Dear JetBrains, Don't mess with your UI
neil.computer
neil.computer
> - Homogenization. JetBrains wants to be like VSCode
> - Animations. Haven't seen them but it will be filled
Seriously, HN? This is the quality that you reward with hundreds of points? Unlike the author, I happen to have used this UI for the past five months and it solves many of my longstanding gripes with IntelliJ:
- Gutters are now much better at highlighting local changes, while also being less visually obtrusive.
- The shaded tool window background nicely separates them from the editor
- Showing the git branch next to the title makes it so much easier to understand where I am, when I work with multiple projects and branches.
- Breadcrumbs in the Statusbar save some space and look much cleaner
---
Also, many of the complaints that I've seen in this thread are either wrong or exaggerated:
- "Monochrome icons" - Only the toolbar icons are monochrome, everything else still uses coloured icons.
- "The toolbar has less icons now" - It's customisable, you can put whatever you want in there. Yes, with colourful icons.
- "More Whitespace" - The vertical space from the top of the screen to the editor is the exact same. The Project view actually looses some vertical space, but only because the font size is bigger (Which you can change obviously). The spacing between the list items is the same according to a designer on Slack.
- "This is just some incompetent designer pushing their agenda on us, breaking my workflow because something something trends" - go touch grass.
And finally, just to be clear: no, there are no animations - everything is instant.
To be fair, this tendency to prioritize your own values above those of your users is not unique to designers. For example, Linus' insistence that kernel developers "don't break userspace."
Wait, what? Is that meant as a counterexample?
Maligning the devs? Not so much. Wanting them to show their work and explain why they think looking more like VSCode is a good idea? Very much so.
Not JetBrains, but I worked at a small app company that embarked on multiple UX re-dos for the vaguest of reasons. If your company has a charismatic VP of Design who has the CEO's ear, you're going to be spending your time on UX rewrites yearly. "It's not modern or sleek! It doesn't pop enough! We need to change our branding so it's fresher! Or, Users are confused with our current design! (without qualitative or quantitative user studies that show this)" It's never a measured attribute that they want to change. If you were a digital artist or interaction designer, you would have been able to build a great portfolio there in a matter of a few years.
Worked for a company that went full circle multiple times on its UI for what was ultimately a very large form between a single page with collapsible modules and tabbed UI layouts as the UX feedback oscillated between "Users don't know all these options are there and we want them to discover them because they keep asking us to implement features we already have" and "Users are overwhelmed by the number of options on screen"
I aimed for something that _wouldn't_ need to change visually with the times. UI Design is often the exact opposite: showcase that your property is on the bleeding edge of design.
I wonder what they'll do once they realize you can only round a corner so much.
I suppose it is human nature to randomly change things (see the other submission about gardening (continuous small changes are different from big changes time to time, still equally annoying)), also a result of the free market, there are too many UX people for coming up new products and make them final.
I bet if architects could change the buildings in minutes and for free, now we all lived in constantly changing (but obviously not improving) buildings.
Why does this surprise you when it happens all of the time? As a UX professional myself this is one of the banes of my existence.
Consider it's less often a "team of UX professionals" who make these decisions, and more often the manager of the month passing out a resume item.
But it is a giant problem that the industry has been "simplifying" everything so mindlessly for so many years that today my iOS device gave me a legitimate prompt that just didn't contain enough information to decide about at all.
Something along the lines of "Change in app administration [...] will you allow [...] to administer Gmail [...] Your data will be administered."
Now tell me: is this thing going to administer my private mails? 2 factor verification? (Yes, Google often use their mail app for that.) etc etc.
At this point it is utterly meaningless: the only correct thing to do with such a vague prompt is to say no until an explanation has been given.
But I can surely imagine UX developers patting themselves and each others on the back for such a brilliant simple design.
To be fair, good UX designers exist, but everytime I hear about ux-designers near any of my products I get a little scared :-)
It's not "just for fun". If the UX doesn't get updated, then they are out of a job.
I’d say they were doing it because they were getting paid for doing it. But that doesn’t mean it actually needed doing. What I have found in tech especially is that a product may get to a point where it’s meeting all its functional goals and really only needs small tweaks and fixes. But this is at odds with company growth. I’m not saying this is the case here but it’s prevalent in this industry.
I only hope nephew didn't mention Electron backend.
So yeah, fuck giving them the benefit of the doubt. Been there, done that, have the t-shirt.
