I remember reading about JetBrains in terms of profitability and capital. If I remember correctly, JetBrains has never raised any money from outside sources and has become profitable very quickly.
In terms of IDE, whenever I get into unfamiliar territory, JetBrains products have always been better at onboarding me into it without having to configure much. For getting into Clojure for example, IntelliJ+Cursive was my preferred choice.
I agree with the sentiment entirely, but reflecting on what you wrote it's so strange that we have to call this out as a slightly
novel way to do business.
It almost feels quaint but very comforting. I wonder if there will be a pushback against companies with lots of investor money. It seems they are all doomed to grow quickly at any cost, shady or not.
Who knows for how long, they are in tough spot (hence the 'reimagine' in the title I guess), vscode is eating their market share fast, they probably already lost the web dev market they had and, for Java/C# is just a matter of time, a few years but today is decent.
Can only speak about technologies I work with so for PHP, JetBrains PHPStorm is probably the only decent IDE. WebStorm is also better for web development, primarily you get all the features out of the box without having to configure anything. I tried VSCode but found that I need to use plugins for everything and there are multiple plugins for the same feature, which means I have to try all of them to see what suits me. This is a time consuming process and I'd rather I use WebStorm instead :)
I used vscode as my daily driver for a long time, I find webstorm is a much better experience, particularly because vscode is so centred around javascript and non-js plugins are lacking. I work in consulting, so my tech stack changes every few months. Being able to swap from pycharm to phpstorm for example and have familiar keybindings is nice.
Wut...? Vscode is “eating” the cheapest end of the market, where they were present (with the Community editions) but not seriously competing.
They are fine. Their financials are sound and they are plenty profitable. Their real threat is Github and similar services, who try to persuade users that they don’t need an IDE at all. Hence why they’ve come up with a collab alternative that they presented with great fanfare. Imho they should just buy gitlab and integrate with it a bit better.
Yes, it is - I develop Cursive and I don't work for JetBrains.
That said, even though Cursive is third party, the way that IntelliJ is designed means that language extensions all tend to work in consistent ways. Rather than totally reinventing the wheel, I basically plug into standard IDE workflows for different types of extensions, and make them work for Clojure. This means that Cursive is able to feel like a JetBrains product, although sometimes less polished because I work on it on my own.
It's one of the things I love most about IntelliJ, all the support for different languages basically works in the same way and it's very easy for me to pick up a new language if there's support for it.
JetBrains IDEs are great. Each update seems to be a windfall of "wow that's really useful" and "yeah I've always wanted that".
But man I wish they'd improve the indexing. Pretty much the only thing that can thrash my CPU hard enough to cause noticable hanging. Perhaps their machine learning can do something about that... Lazy indexing based on recent areas of focus or something.
While I'm sure it's useful, JVM advice like this is usually less helpful than how much ram you're giving it (Xmx, xms). Particularly because you'll have an entirely different experience between an RPi, laptop, or desktop.
> "Even Microsoft has made its long-living Visual Studio available for free."
Microsoft has other products beyond Visual Studio. People using Visual Studio are likely using other Microsoft products. The business case for JetBrains is different. JetBrains' business is selling their IDE products.
Sounds great. I’ve been a delighted Jetbrains user since I started using IntelliJ in 2002, when it was clearly head and shoulders above every other Java IDE, and I’ve been pleased to see the company’s continued growth and success over the past 18 years.
They have to be the only company shipping a significant desktop Java/Swing app today? I wonder how they feel about the future of that platform.
I'm pretty sure they are mainly developing in (and converted) a significant portion of the codebase to Kotlin (developed in-house to make it less painful).
He was probably referring to the Swing GUI toolkit, which allows you to develop a GUI once and run on Windows, Mac, and Linux and keep mostly the same look and feel.
Swing was popular briefly back in the Applet days on the internet and also for a lot of internal corporate apps, but never saw much use in commercial and open source apps.
I think the main GUI libs were removed in version 11 of Java released a year or so ago, and now they're packaged separately.
