zulip/docs/overview/gsoc-ideas.md

28 KiB
Raw Blame History

:orphan:

Google Summer of Code

Zulip has been a GSoC mentoring organization since 2016, and we generally have 10-15 GSoC students each summer, depending on how many high-quality applications we receive. We have some of the highest standards of any GSoC organization; successful applications generally have dozens of commits integrated into Zulip or other open source projects by the time we review their application. See our contributing guide for details on getting involved.

About us

Zulip is a powerful, open source team chat application. The core web app is written in Python and uses the Django framework. We also make a cross-platform mobile app for iOS and Android, a cross-platform desktop app, and over 100 native integrations, all open source.

Zulip has gained a considerable amount of traction since it was released as open source software in late 2015, with code contributions from over 450 people from all around the world. Thousands of people use Zulip every single day, and your work on Zulip will have impact on the daily experiences of a huge and rapidly growing number of people.

As an organization, we value high-quality mentorship and making sure our product quality is extremely high -- you can expect to learn a lot from disciplined code reviews by highly experienced engineers. Since Zulip is a team chat product, your GSoC experience with the Zulip project will be highly interactive, with a real focus on teaching you the concepts and reasoning behind how Zulip is engineered and how to make it better.

As part of that commitment, Zulip has over 130,000 words of documentation for developers, much of it designed to explain not just how Zulip works, but why Zulip works the way that it does.

Zulip participated in GSoC 2016 and mentored three successful students officially (plus 4 more who did their proposed projects unofficially). We had 14 (+3) students in 2017 and we had 10 (+3) students in 2018. We've also mentored five Outreachy interns and hundreds of Google Code-In participants (several of who are major contributors to the project today).

Expectations for GSoC students

Our guide for having a great summer with Zulip is focused on what one should know once doing a summer project with Zulip. But it has a lot of useful advice on how we expect students to interact, above and beyond what is discussed in Google's materials.

What makes a great Zulip contributor also has some helpful information on what we look for during the application process.

We also recommend reviewing the official GSoC resources -- especially the student manual.

Finally, keep your eye on the GSoC timeline. The student application deadline is April 9, 2019.

Getting started

We have an easy-to-setup development environment, and a library of tasks that are great for first-time contributors. Use our first-time Zulip developer guide to get your Zulip development environment set up and to find your first issue. If you have any trouble, please speak up in #GSoC on the Zulip development community server (use your name as the topic).

Application tips, and how to be a strong candidate

You'll be following GSoC's application process instructions. And we'll be asking you to make at least one successful pull request before the application deadline, to help us assess you as a developer. Students who we accept generally have 5 or more pull requests merged or nearly merged (usually including at least a couple that are significant, e.g. having 100+ lines of changes or that shows you have done significant debugging).

Getting started earlier is better, so you have more time to learn, make contributions, and make a good proposal.

Your application should include the following:

  • Details on any experience you have related to the technologies that Zulip has, or related to our product approach.
  • Links to materials to help us evaluate your level of experience and how you work, such as personal projects of yours, including any existing open source or open culture contributions you've made and any bug reports you've submitted to open source projects.
  • Some notes on what you are hoping to get out of your twelve-week project.
  • A description of the project you'd like to do, and why you're excited about it.
  • Some notes on why you're excited about working on Zulip.
  • A link to the initial contribution(s) you did.

We expect applicants to either have experience with the technologies relevant to their project or have strong general programming experience. We also expect applicants to be excited about learning how to do disciplined, professional software engineering, where they can demonstrate through reasoning and automated tests that their code is correct.

While only one contribution is required to be considered for the program, we find that the strongest applicants make multiple contributions throughout the application process, including after the application deadline.

We are more interested in candidates if we see them submitting good contributions to Zulip projects, helping other applicants on GitHub and on chat.zulip.org, learning from our suggestions, trying to solve their own obstacles and then asking well-formed questions, and developing and sharing project ideas and project proposals that are plausible and useful.

Also, you're going to find that people give you links to pages that answer your questions. Here's how that often works:

  1. you try to solve your problem until you get stuck, including looking through our code and our documentation, then start formulating your request for help
  2. you ask your question
  3. someone directs you to a document
  4. you go read that document, and try to use it to answer your question
  5. you find you are confused about a new thing
  6. you ask another question
  7. now that you have demonstrated that you have the ability to read, think, and learn new things, someone has a longer talk with you to answer your new specific question
  8. you and the other person collaborate to improve the document that you read in step 3 :-)

This helps us make a balance between person-to-person discussion and documentation that everyone can read, so we save time answering common questions but also get everyone the personal help they need. This will help you understand the rhythm of help we provide in the developers' Zulip livechat -- including why we prefer to give you help in public mailing lists and Zulip streams, instead of in one-on-one private messages or email. We prefer to hear from you and respond to you in public places so more people have a chance to answer the question, and to see and benefit from the answer. More about that in this blog post.