If it turns out not to be 90% gratuitous changes it will be the anomaly, not the rule. They don't deserve that benefit of the doubt at this point.
Not fun, but giving themselves work.
Unfortunately, no UI team ever got rewarded for leaving a good product alone. They're hired to work, so they work.
Anyone old enough has seen the quality of certain products peak and then decline, after the company just couldn't leave well enough alone, and ruined excellence with needless tampering. I've hung on to various applications long after their EOL because the new versions were really worse. History is replete with great websites absolutely ruined by redesigns.
This field needs more of an old traditional craftsman appreciation for excellence.
I still miss Office 2003, it was so much better than new versions
It's not offensive to question it, and you trying to police people with legitimate grievance is disgusting.
And it's experience that says that they do. Maybe not for fun, but for another 3, 6, 9 mos of pay.
“I imagine it’ll suck but I haven’t used it, and they’re breaking the UI for the wrong reasons”.
Such a quick transition from imagining someone doing something and then getting angry at your own imagination.
Some anecdata, three times in my career, I have worked on massive 1M+ line of code Java projects -- single repo. The initial Apache Maven download of dependent JARs plus indexing might take one hour or so. After that: Lightspeed. Yes, I do agree: When you are forced to invalidate your indexes, the next restart is painful.
More anecdata: One of my current projects is Python 3+. The indexing is much slower and less accurate. Plus, we use lots of binary libraries (compiled native code), which seems to trigger indexing on each restart. But whatever, get a 5GHz machine, give IntelliJ 16GB of RAM and let it run. In 2022, are there any serious alternatives besides a Titanic-sized IDE for enormous complex codebases? I don't see any that average developers (myself) will use.
Last point to mix big IDEs with C++: Visual Studio (as much as a cringe writing anything positive about Microsoft) is a phenom platform. For years, it absolutely dominated C++ IDEs. IntelliSense was mind-blowing in C++ and force multiplier that was unmatched for many, many years. It is still very impressive to me. I am sure they feel JetBrains' CLion biting at their heels!
- "The toolbar has less icons now" - It's customisable, you can put whatever you want in there. Yes, with colourful icons.
I was wondering where the back / forward buttons were, but realised after a while that they can be put back in, just with some configuration.
Many read HN just for the comments, so will upvote a low-quality article with a discussion-driving headline (often without reading the article), purely to generate discussion on a topic they're interested in.
It's something of a pity as articles like this with absolute garbage takes do a disservice to those users who share the sentiment for valid reasons. But at least we get to read some of those more well-reasoned arguments in the comments here.
Furthermore, changes need to be evaluated against retraining millions of people with new UI. This part seems to be missing in the calculation.
This happens but stating that the Jetbrains changes fall under this category is baseless speculation.
> This part seems to be missing in the calculation
What makes you say that?
I don't think anyone commenting here is in any doubt that needless/bad UI changes do happen in products. Noone's debating that general broad point. The thesis put forward by the article is that Jetbrains' changes specifically are an example of this trend. And it's very clear that the article is so lazily researched that none of the points listed do anything to support that specific thesis.
But a professional tool needs to follow UI changes that are not justified beyond flimsy reasons? Used by millions to make a living?
The onus is on designers to justify their aesthetic/fashion instincts, to justify their tendency to churn user interfaces, the tendency to turn highly functional and dense tools into fisher price toys in the name of minimalism and infantilization users.
UI in general is regressed continually - completely agree with you there, yes.
The article however is specifically about Jetbrains. Jetbrains may be regressing their UI. Perhaps someone should spend some time looking at Jetbrains' new UIs and analyse whether it does in fact constitute yet another example of regressed UI.
This article doesn't do that - it makes a very lazy & exceptionally poor case.
More spanking is needed, there cannot be enough.
As someone who uses Rider, RubyMine, WebStorm (and sometimes PHPStorm) on a daily basis am worried that I now need to customize 4 IDEs (will they sync this config?). So until now I never wanted (or needed) to do that, now I have half a day work to get something I already have. I don't think that's progress.
Why not make it customizable, but let the default be the current state?
I enjoy the new monochromatic icons, I don't mind the extra gap, but I usually hide the panels and use shortcuts to access the panels.
I just wish they would invest more money into things that are actually a problem currently like how slow indexing is and how disruptive it is when you do a clean re-install and it indexes all your dependencies again.
