spec tests, this is different than that.
- Claudio: If you work on real implementations, you will learn how to work with people. If you want to go into the industry, it is really valuable to work on real problems with real people. I work at Igalia, finding people with real world browser companies, so it would be good for us. We have this project called ["coding experience"](<https://www.igalia.com/coding-experience/>), where we are mentors and help students.

Q: What are the incentives for browser vendors, engine implementers, standards groups, and universities to participate?

- Presentor (Mikhail): Some of the students liked the most the experience of contributing to an open source projects. At the University of Bergen, we value participation in open source and standards work.
- Tom: Always useful to have an implementation you only have to review instead of write.
- ??: I like it when I'm working with/mentoring students.
- sideshowbarker: Yeah its always hard to hire. There are specs that need help. Like the [Scalable Vector Graphics (SVG)](<https://www.w3.org/Graphics/SVG/>) spec, it needs maintanence/work. UI events are underspecified.
- Tom: Something else I wanted to add, I think you mentioned you write a contribute guide.
- Val: I wonder if something like [outreachy](<https://www.outreachy.org/>), the organizers of outreachy go to opensource projects and request projects, a lot of individual contact/connections.
- ???: Mobile development uses miniapps(?), and most miniapps uses webEngines, so when we trie to hire people, either in China or elsewhere it is always hard to find poeple with relevant experience (I think this is along the lines of what said).
- Sideshowbarker: to find a place where there is real time discusison, where to the project maintainers hang out. Go to the place or github. Webkit and chromium have a slacks, but you need a invite.
- Morgan: Developers are building and running things every day, often we write docs but they get out of date, if you contribute/fix docs, you will get a lot of appreciation/respect.
- vhilla: New at mozilla, contributed as student before. Would be good to have central points of contact at teaching chairs that can build relationships over years with maintainers and gain an understanding of their processes. To find tasks to work on. You can look at open standard issues. We have [bugzilla](<https://bugzilla.mozilla.org/home>) to see bugs in mozilla, good places to start are "[parity bugs](<https://bugzilla.mozilla.org/buglist.cgi?list_id=17578417&product=Core&bug_severity=S3&bug_type=enhancement&resolution=---&keywords=parity-chrome%2Cparity-safari%2C%20&classification=Client%20Software&classification=Developer%20Infrastructure&classification=Components&classification=Server%20Software&classification=Other&keywords_type=allwords&query_format=advanced>)", or [good-first-bugs](<https://bugzilla.mozilla.org/buglist.cgi?resolution=---&keywords=good-first-bug%2C%20&classification=Client%20Software&classification=Developer%20Infrastructure&classification=Components&classification=Server%20Software&classification=Other&query_format=advanced&bug_severity=S3&product=Core&list_id=17578410&keywords_type=allwords&bug_type=enhancement>). Feel free to request a "needinfo" from people associated with the bugs to assess whether this is something you could work on, and what mentorship is possible.

Q: how much involvement is needed from the browser implementors, and how do we make it sustainable?

- Presentor (Mikhail): Mentoring was an hour to 1/2 hour a week for 3-4 months. How to make is sustainable, without mentorship. That is why we had the idea to write very detailed tutorials. ... How much involvment is managable as a implementer/maintainer?

- Val: Maybe you could find advocates within companies who would create programs in the company like the Code Experience program at Igalia, where people at the company are allowed to spend some working hours on mentoring.

- SideshowBarker: Do you know of [Pavel Panchekha](<https://pavpanchekha.com>) at the University of Utah? Has a course in [browser engineering](<https://browser.engineering>). They have written a book about browser engines, not about the JS part, unfortunately. Is it possible to get in contact with Pavel, and brainstorm/discuss as they might have the same challenges?

- (missed)

- ?????: real contributions, like experimenting with how effective fingerprinting. So many interesting research areas, in security.
morgan: I would also +1 fingerprinting research in general, browser specific thing. A lot of contributions as project maintainers value doing debugging for whatever problem they are experiencing

- Val: You have a course in compilers/engines, do you want to work on the other kinds of problems we have been discussing? Like svg/fingerprinting/web accessibility.

- Presenter (Mikhail): We have a separate course on compilers, but no specific course on web engines. For the projects that I described, we have been using a special course that interested students can take to work on this kind of stuff.

- Presenter (Mikhail): How to inspire other universities to do something similar is the question? The University of Bergen has been involved in Fortran, C++ and JS development for a long time already. And other universities aren't necessarily as inclined to do this. This maybe be perceived by some universities as a too practical work.

- Presenter (Mikhail): There have been some collaborative work and different projects with other universities about standards work, see slides. How do we involve researchers, not students, in this work? TC39 has 6 meetings a year, 3 in person. For the in person ones, we do a [workshop](<https://github.com/tc39/tg5/tree/main/workshops>) where we reach out to a local university and try to have them present some of their work and collaborate.

- Val: For WC3 in persons meeting, we try to present for local universities would be cool to do!

- Lola: I'm on the technical committee of the W3C I've done lots of presentations for students.

- Stephen: From a scalability perspective, browser engineers are few and students are many, so gathering students togheter in groups where they can work togheter and collaborate on work outside of the course schedule is powerfull and probably something that is more scallable.

- Sideshowbarker: This academic [Pavel Panchekha](<https://pavpanchekha.com>) says: No student should go through a CS or Software Engineering degree without taking a course in browser engineering.