What’s Next: IntelliJ IDEA 2019.3 Roadmap

Last month we released IntelliJ IDEA 2019.2 with a load of awesome new features such as Java 13 Preview features, Profiling Tools, Services Tool window, and much more. Actually, the first bug-fix update for it – IntelliJ IDEA 2019.2.1 – is already available for you to download.

Now it’s time to shed some light on what we are planning to do next. The upcoming major release of IntelliJ IDEA 2019.3 is going to be different from all of our previous major releases. It will be focused more on things like performance and quality, rather than adding new functionality.

Naturally, each major release comes with myriad bug-fixes, performance fixes, and little improvements – not just new shiny features. And yet, with v2019.3 we plan to take this to a whole new level. We’re going to host a quality marathon, during which almost the whole IntelliJ team will be working on fixing the issues with the highest user impact.

We can say that this release is a bit of an experiment for us and we hope that you’ll like and support the idea of a quality-targeted release!

So let me talk in a bit more detail about the main areas that we are going to be focusing on. 

First of all, we’re going to improve the overall performance and responsiveness of IntelliJ IDEA.

Second, we are investing lots of effort into fixing issues that affect a significant number of users and into eliminating pain-points in different subsystems all over the IntelliJ Platform.

Make sure to visit our issue tracker and upvote the issues that affect you the most. This will go a long way towards helping us better define our priorities for v2019.3.

That said, it seems we will have enough resources to do things like extend support for GitHub pull requests, introduce support for several JVM Microservices frameworks, and improve Java 13 support. Still, performance and usability will receive the lion’s share of our attention and focus.

That’s about it. Stay tuned as we’re going to launch the Early Access Program for IntelliJ IDEA 2019.3 pretty soon.

Happy developing!

About Zlata Kalyuzhnaya

IntelliJ IDEA Marketing Manager at JetBrains. twitter: @ZlataKalyuzhnay ‏
This entry was posted in News & Events and tagged , . Bookmark the permalink.

54 Responses to What’s Next: IntelliJ IDEA 2019.3 Roadmap

  1. Andrew says:

    Good news! Sounds like a plan… My personal impression of Intellij is that it has pretty much stagnated in terms of providing good user experience and fixing bugs that affect usability. Good luck with that, it will be interesting to see the outcome.

  2. Christopher Rucinski says:

    I smell Android’s Project Marble influence here!

  3. Charlie Hayes says:

    I’m very excited! I’m already seeing a number of long-standing tickets getting action.

  4. Mike Mitterer says:

    A very good decision!!!!

  5. Pierre Liebsch says:

    More and more features will related in more bugs and if you don’t do such performance releases from time to time you can directly use Eclipse

    But hopefully you extend the Pull Request feature also to Gitlab

  6. Christ_off says:

    Instead of new features we will have better performances and bug fixes….
    That’s great !
    Making and IDE faster and more reliable is IMHO the first and best thing to do

  7. W says:

    Sounds awesome!

  8. Mike says:

    People aren’t resources.

  9. Alex Saveau says:

    Yes! Thank you Google for having the guts to do Project Marble.

    PS: please merge Android Studio 3.5.

  10. David says:

    This is great.
    It think it is better to fix bugs for existing features instead of building new features with more bugs.

  11. François says:

    Great news, I couldn’t agree more with this decision!
    IntelliJ is already incredibly rich with features, the main axis to improve today would clearly be its many small bugs, quirks and performance issues.
    Can’t wait to see the EAP releases flowing in :)

  12. Arkadiusz says:

    2 years too late, but good move

  13. Hello all, could you allow for more integration with external terminals (zsh and others). I love IntelliJ!

    • Dmitry Jemerov says:

      What exactly do you mean? zsh is a shell, not a terminal, and you can use it together with IntelliJ IDEA’s embedded terminal.

  14. Suresh G says:

    >introduce support for several JVM Microservices frameworks,

    Awesome to see performance focused release. So for adding new JVM micro services support, is Micronaut framework there in the plan?

  15. Pim says:

    No better features than bug fixes 😉

    I’m a frequent user of the EAP and perhaps I don’t experiment enough with it, but in my experience it’s already quite stable. I’d love to see it improve on stability further and performance.