I have tried spacemacs which gives a lot of functionality out the box with less configuration needed than DIY vim or emacs but it still just doesn’t compare to the ease of use of Jetbrains IDEs across languages (I use Java, Python, Ruby and Typescript every week).
I've been waiting for indexing improvemence for over 10 years, not holding my breath.
But for some reason, whenever I reopen a project, it needs to reindex it. Even when I open the same project every day, not having changed anything in between. I understand from a friend it doesn't do that with Java projects.
Sometimes it does. I have no idea what causes it.
You installed an npm package and it seems to depend on all the things? Let's start indexing on node_modules and give you no chance to even uninstall the **er. You used the CLI to work around our tooling? Hey let's start indexing immediately after this one finishes so it's a great idea for a coffee break, just never pull directly after coming back so you may actually do some work until we bog down all the cpu cores.
Don't get me wrong, I love all the JB IDEs, but indexing has been a deal-breaker for many.
I wish it bogged down all cores.. I work on relatively small Java projects (maybe 10 KLOC, 20 or so items in the pom.xml (although these obviously have transitive dependencies)). IntelliJ 2018 would re-index everything in a few minutes using all 20 vcores. I specifically spent about $8k on an iMac Pro back in the day to make this faster.
IntelliJ 2022, on the other hand, uses only 4 cores, and takes about 10 minutes. That's a stupid default (if I buy a powerful computer I want to use it..).
There is a setting to override it, but in my case it just doesn't work. CPU goes to approx 400% and just stays there. I filed a bug report, no movement.
Edit: I should say 12-thread, not 12-core, just to be clear on the performance level we're expecting here. It's a 6c/12t i7.
I haven't worked with Java in ~2 years so I'm not sure how bad it is now. I might start a NiFi index job just for fun...
I think their developers have no idea how it works either and just gave up some time after the useless shared indices came out.
That's both on my 16-inch M1 Max and my M2 Air with Go in Goland.
Found that while searching for the same problem
I did make an attempt to be the million line vimrc, t480, arch Linux guy (from reading HN) for a month or two but I found it was taking up so much time it was basically a hobby. I would rather just have something that works out the box (Mac + jetbrains) and be able to spend work and hobby time coding interesting stuff.
The fancy vim configuration just becomes more and more frustrating the more languages that you use and you start to run into weird things like when you start a new line in markdown the cursor doesn’t start at the same bullet indentation level so now you need to debug which plugin is messing that up.
Or if your Linux laptop doesn’t actually go to sleep (and goes flat) and PyCharm needs to index _again_ on start!
We are not fine with it. It really sucks. It's just out of our hands.
To this day I hate the gmail interface. It used to be so much better back in the day.
It was more pleasant to look at and provided more useful functionality on the UI.
The modern trend of wasting space and erasing colors is really bad.
The only new feature I like is the "tab" categories (but only on desktop). Though they really botched it because now I'm confused about the difference between categories and labels so I don't use labels anymore. IMO they should have used the existing labels system for that.
It was also pretty easy to find and read your email, which it currently is not.
I wish there was a better balance between both UIs. The "better looking" one wastes too much resources (CPU, memory, screen space).
It's not about resource cost for me, I use high end hardware and fast connections (coupled with ad and script blockers, both in-browser and at network level, which are probably adding to the frustration) and not only does it load and run like complete crap, I literally couldn't find an email I knew was there in my "inbox" last time I opened it.
I only use Gmail/Workspace for family/personal stuff so I just use mobile apps 99% of the time now, but if I had to use it more regularly I'd be quite annoyed.
Not that Office 365 which I use for work is any better in a web browser.
The "basic HTML" version is probably not exactly as any previous version but it's a good approximation. I think it makes much better use of color and space, and the buttons have text labels instead of using an arcane symbol language I have to relearn every time I have the misfortune to use it.
Actually I did use the web client one time recently and I literally couldn't find an email I knew I'd just received; my inbox is kept at 0 unread; so in theory it should have been impossible NOT to see it, right? After the initial confusion then annoyance came the realisation and I just went and found my phone to read it on.
> Nah I'm fine with it, because I never go to the website and just use mail
> clients; I can tolerate you its existence, out of sight out of mind and all that.
I'm neither fine with Gmail, nor with major changes to Jetbrains IDE which I've been using every day for years now, and totally agree with all of the points in the article (and by extension this thread).