Mentors

We have more than a dozen Zulip contributors who are interested in mentoring projects. We usually decide which contributors are mentoring which projects based in part on who is a good fit for the needs of each student as well as technical expertise. You can reach us via #GSoC on the Zulip development community server, (compose a new stream message with your name as the topic).

Zulip operates under group mentorship. That means you should generally post in public streams on chat.zulip.org, not send private messages, for assistance. Our preferred approach is to just post in an appropriate public stream on chat.zulip.org and someone will help you. We list the Zulip contributors who are experts for various projects by name below; they will likely be able to provide you with the best feedback on your proposal (feel free to @-mention them in your Zulip post).

However, the first and most important thing to do for building a strong application is to show your skills by contributing to a large open source project like Zulip, to show that you can work effectively in a large codebase (it doesn't matter what part of Zulip, and we're happy to consider work in other open source projects). The quality of your best work is more important to us than the quantity; so be sure to test your work before submitting it for review and follow our coding guidelines (and don't worry if you make mistakes in your first few contributions! Everyone makes mistakes getting started. Just make sure you don't make the same mistakes next time).

Once you have several PRs merged (or at least one significant PR merged), you can start discussing with the Zulip development community the project you'd like to do, and developing a specific project plan. We recommend discussing what you're thinking in public streams on chat.zulip.org, so it's easy to get quick feedback from whoever is online.

Project ideas

These are the seeds of ideas; you will need to do research on the Zulip codebase, read issues on GitHub, and talk with developers to put together a complete project proposal. It's also fine for you to come up with your own project ideas. As you'll see below, you can put together a great project around one of the area labels on GitHub; each has a cluster of problems in one part of the Zulip project that we'd love to improve.

We don't believe in labeling projects by difficulty (e.g. a project that involves writing a lot of documentation will be hard for some great programmers, and a UI design project might be hard for a great backend programmer, while a great writer might have trouble doing performance work). To help you find a great project, we list the skills needed, and try to emphasize where strong skills with particular tools are likely to be important for a given project.

For all of our projects, an important skill to develop is a good command of Git; read our Git Guide in full to learn how to use it well. Of particular importance is mastering using Git rebase so that you can construct commits that are clearly correct and explain why they are correct.

Focus areas

For 2019, we are particularly interested in GSoC students who have strong skills at visual design, HTML/CSS, mobile development, performance optimization, or Electron. So if you're a student with those skills and are looking for an organization to join, we'd love to talk to you!

The Zulip project has a huge surface area, so even when we're focused on something, a huge amount of essential work goes into other parts of the project. Every area of Zulip could benefit from the work of a student with strong programming skills; so don't feel discouraged if the areas mentioned above are not your main strength.

As a data point, in Summer 2017, we had 4 students working on the React Native mobile app (1 focused primarily on visual design), 1 on the Electron Desktop app, 2 on bots/integrations, 1 on webapp visual design, 2 on our development tooling and automated testing infrastructure, and the remaining 4 on various other parts of the backend and core webapp.

Full stack and web frontend focused projects

Code: github.com/zulip/zulip -- Python, Django, JavaScript, and CSS.

  • Fill in gaps in Zulip's library of native integrations. We have about 100 integrations, but there are a handful of important integrations that are missing. The the integrations label on GitHub lists some of the priorities here (many of which are great preparatory projects); once those are cleared, we'll likely have many more. Skills required: Strong Python experience, will to do careful manual testing of third-party products. Fluent English, usability sense and/or technical writing skills are all pluses. Expert: Eeshan Garg.

  • Fill in the gaps in Zulip's REST API documentation. Zulip has a nice framework for writing API documentation built by a student last summer based on the OpenAPI standard with built-in automated tests, but there are a few dozen endpoints that are missing, several of which are quite important. See the API docs area label for good starter projects and this issue for a relevant TODO list.

  • Make Zulip integrations easier for nontechnical users to setup. This includes adding a backend permissions system for managing bot permissions (and implementing the enforcement logic), adding an Oauth system for presenting those controls to users, as well as making the /integrations page UI have buttons to create a bot, rather than sending users to the administration page. Skills recommended: Strong Python/Django; JavaScript, CSS, and design sense helpful. Understanding of implementing Oauth providers, e.g. having built a prototype with the Django Oauth toolkit would be great to demonstrate as part of an application. The Zulip integration writing guide and integration documentation are useful materials for learning about how things currently work, and the integrations label on GitHub has a bunch of good starter issues to demonstrate your skills if you're interested in this area. Expert: Eeshan Garg.