  16. Mike says:

    Right decision, IDEA has already all necessary and sufficient features to be the top IDE on the market at H2 2019, now it’s a great time to focus on stability, reliability, and performance.

    The key directions:
    * keep standards support up-to-date, e.g. language versions, protocols, etc.
    * update 3rd-party libs dependencies
    * stability
    * performance
    * usability
    * clean-up the legacy / drop irrelevant features

  17. Mike says:

    Also, it will be great to see a report with performance benchmarks (before-after) and which % of improvements from Project Marble have been imported to the genuine IDEA code base.

    • Dmitry Jemerov says:

      Out of Project Marble, not that much is applicable to IntelliJ IDEA. For the issues that the Android Studio team found in the platform code (such as memory leaks), they have contributed fixes directly to us and they’re already integrated in 2019.2. Almost all of other changes that we could take (Windows Defender checking, default memory usage configuration, memory use analysis) are in some form already in 2019.2.

  18. Baron says:

    Two questions regarding the 2019.3 release:

    1. Will you be looking into improving indexing performance?
    2. Will you continue your efforts to slim down the release (IJ Ultimate was over 800M!)?

    Looking forward to the great improvements to come in 2019.3!

    • Dmitry Jemerov says:

      1) Yes, but it’s not our primary focus right now. Our primary focus is on issues that affect UI responsiveness, and indexing is a background activity.
      2) There won’t be any significant changes in 2019.3. We plan to stop bundling a few more plugins, but it won’t be much in terms of percentage. You can expect much more significant size reduction in future releases.

      • Roman says:

        IDEA platform needs DRASTIC changes if want stay relevant for another decade

        Software in general is becoming slower and slower, year after year, this is NOT acceptable

        Java show its weakness when it is moved from its natural habitat: servers where users doesn’t realize how sluggish it is

        • Joe says:

          You are delusional if you think Java on the desktop is sluggish. A Java desktop app will easily out-perform an Electron app, Electron isn’t even in the same ballpark performance wise. Just look at VS Code, which I used for OpenSCAD and Arduino stuff, when I rename a variable I have to type slow or the multiple cursors get out of sync…it’s laughably bad.

          As another example take a look at Cura (a 3d printing slicing program). Its interface is python+QT and python really struggles to keep up, the interface can be painful to use (the slicing engine is c++).

          I really have no idea what people are complaining about with IntelliJ. It is quite snappy for me and I am on a 7 yr old laptop.

          • Jonas says:

            I can keep 10 different projects open in as many windows in vs code, while I can barely keep two of those projects open in phpstorm. So I don’t believe a word of that. I wish you were right though cause phpstorm rocks everything apart from memory usage.

  19. Mark Hodgson says:

    It would be great to see some detailed blog posts about the process or specific fixes you’ve made and the lessons learned.
    Whether it lead to any new inspections or functionality to make this process easier.
    That sort of thing..

  20. Raman Gupta says:

    Amazing news. As a long time user of IntelliJ IDEA there is no question in my mind that performance has dropped and buggy behavior has increased over time. Congratulations on this brave decision.

  21. Stephen Friedrich says:

    Fantastic!!

    Really a couple of years overdue – but better late than never.
    I more and more stopped being passionate about IDEA during the last two years or so.
    For me, it turned from “Develop with pleasure” to “Develop with getting annoyed slightly less than with other IDEs”.

    Performance is really very important. Major bugs, of course. But for me what once was the strength of IDEA was the “smoothness” and orthogonality (what worked in one context works in others, too). So, please iron out all the tiny glitches and inconsistencies, too.

  22. Josh says:

    >GitHub pull requests

    I would love to see inline pull request reviews where I can leave a comment, or something. Would be awesome :)

    But loving this idea – giving IntelliJ some much needed TLC will make us all happier developers :)

  23. Anton says:

    I have an 8-year old minor bug, may you will fix it? Just for fun. https://youtrack.jetbrains.com/issue/IDEA-73911.
    But I’m ok to wait another 10 years 😉

  24. ADEV says:

    Today i did small experiment, i tried to run Rider with GraalVM

    And the result is pretty interesting, IDE feels snappier and consume less memory, the problem was font rendering is broken..