> I'm afraid my British humour (sarcasm) may have been lost here.
I actually really like it. I've been looking at IntelliJ for 8 years, the refresh is really very welcome XD. I also haven't noticed any lost functionality.
https://mail.google.com/mail/u/0/h/
Still points to the "basic HTML" version, which is at least faster than the standard one. (I use it, and I'll keep using it, until I can, I guess.) https://support.google.com/mail/answer/15049
JetBrains is good for stuff I rarely do in the editor, like certain types of debugging, but otherwise feels too heavy, expensive, and hard to configure across installs. Finding a particular config option is always a hunt.
My dad was a vim user since the 90s and he ended up switching to IntelliJ a few years ago as well. It’s just significantly less hassle and has way more intelligent intellisense/autocomplete.
I used to meticulously maintain a vim config for the 8 or so languages I switch between, but eventually realised that for me the good part of vim is modal editing, and that IdeaVIM gives me that without having to fuss with it. My vimrc is now empty and it’s relegated to commit messages and fixing merge conflicts…
Do you know you can search for an option in the settings menu? Actually, you can search just about everywhere.
You can also set your IDE to store your config in a git repo so that it's the same everywhere you use it.
You should be debugging alongside development to gain deep understanding and reduce errors. No amount of "test" coverage can replace that.
I'm always skeptical about people who don't use IDEs and I'm almost always right about them. And it's not the good kind of being right.
Can it:
- assign (scenario, subsystem, level, description) to breakpoints and watches, so they can be turned on/off or selected in groups for debugging without removal?
- commit these scenarios of bps and watches into a repository?
- edit/read them in a textual form to patch or share over an IM?
- store traces of previous sessions and diff them to find out what changed?
Does anything even remotely similar exist? Most debuggers are just test-pause-and-inspect tools with “integration” in the form of eval on mouse hover. Their ux and dx sucks and the only reason I could find to use them were slow build times and too low-level runtimes, which are no more today.
Pretty sure most people avoid debuggers not because they are too hard to learn (they aren’t), but because they are too dumb to be useful, given that you can ‘if (cond) log.level(…)’ and restart instantly.
In Jetbrain's tools they definitely can. I can
* create a bunch of breakpoints
* give them descriptions
* only have them break on execution under certain conditions (conditions stated via the programming language being debugged, so `my_struct.some_func() == 5`)
* disable and enable breakpoints with a click of a button
* have a breakpoint only become active when another breakpoint got hit (and therefore only cause the debugging logic to execute in certain code paths
* Have a breakpoint not stop execution when it gets hit, which is useful because I can then
* Log when a breakpoint gets hit, including it's stack trace (so I can see exactly the flow that got me to that breakpoint)
* Evaluate some variable when a breakpoint is hit and log that evaluation.
All of this with an extremely simple UI that can be manipulated without any recompiles and while the application is running. I've gotten tons of value of debugging complex scenarios in both C# and Rust via this constructs.
So while I'm not going to convince you that debuggers are worthwhile, to call them "too dumb to be useful" is pretty naive.
The problem that debuggers solve is simply not a big problem anymore, in my experience. Definitely not worth the setup hassle when actively working on multiple platforms.
As a fellow debugger lover, having to basically use log statements instead of being able to reach in and debug was a rude awakening. Some people probably don’t know that debugging can be easy with the right combination of IDE and language. Or the billion software layers mean that the debugger fails to attach for some reason which can’t be figured out.
Debugging and using IDEs are entirely orthogonal. DTrace and lldb are far more valuable to me than a graphical step debugger, which I effectively never use. As it happens I use JetBrains IDEs full time.
One option is to use their sync service with your Jetbrain account, but you can also use git[1]
It definitely makes it easier to move machines when your settings and plugins are kept in sync with it.
[1] https://www.jetbrains.com/help/idea/sharing-your-ide-setting...
Apologies if you have accessibility needs and that’s the reason, but otherwise what is so unique about your editing requirements that requires such extreme customisation? I just use tools as they come. I try not to change anything. Less entropy in my life.
(Emacs, of course, has best in class discoverability via M-x and C-h, which allow you to find and introspect basically everything that's available to you, but then Emacs is an OS, not an editor)
It’s an uncomfortable feeling to see something new, and sometimes it genuinely is a step backwards, but I feel most of the time it’s progress and you get over it after a month.
I kid, I kid.
But not really.