  • Build a meta-integration that converts the Slack incoming webhook API to post messages into Zulip. Zulip has several dozen native integrations (https://chat.zulip.org/integrations/), but Slack has a ton more. We should build an interface to make all of Slacks numerous third-party integrations work with Zulip as well, by basically building a Zulip incoming webhook interface that accepts the Slack API (if you just put in a Zulip server URL as your "Slack server"). Skills required: Strong Python experience; experience with the Slack API a plus. Work should include documenting the system and advertising it. Expert: Steve Howell.

  • Visual design work on Zulip's logged-out pages, including /help, /apps, /integrations, /api, /login, /register, the zulip.org website, etc. We'd love to make these look nicer both through polish and potentially through adding fun illustrations to make the product more friendly. A project could include work on Zulip's logged-in UI as well. Skills required: Design, HTML and CSS skills; JavaScript and illustration experience are helpful. A great application would come with mockups for specific changes, and/or a set of PRs making small improvements to the logged-out UI. Expert: Tim Abbott.

  • Build support for outgoing webhooks and slash commands into Zulip to improve its chat-ops capabilities. There's an existing pull request with a lot of work on the outgoing webhooks piece of this feature that would need to be cleaned up and finished, and then we need to build support for slash commands, some example integrations, and a full set of documentation and tests. Recommended reading includes Slack's documentation for these features, the Zulip message sending code path, and the linked pull request. Skills required: Strong Python/Django skills. Expert: Steve Howell.

  • Build a system for managing Zulip bots entirely on the web. Right now, there's a somewhat cumbersome process where you download the API bindings, create a bot with an API key, put it in configuration files, etc. We'd like to move to a model where a bot could easily progress from being a quick prototype to being a third-party extension to being built into Zulip. And then for built-in bots, one should be able to click a few buttons of configuration on the web to set them up and include them in your organization. We've developed a number of example bots at contrib_bots/ in the main Zulip repository that can be used for testing; the design document for the deployment part of this vision (likely part 1) is here. Skills recommended: Python and JavaScript/CSS, plus devops skills (Linux deployment, Docker, puppet etc.) are all useful here. Experience writing tools using various popular APIs is helpful for being able to make good choices. Experts: Steve Howell.

  • Improve the UI and visual design of the existing Zulip settings and administration pages. Last summer, a student built a nice framework for many common elements (e.g. checkboxes, dropdowns, etc.) and migrated the codebase to use it, but the tables settings screens with tables are inconsistent, only one of them has a convenient sorting feature, etc. You can get a great sense of what needs to be done by playing with the settings/administration/streams overlays in a development environment. You can get experience working on the subsystem by working on some of our open settings/admin issues. Skills recommended: JavaScript, HTML, CSS, and an eye for visual design. Expert: Shubham Dhama.

  • Optimize web frontend performance and scalability. Zulip is already one of the faster webapps out there, but there are a bunch of ideas for how to make it substantially faster. This is likely a particularly challenging project to do well, since there are a lot of subtle interactions to understand. Skill recommended: Strong debugging, communication, and code reading skills are most important here. JavaScript experience; some Python/Django experience, some skill with CSS, ideally experience using the Chrome Timeline profiling tools (but you can pick this up as you go). Expert: Steve Howell.

  • Build out the administration pages for Zulip to let admins set a retention policy for when old messages should be deleted, audit data, etc. ... the sorts of things needed for Zulip to be used at larger organizations. We get constant requests for these kinds of features from Zulip users. The Zulip bug tracker has almost 50 open issues( settings (admin/org), settings UI, settings (user), stream settings ) in the space of improving the Zulip administrative UI. Many are little bite-size fixes in those pages, which are great for getting a feel for things, but a solid project here would be implementing 5-10 of the major missing features. The first part of this project will be refactoring the admin UI interfaces to require writing less semi-duplicate code for each feature. Skills recommended: A good mix of Python/Django and HTML/CSS/JavaScript skill is ideal. The system for adding new features is well documented. Expert: Shubham Dhama.

  • Rebuild the Zulip web UI using a modern reactive layer like vue.js. Strategically, we'd start with self-contained, messy pieces (like the presence layer), then move on to more complex pieces (like the subscriptions page), and finally attach the main UI. Definitely worth reading the vue.js documentation and reading how Zulip's real-time sync works. Skills recommended: Strong JavaScript experience, good communication skills and an eye for detail. We think this would be an awesome project, but rewrite projects often introduce lots of bugs, so we're interested in particularly careful candidates who have the discipline to redo a small component at a time and carefully test for regressions. Good ways to demonstrate qualification for this are finding and reporting bugs using Zulip's manual UI testing guide and doing refactoring projects. Expert: Tommy Ip, Tim Abbott.

  • Work on Zulip Terminal, the official terminal client for Zulip. zulip-terminal is already a basic usable client, but it needs a lot of work to approach the webapp's quality level. We would be happy to accept multiple strong students to work on this project. Skills required: Python 3 development skills, good communication and project management skills, good at reading code. Experts: Aman Agrawal, Neil Pilgrim.