    Maybe Jetbrains could fork GraalVM and apply their patch there instead of the regular OpenJDK, since GraalVM introduce an AOT compiler, i think this is the perfect VM for a desktop application

    • Joe says:

      AOT would actually slow down IntelliJ. An application compiled with AOT doesn’t get on-the-fly optimizations like apps compiled with a JIT compiler do.

      The latest release of GraalVM did just add support for being able to give profiler output as input to the AOT compiler which lets it do some optimizations. One day this method may produce compiled code as fast as JIT compiled code, but not yet.

  25. Mike says:

    Performance issue can be also subjective complaint, which is strongly depends on project size, hardware, and software configurations, therefore hard to estimate.

    The real problem is support of JS project, which uses not so common JS libraries, e.g. React , Angular, and Vue.js. For instance, using UI5 (SAPUI5/OpenUI5), developer barely gets some benefits from the IDE:
    * IDEA can’t correctly decide wherever the method is used somewhere or not.
    * It feels like that code suggestions (IntelliSense) are less solid and wise in JS vs. Java, sometimes I get impression it works just by dictionary.
    * During the client-side debugging IDEA is less helpful than Chrome DevTools, where I can easily edit structure with auto-completion and edit code on the flight.

    Perhaps, under some configurations/extensions IDEA, is capable offering similar level of development experience, but DevTools do it out-of-box without any additional configurations/extensions.

  26. Mehmet Esen says:

    Finally you are headed to right way…

    I’m the only guy who uses Intellij in my company because I started programming with Java/Intellij and it has very special place for me but now working with Dart/Python mostly.

    We have an IDE that flexible and powerful enough already, let’s make it efficient/performant and make the developers really happy.

    I’m hundred percent sure, just in our company there are many developers(I can’t imagine rest of the world) who will move from VSCode to Intellij if you achieve this.

    I hope, I won’t have blow-drying fans anymore, Thanks!

  27. Kishore says:

    Awesome! Great to see a release targeted at performance and reliability! Thanks Jetbrains!

  28. Fabio says:

    Finally, what we were all hoping for! That’s awesome news.

  29. Hans says:

    Great decision. Please consider making this a regular thing, if possible (maybe twice a year or more).

  30. zgq says:

    Intellij idea’s only weakness is performance.

  31. Joe says:

    I don’t understand all the complaints in here about IntelliJ’s performance. I run it on a 7 yr old laptop with 16 gigs of RAM and performance is just fine. What kind of hardware are you all trying to run it on?

    Give it a couple gigs of heap and turn off any bundled plugins you don’t need. Problem solved.

    • Mehmet Esen says:

      And I have 2013 15 Macbook Pro with 16GB RAM.
      Just try it with multitasking. You have Chrome with multiple tabs, emulator, postman etc.

      If you have big project, when you open Intellij fans start spinning instantly and your laptop is always in high temp.

      People look at VSCode and see how lightweight and powerful it is, then they look at Intellij how extremely powerful and cumbersome it is. Mostly they prefer VSCode. This is the problem.

      Not only Java devs using Intellij these days. I was Java dev, now I’m rarely developing Java. It’s mostly Dart, Python, Rust. I saw some guys using Intellij for Elixir :) The years Java losing it’s popularity and VSCode being dominant, I think they should move from the old way.

      • zgq says:

        You talk a lot, yet Java is still the most popular language, Intellij idea is the best IDE in the world, lol

      • Joe says:

        2013 15 Macbook Pro with 16GB RAM is the exact laptop I have as well. IntelliJ performs just fine on it.

        I use VS Code for arduino and OpenSCAD and VS Code isn’t even in the same league as IntelliJ.

  32. Charlie says:

    Really looking forward to this. Honestly IntelliJ already has most of the features I’m looking for and I love the improvements, but if the performance could be dialed in to run more like a native app I’d be super happy with that if it was the focus for the next six months.

  33. Martin Jones says:

    Any chance you’ll look into this evergreen, please?
    https://github.com/JetBrains/intellij-community/pull/799

  34. Dmitry says:

    Awesome news, thanks a lot! I hope this experiment will be successful.

Leave a Reply

Your email address will not be published. Required fields are marked *