I wonder how many percent of users are sticking with the old UI.
Even now when I look at Gmail 2004 I find it superior to the current UI.
It might look a lot prettier now, but it's much harder to actually do things from zero knowledge, I do fear we've regressed.
”New packaging, same great taste” the professional liars proclaim; as the box got emptier, the false bottom bigger, the air more voluminous, the filler lower quality. No, no. Worse things are in fact good for you. The party proclaimed it, so you are required to believe it. Or else.
Even more controversial is both the cause and the remedy, for which the overwhelming number of people, especially on a site like this, have absolutely no interest in facing; so I will refrain from expanding on that. It’s irrelevant anyways, because the consequences are as inevitable as they were clearly predicted and suppressed.
The abusive relationship that the western world in particular is currently in, remains largely furiously defended, like any other abusive relationship the abused have not yet faced reality about.
That results in continued comments one can also see in these threads, proclaiming essentially how “no, no, degraded features and lower information density against my will is actually a good thing”, which is no different than “I would not get beaten if I just made smarter decisions”.
The problem is precisely what you say: UI """designers""" don't make evidence-based decisons and don't do research or look for user feedback, they simply go with whatever they subjectively think best.
Don\t get me started on GNOME-iffication of apps.
Or stop the damm thing to run indexing, or avoid doing 10 finger combos, showing code problems without running inspections, doing Java compilation on save without using Eclipse's own compiler,...
> doing Java compilation on save without using Eclipse's own compiler If you are saving files with Intellij. Configure Intellij to save when focus is lost. I haven't pressed Ctrl-S in 15 years in Intellij. Regarding the automatic compiling whenever you save, like in Eclipse. I am really happy it doesn't do this. Because it means I can work slow compiling projects at all. I can work with refactoring features while my code is in the middle of larger refactoring and it doesn't compile. And so on. But if you prefer Eclipse's workflow, why not use Eclipse?
Incrementally compiling on save, is hardly noticeable versus continuous indexing.
Eclipse has incremental compile. Only parts you changed are compiling and it is not slow at all.
> I can work with refactoring features while my code is in the middle of larger refactoring and it doesn't compile.
That works in Eclipse perfectly well - except that code completion and suggestions are much better then the ones you get in Idea.
Especially in cases when much more important tasks are at hand, as you rightfully note.
Sadly it’s not actually just a UI refresh but a UI downgrade because they’ve actually removed key functionality, not just tweaked some visual styling.
It’s a completely different tool window layout mechanism with far fewer tool windows that can be open at the same time. I rely on having my source control, embedded terminal, problems window, commit window + more all simultaneously visible. It’s key to my workflow to have a number of key windows open and their new layout throws this all away.
So far its the most disruptive and user hostile “UI refresh” I’ve ever seen and will see me moving back to Visual Studio if they persist with it.
They dont, not compared to eclipse and java. They are more lightway and look nicer, but it does much less for you. (The default call hierarchy is just atrocious for example, you can see only problems from small part of project etc)
I'm not fine with it. It's still an accessibility DF for me, and I use HTML mode, despite it being semi-broken.
If they force this update, I'll just not update it. Not to mention, since I now will not be updating it, I can cancel my yearly, recurring subscription too.
Hey Jetbrains: You just have to do one thing. And that is: to NOT mess up existing thing.
Even if you stop launching new features into your product for next few years, I'll still happily pay you folks just because your product is so damn perfect as of now.
Please don't ruin it. Forget what VSCode is doing. They are far far far behind you folks. Don't let THEM ruin your product.
Except for remote development extension. Do that. It's the killer feature of VSCode.
I gave up on remote development with PyCharm and went with VSCode.
Based on your comment it seems they still haven't figured it out?
[1] https://www.jetbrains.com/pycharm/whatsnew/2021-3/#remote-de...
> Except for remote development extension.
I never really appreciated a remote dev server until I had to work on a behemoth Java project and then a behemoth C++ project. Having the extra memory and the extra cores is great. Never having to worry about your laptop heating up or having your battery die due to CPU load is even better.
I want to be able to hit Command+Plus and scale up the whole interface for presenting something to the team or when I have to code pair with someone who doesn't have as large of a monitor as I do.
Presentation mode is an absolute mess, and does not help with code pairing or code review presentations, and font scaling only affects the one file and not all the fonts, so the project navigation is still small and I have to repeat it for every file and again in reverse to reset it.