According to Oracle, support will be provided for a few more years 'commercially' which means you'd have to pay them if you want security updates. They're apparently trying to push off the stewardship of the library to the community, like they've already done with a lot of Java technology that they've ignored since buying Sun (e.g. Netbeans platform).
I'm not sure what that means for the OpenJDK versions which is what most of the world now uses instead of Oracle.
It is without a doubt the best $100 you can spend as a .NET developer. The only complaint I have is that it thrashes the disk because VS can't be arsed to make a 64-bit version, and so there's a stupid memory ceiling.
I actually prefer JetBrains Rider these days. Works great for general C# development, cross-platform, and doesn't thrash my disk like crazy.
Granted, I just spent about two years doing Java dev in IntelliJ and recently moved back over to working with C#. I'd gotten pretty used to working with IntelliJ so that probably colors my viewpoint a little, but the JetBrains IDEs do have some features I really like - the class structure pane, for example, is really useful.
Great line of products. I love GoLand, CLion and Datagrip. I mostly use VSCode now, but I've always preferred the debugging experience and tooling with JetBrains products for my more focused projects.
There are several reasons I’ll never touch a JB product:
1. I hate the idea of paying for the thing I write code in. I may not be a FS nut, but I’m a big believer in supporting open source by using open source.
2. IDEs tend to be optimized for the write path, when it’s the read path that really matters. Giving people too many tools to rush through writing code is a bad idea. That’s basically how Java became unusable. Writing code is bad. The less code you can write to solve your problem the better.
3. Most of the people I admire at work and in open source don’t use an IDE. Clearly, an IDE is not a necessity to being a great programmer, so why shell out money for one.
4. Not being able to use my text editor I’m a terminal is just a nonstarter for me. It’s immensely convenience, and some workplaces can’t even handle local development because of monorepos.
5. IDEs rob you of the knowledge of the underlying tools, locking you into an ecosystem that you can’t take with you all the time.
6. I’ve worked in plenty of places that didn’t have JB stuff available. Never worked at a place you couldn’t use vim or emacs.
My philosophy on text editors is that you can get 90% of the benefit of an IDE with way less tooling. I use vim, ALE, and fzf/rg and I’ve never felt the least bit slowed down by my tooling.
I have to disagree strongly with point 2. IDEs are much better at the read path than simple text editors are. Critical to reading code quickly is being able to navigate it quickly, answering questions like "What are the subclasses of this?", "Where is this called?", "What are the methods in this interface?", "What does super do?", etc. Grep can only take you so far. Language server protocol is definitely improving things, but it still can't answer questions like "What is the chain of function calls linking the creation of this object to its eventual use?"
You’ve been able to do this with ctags for a long time, which is also the only way you’re ever going to get goto def and dec in a large project like the Linux kernel. Vim even has built in goto, search, etc for ctags.
But really, if an IDE is allowing you to write code you can’t understand without the IDE that’s a bad thing. Forcing yourself into a situation where you have to compartmentalize is going to make you a better programmer and lead to better code.
Strong disagree with your thesis here. Most navigators of code are looking for something, and that something is not the most elegant refactor to make to cut some particular Gordian knot.
Navigation tools are essential especially because the (vastly) larger audience is uninitiated.
>1. I hate the idea of paying for the thing I write code in.
Do you hate the idea of being paid for the code you write? Or it's only other people's software that should be free?
>2. IDEs tend to be optimized for the write path, when it’s the read path that really matters.
Actually it's the exact opposite. Plain editors like vim are optimized for the write path -- they have all kinds of text manipulation options built in, and they're 99% about text manipulation. IDEs have code navigation options, source control integration, refactoring tools, even collaboration options in some cases, all of which are about the read path.
>3. Most of the people I admire at work and in open source don’t use an IDE.
Some of them use an Emacs so loaded that it's just a Greenspun-10th-rule-style re-implementation of an IDE. And tons of great developers use IDEs, from Notch to John Carmack, to Ken Beck. If you constrict the "people you admire" to old school UNIX neckbeards, then sure.