  • Write cool new features for Zulip. Play around with the software, browse Zulip's issues for things that seem important, and suggest something youd like to build! A great project can combine 3-5 significant features. Experts: Depends on the features!

  • Work on Zulip's development and testing infrastructure. Zulip is a project that takes great pride in building great tools for development, but there's always more to do to make the experience delightful. Significantly, a full 10% of Zulip's open issues are ideas for how to improve the project, and are in these four labels for tooling improvements. A good place to start is backend test coverage.

    This is a somewhat unusual project, in that it would likely consist of dozens of small improvements to the overall codebase, but this sort of work has a huge impact on the experience of other Zulip developers and thus the community as a whole (project leader Tim Abbott spends more time on the development experience than any other single area).

    A possible specific larger project in this space is working on adding mypy stubs for Django in mypy to make our type checking more powerful. Read our mypy blog post for details on how mypy works and is integrated into zulip. This specific project is ideal for a strong contributor interested in type systems.

    Skills required: Python, some DevOps, and a passion for checking your work carefully. A strong applicant for this will have completed several projects in these areas.

    Experts: Tim Abbott (provision, testing), Steve Howell (tooling, testing).

  • Write more API client libraries in more languages, or improve the ones that already exist (in python, JavaScript, PHP, and Haskell). Skills required: Experience with the target language and API design. Expert: Depends on language.

  • Develop @zulipbot, the GitHub workflow bot for the Zulip organization and its repositories. By utilizing the GitHub API, @zulipbot improves the experience of Zulip contributors by managing the issues and pull requests in the Zulip repositories, such as assigning issues to contributors and appropriately labeling issues with their current status to help contributors gain a better understanding of which issues are being worked on. Since the project is in its early stages of development, there are a variety of possible tasks that can be done, including adding new features, writing unit tests and creating a testing framework, and writing documentation. Skills required: Node.js, ECMAScript 6, and API experience. Expert: Cynthia Lin, Joshua Pan.

React Native mobile app

Code: React Native mobile app. Experts: Greg Price, Boris Yankov.

The highest priority for the Zulip project overall is improving the Zulip React Native mobile app.

  • Work on issues and polish for the app. You can see the open issues here. There are a few hundred open issues across the project, and likely many more problems that nobody has found yet; in the short term, it needs polish, bug finding/squashing, and debugging. So browse the open issues, play with the app, and get involved! Goals include parity with the webapp (in terms of what you can do), parity with Slack (in terms of the visuals), world-class scrolling and narrowing performance, and a great codebase.

A good project proposal here will bundle together a few focus areas that you want to make really great (e.g. the message composing, editing, and reacting experience), that you can work on over the summer. We'd love to have multiple students working on this area if we have enough strong applicants.

Skills required: Strong programming experience, especially in reading the documentation of unfamiliar projects and communicating what you learned. JavaScript and React experience are great pluses, as are iOS or Android development/design experience is useful as well. You'll need to learn React Native as part of getting involved. There's tons of good online tutorials, courses, etc.

Electron Desktop app

Code: Our cross-platform desktop app written in JavaScript on Electron. Experts: Akash Nimare, Abhighyan Khaund

  • Contribute to our Electron-based desktop client application. There's plenty of feature/UI work to do, but focus areas for us include things to (1) improve the release process for the app, using automated testing, typescript, etc. and (2) making it easy to install the desktop app on a large number of machines for enterprise deployment. Browse the open issues and get involved!

Skills required: JavaScript experience, Electron experience. You can learn electron as part of your application!

Good preparation for desktop app projects is to (1) try out the app and see if you can find bugs or polish problems lacking open issues and report them and (2) fix some polish issues in either the Electron app or the Zulip web frontend (which is used by the electron app).

Circulating proposals (February-March)

If you're applying to GSoC, we'd like for you to publicly post a few sections of your proposal -- the project summary, list of deliverables, and timeline -- some place public on the Web, sometime in February or March. That way, the whole developer community -- not just the mentors and administrators -- have a chance to give you feedback and help you improve your proposal.

Where should you publish your draft? We prefer Dropbox Paper or Google Docs (or even just a message in Zulip), since those platforms allow people to look at the text without having to log in or download a particular app, and you can update the draft as you improve your idea. In either case, you should post the draft for feedback in chat.zulip.org.

Rough is fine! The ideal first draft to get feedback from the community on should include primarily (1) links to your contributions to Zulip (or other projects) and (2) a paragraph or two explaining what you plan to work on. Your friends are likely better able to help you improve the sections of your application explaining who you are, and this helps the community focus feedback on the areas you can most improve (e.g. either doing more contributions or adjusting the project plan).

We hope to hear from you! And thanks for being interested in Zulip. We're always happy to help volunteers get started contributing to our open source project, whether or not they go through GSoC.