And the best past is, they could do it without impacting anyone else who doesn't need it, unlike this proposal.
Oh man I'm still mad that if I want to present my Jetbrains IDE screen and make it readable, I have to change 2 fonts, one for the editor, one for the interface.
Usually I get a lot of aggressive responses about how I don't understand presentation mode. I do, and it's crap.
(Replace shift shift with whatever you have the “everything” search bound to).
For presenting, it is very specific on how it is meant to be used, which is not how I would present during a screen sharing exercise.
For the others… well, it’s not called screen sharing mode or code pairing mode.
So you agree that it doesn't work that great for the scenarios that I've been talking about?
Sometimes, especially when I'm pairing with a junior that has a smaller screen, I want to scale everything because I want them to see everything. In this case scaling the whole UI is a benefit even if it is a "pro" application.
YMMV, and having alternative but consistent scaling options (such as all text in the file editor rather than piecemeal) would be a massive step up from the chaos at the moment. As already mentioned, the best part of this feature rather than a full UI redesign is that you don't have to use it!
Fascinating. Their flat, contrast-less and confusing UI is one of my biggest chafing points with their product. If I want to resize an UI element, I often have to make several guesses before I find the right border, let alone pixel.
It looks like they're copying VS Code's UI style which is a huge step backwards in my opinion. This UI minimalism looks great on paper but it's terribly confusing to actually work with and it can take weeks to get around it through sheer brute force and muscle memory.
I agree that their UI needs an overhaul but this isn't the solution.
The mental overhead in trying to learn and remember what to click to achieve what I want is exhausting, and somehow, in far too many apps, there's no way to look-up what the shortcut keys are.
- Increased margins/padding: At first, I thought it was a waste of space - but things are now easier to click on. I realised this when loading an older version on another machine. I'd still like to be able to tweak this - the document tabs are too padded.
- The main menu was hidden inside a hamburger menu. This was nasty - fortunately there's a "Show main menu in toolbar" option. This really should be on by default for non-mac.
- I wasn't keen on the darker colour scheme, I switched back to the old dark - it took less than a minute to learn I actually prefer the new one.
- Can only have 4 docked tool windows open at once (2 side, 2 bottom).
- Toolbars are easier to customise, but I wish I could combine the navigation bar into it, as I only have a few buttons on the right.
- The bright blue "run/debug" toolbar is distracting
- I've not noticed any removal of colour from the menu icons, and there are no animations
Overall, there isn't much to be upset about - took me minutes to get used to, no nowhere near as bad as the Visual Studio 2013 UI refresh.
You can still have them floating, then whatever layout you want - but it's not the same.
This is the worst "improvement", by far.
Also one of the reasons I'm not that thrilled about Fleet. I get it's easy to just in general be negative to change. But these are professional tools. I've used it for over a decade. I don't need it to be dumbed down.
Right now I'm comfortable and never enjoyed VSCode and others. But if I'm forced this change, what's stopping me then from trying a completely different IDE?
Will Fleet replace any existing JetBrains IDEs?
No, it will not. We are fully committed to continuing active development of our IntelliJ-based IDEs. With Fleet, we’re aiming to offer an alternative view on how an IDE can be organized, and this is impossible to implement within our current product line without going against the expectations of developers currently using those other IDEs. This means Fleet will co-exist with our established products.
Additionally, Fleet heavily relies on the IntelliJ code-processing engine for its smart code-editing features, such as project-aware and context-aware code completion, navigation to definitions and usages, refactorings, on-the-fly code quality checks, and quick-fixes.This means it will take resources away from the development of the established products. You cannot have the cake and eat it.
> It is not my our job to micro manage their dev resources.
While Fleet is a new product representing our attempt to rebuild the entire IDE from scratch, the new UI is a redesign of the existing JetBrains IDE product line.
Another reason icons suck is that different apps use the same or similar icons for different things, my memory just doesn't "click" with simplistic icons.
The old UI is information dense but easy to work with. It is one of the reasons I went to and paid for Jetbrains on my own dime, rather than using VS Code like others at my workplace.
The new UI looks nice but it won't work well for me. If they want everything to be so similar to VS Code then I might as well use that.
The current UI has text (!!) tabs instead of bland non descriptive icons. It has toolbars and menu items that I can configure.
I understand that the current generation of new developers are intimidated by so much raw power at their hands ;) But we have got to stop dumbing things down. It is killing a lot of experience and knowledge and will come back to haunt us as a society.