>4. Not being able to use my text editor I’m a terminal is just a nonstarter for me. It’s immensely convenience, and some workplaces can’t even handle local development because of monorepos.
Monorepos have nothing to do with local development or not. You can export a specific project out and work on it. Googlers use all kinds of IDEs, despite the monorepo.
Not being able to run in a terminal being a non-starter, is dubious, as 99% of the time a programmer doesn't need to work in a remote shell session, exception if they're a sys admin.
>5. IDEs rob you of the knowledge of the underlying tools, locking you into an ecosystem that you can’t take with you all the time.
Take with you were? You aren't going anywhere. You can always use an IDE whatever the circumstances. People use IDEs in games, data science, application programming, embedded development, aviation, web programming, etc. Not sure which realm would prohibit the use of an IDE.
>6. I’ve worked in plenty of places that didn’t have JB stuff available. Never worked at a place you couldn’t use vim or emacs.
There are tons of companies that mandate a specific corporate-decided development environment for all programmers, where you can't just "have vim or emacs".
Try using your favourite language will lead you to search for a plug-in. Then you discover you need the ultimate edition. Parts of it are open source but all of the parts you need aren't.
regarding #1: As others mentioned, IntelliJ (the platform) is open source, but even if it weren't, it would be a superior model.
I'm a user of JB products for some years now. last week I had to open eclipse again, which is probably THE foss IDE for java. Developing in it is mostly fine, but if you get to the tiny details then IntelliJ is far better (multiple Carets, a proper dark mode, good vcs integration, UI is to the point and not as crowded as Eclipse, integrated terminal, etcpp).
I think that it shows the limits of (pure) open source and that the paid model really can polish a complex product. I imagine that contributors and drive-by patches are not really able to create a coherent product which maybe needs a proper plan.
(This really reminds me of that Bazaar vs Cathedral argument)
In terms of IDE, whenever I get into unfamiliar territory, JetBrains products have always been better at onboarding me into it without having to configure much. For getting into Clojure for example, IntelliJ+Cursive was my preferred choice.
They are fine. Their financials are sound and they are plenty profitable. Their real threat is Github and similar services, who try to persuade users that they don’t need an IDE at all. Hence why they’ve come up with a collab alternative that they presented with great fanfare. Imho they should just buy gitlab and integrate with it a bit better.
Their recent effort on EduTools was impressive way to teach many programming concepts.
That said, even though Cursive is third party, the way that IntelliJ is designed means that language extensions all tend to work in consistent ways. Rather than totally reinventing the wheel, I basically plug into standard IDE workflows for different types of extensions, and make them work for Clojure. This means that Cursive is able to feel like a JetBrains product, although sometimes less polished because I work on it on my own.
It's one of the things I love most about IntelliJ, all the support for different languages basically works in the same way and it's very easy for me to pick up a new language if there's support for it.
But man I wish they'd improve the indexing. Pretty much the only thing that can thrash my CPU hard enough to cause noticable hanging. Perhaps their machine learning can do something about that... Lazy indexing based on recent areas of focus or something.
It would make even more users. Even Microsoft has made its long-living Visual Studio available for free.
Microsoft has other products beyond Visual Studio. People using Visual Studio are likely using other Microsoft products. The business case for JetBrains is different. JetBrains' business is selling their IDE products.
They have to be the only company shipping a significant desktop Java/Swing app today? I wonder how they feel about the future of that platform.
It is still Java underneath, though
Swing was popular briefly back in the Applet days on the internet and also for a lot of internal corporate apps, but never saw much use in commercial and open source apps.
I think the main GUI libs were removed in version 11 of Java released a year or so ago, and now they're packaged separately.
According to Oracle, support will be provided for a few more years 'commercially' which means you'd have to pay them if you want security updates. They're apparently trying to push off the stewardship of the library to the community, like they've already done with a lot of Java technology that they've ignored since buying Sun (e.g. Netbeans platform).
I'm not sure what that means for the OpenJDK versions which is what most of the world now uses instead of Oracle.
* https://www.oracle.com/technetwork/java/javase/javaclientroa...