Jetbrains is currently THE reference when it comes to professional IDEs, changing what they’re doing will not go well for them :-(
The new generation is fucking things up for us. Text > icons.
However, I do want to point out that I don't think the author has actually used the new UI considering they mention the overuse of animations, considering that the new UI doesn't have _any_ as far as I can tell.
I mostly use keyboard shortcuts and when I don't know the combination, I rely on the Search Anywhere (double shift) and Run Anything (double ctrl) functionalities.
I find the new UI less distracting and the new VCS menu is a nice addition.
I also wonder how much technical debt has accumulated in the previous UI that is being addressed by a refresh.
That, happens to be exactly what I want (icon + label for most things).
I'm a very visual and a design oriented person and I want my tools to look good. If I'm going to spend countless hours using a product/tool I want to enjoy it's looks as well. Too many tools in my opinion are designed by people who want to maximize productivity with very utilitarian and practical user interfaces (which is great!) but don't seem to have enough care for making something that looks beautiful at the same time.
As with all things design, it's open to a large degree of subjectivity.
Beautiful instead of productive and utilitarian : absolutely NOT.
Jetbrains should keep working on their own thing instead of trying to change because of someone else. This holds true for people as well.
That said I was talking to a friend about how great it was and he was like "yeah all these things you describe seem like they were lifted from magit" so I suppose once again the answer is "magit is still the best".
Maybe if I ever have the motivation to actually set up an editor I'll try Spacemacs again but I get frustrated when things don't just work and end up going down the configuration rabbit hole anyway.
So they weren't timeless. If they were, then the appreciation wouldn't have been lost.
Anticipate it, create UIs that allow for large scale changes without breaking the old user flow and automatically prioritize hotkey-commands by user usage if there is a conflict. You also need to stop the Designers from going amok, rendering the whole tutorial section of your product on youtube obsolete. It will always be a paint point, but as you anticipated the bloat on success you integrated it into the architecture.
b) Phoenix Lifecycle it. When it all becomes to much, a smaller, more lightweight version of your product emerges out of the ashes, capable of integrating the newest, shiniest changes, without having the weight of all the bloat, it is first ignored, laughed at, finally emulated and it becomes your new product. There it slowly aggregates the same bloat, waiting for a challenger to appear. Its inevitable.
Is your IDE still a program, or is it in fact a OS for interacting addons and a really hungry text editor?
Why?? For more white space.
Software is getting very depressing as designers decide that everything needs to be shades of grey. It's really sad.
(Similarly, this also bothers me with the PS5 controller. The triangle, circle, square and X buttons no longer have different colors. This might not be an issue if you're already familiar with the button layout, but I bet it makes it more difficult to learn for somebody that is new to the controller.)
I’d love to see their IDEs re-platformed on Electron. I’m not ordinarily a fan, but VS Code has shown us how good an Electron-based IDE can be (while Atom showed us the opposite).
yes, if you invest hundreds of millions of dollars and spend hundreds of man years on trivial things like "we can't make terminal output fast because web tech".
Do not mistake a single success story achieved through insane effort and unlimited money with "this is the way".
IntelliJ simply could do C/C++ but they chose to push people towards a new IDE... C Lion. That was a huge blow for me, a polyglot.
Then I had a couple of bug opened with them... it took years for them to fix it. Except they marked it as fixed and it was indeed not fixed... Years later!
I kept paying until that point, I just couldn't do it anymore.
The company is a victim of its own success if you ask me. And that's a shame.
VSCode isn't perfect, but it does 90% of what IntelliJ does, and arguably 40% of it better. Also VSCode is faster which still boggles my mind as to how that's possible. I suppose the buttload of cash that Microsoft has thrown at it probably helps there... Since IntelliJ went through their "milliseconds latency blabla" years ago, it's gone downhill.
I imagine that retina pixel displays + java 2nd class child on the macOS platform = big oof (pre-metal).
I dream of a native vscode-like IDE...
https://github.com/intellij-rust/intellij-rust#compatible-id...
I understand the same goes for yet a third branch, which is .Net, and possibly a fourth, which is AppCode. But I don't have any contact with those ecosystems, so don't really know.
You had one job JetBrains -> speed up developers. I dont need half of the features available. Terminal does a better job at most of them and will never change.
But for the stuff I need and is used by millions if not tens of millions of developers.. just keep it working.