Granted, I just spent about two years doing Java dev in IntelliJ and recently moved back over to working with C#. I'd gotten pretty used to working with IntelliJ so that probably colors my viewpoint a little, but the JetBrains IDEs do have some features I really like - the class structure pane, for example, is really useful.
1. I hate the idea of paying for the thing I write code in. I may not be a FS nut, but I’m a big believer in supporting open source by using open source.
2. IDEs tend to be optimized for the write path, when it’s the read path that really matters. Giving people too many tools to rush through writing code is a bad idea. That’s basically how Java became unusable. Writing code is bad. The less code you can write to solve your problem the better.
3. Most of the people I admire at work and in open source don’t use an IDE. Clearly, an IDE is not a necessity to being a great programmer, so why shell out money for one.
4. Not being able to use my text editor I’m a terminal is just a nonstarter for me. It’s immensely convenience, and some workplaces can’t even handle local development because of monorepos.
5. IDEs rob you of the knowledge of the underlying tools, locking you into an ecosystem that you can’t take with you all the time.
6. I’ve worked in plenty of places that didn’t have JB stuff available. Never worked at a place you couldn’t use vim or emacs.
My philosophy on text editors is that you can get 90% of the benefit of an IDE with way less tooling. I use vim, ALE, and fzf/rg and I’ve never felt the least bit slowed down by my tooling.
But really, if an IDE is allowing you to write code you can’t understand without the IDE that’s a bad thing. Forcing yourself into a situation where you have to compartmentalize is going to make you a better programmer and lead to better code.
Navigation tools are essential especially because the (vastly) larger audience is uninitiated.
Do you hate the idea of being paid for the code you write? Or it's only other people's software that should be free?
>2. IDEs tend to be optimized for the write path, when it’s the read path that really matters.
Actually it's the exact opposite. Plain editors like vim are optimized for the write path -- they have all kinds of text manipulation options built in, and they're 99% about text manipulation. IDEs have code navigation options, source control integration, refactoring tools, even collaboration options in some cases, all of which are about the read path.
>3. Most of the people I admire at work and in open source don’t use an IDE.
Some of them use an Emacs so loaded that it's just a Greenspun-10th-rule-style re-implementation of an IDE. And tons of great developers use IDEs, from Notch to John Carmack, to Ken Beck. If you constrict the "people you admire" to old school UNIX neckbeards, then sure.
>4. Not being able to use my text editor I’m a terminal is just a nonstarter for me. It’s immensely convenience, and some workplaces can’t even handle local development because of monorepos.
Monorepos have nothing to do with local development or not. You can export a specific project out and work on it. Googlers use all kinds of IDEs, despite the monorepo.
Not being able to run in a terminal being a non-starter, is dubious, as 99% of the time a programmer doesn't need to work in a remote shell session, exception if they're a sys admin.
>5. IDEs rob you of the knowledge of the underlying tools, locking you into an ecosystem that you can’t take with you all the time.
Take with you were? You aren't going anywhere. You can always use an IDE whatever the circumstances. People use IDEs in games, data science, application programming, embedded development, aviation, web programming, etc. Not sure which realm would prohibit the use of an IDE.
>6. I’ve worked in plenty of places that didn’t have JB stuff available. Never worked at a place you couldn’t use vim or emacs.
There are tons of companies that mandate a specific corporate-decided development environment for all programmers, where you can't just "have vim or emacs".
[1] https://github.com/JetBrains/intellij-community
I'm a user of JB products for some years now. last week I had to open eclipse again, which is probably THE foss IDE for java. Developing in it is mostly fine, but if you get to the tiny details then IntelliJ is far better (multiple Carets, a proper dark mode, good vcs integration, UI is to the point and not as crowded as Eclipse, integrated terminal, etcpp).
I think that it shows the limits of (pure) open source and that the paid model really can polish a complex product. I imagine that contributors and drive-by patches are not really able to create a coherent product which maybe needs a proper plan.
(This really reminds me of that Bazaar vs Cathedral argument)
https://github.com/JetBrains/intellij-community