I feel like its slowly the time to move to VSCode. An army of JS Devs will keep on updating it.
They do a lot of smart engineering around the editor: https://code.visualstudio.com/blogs/2018/03/23/text-buffer-r...
I found the slack support for JetBrains to be absolutely fantastic. Super helpful and very technical. Having written plugins for both VSCode and JetBrains I prefer the latter but I have a heavy Java bias.
I agree that the multi-IDE strategy is problematic. They usually offer an option to just install plugins but with CLion they kind of dropped the ball. Which is surprising because I run their rust plugin on IntelliJ and it works well. Hopefully they'll address that too.
What if those remaining 10% are essential to your productivity ?
Like, will it autocomplete in my Django projects the way Pycharm does with models, up to their relations ? If so I may give it a try.
> The current UI will remain available for at least two years, and we’re not going to remove it until we’ve seen that the vast majority of our users have successfully made the switch to the new UI.
this sounds like you have to actively switch to the new UI if you already have an older version installed, and that they will be monitoring how many users have made the switch. Only after the "vast majority" (>80% ?) have switched will they "shove it down the throat" of the other 20%. Well, let's see how this promise will hold up...
> will remain available
Not to mention how many bug reports with be followed up with "Try in the new UI" as their only reply. We've seen this with Firefox, LibreOffice, KDE, Gnome, and so many other once-useful software.
"We won't force it for two years". We are gonna force it in two years.
"We won't until the vast minority" we will as soon as most people have moved.
Software developers as users: "Don't force me to upgrade"
The hypocrisy is tangible and close to mind boggling.
As Users: “I don’t want to upgrade to the new version. Don’t force work on me.”
It’s not hypocritical to want your cake and eat it too. It’s only hypocritical if you think it’s wrong for companies to force people to upgrade and then doing that. Not liking when your parents forced you to eat vegetables and making your kids eat vegetables isn’t hypocrisy.
(customer since 2010)
I smell the Reddit moment of "well, we see that you hate it so we won't remove the old one, we will just subtly break it in more and more ways till we brownbeat you to move. After all we can't say we wasted money on redesign to our bossess" moment.
they don't have to explain themselves here.
People need not justify the tools they choose to anyone, so correct, should that happen, people will not have to explain themselves.
It wasn't better in any way. In fact, most days I just missed WebStorm, but I soldiered on to see if I could find anything that was truly an improvement. At the end of my self-imposed, one-year trial, I concluded that VS Code had just one "advantage": it was free. Its barrier to entry was zero, from a pricing perspective. But since I had really given VS Code a fair shake, I gladly went back to WebStorm at work.
Fast forward to this year, when I ran head on into a serious WebStorm performance problem: type checking a very large TypeScript code base. The WebStorm logs show that the TypeScript language server is crashing, and I filed a bug with JetBrains many months ago, but zero progress. VS Code, on the other hand, has no problem type checking this huge code base—it happens near instantly. For months I have now had WebStorm open for editing, and VS Code open for type checking, which is ridiculous, and I'd be embarrassed if my team knew I was doing this.
It pains me greatly to say it, but perhaps the time has come to cancel my WebStorm license and just switch to VS Code.
I mean, why not, if JetBrains is just set on copying VS Code anyway? I'll take better performance with a worse UI, and it seems like that's precisely what VS Code can give me at this point. How sad.
Instead of doing automatic updates, I delay updating until the company forces me to choose between updating and not using the software--and only then do I make the hard choice. I wish that every software would provide security updates on a separate track from the usual "UX rewrite" and "Feature cram" downgrades. That way, I can stay on the oldest version (the one that works the best), yet get critical security updates so I can use it safely.
Web software, too! Reddit is the obvious example, but everywhere they're changing for the sake of change and usually making it worse. For example, Vanguard (the brokerage) recently did a site-wide update. They basically removed lots of critical functionality (including the only thing I actually visit their web site to do regularly) and replaced it with whitespace and fluff. For a while, they let you remain on the old site. Then they made it more difficult to find the link to the old site. Finally I got an E-mail saying that they'd charge my account a fee unless I "migrated" my account to the new site. Talk about end-user hostility. Now I just call them on the phone to do my rollovers, likely costing them more money. Good job, morons.
The problem here is that statistics, and feedback, doesn't inform about the negative. All the users out there who are happily using your software in its current state and are perfectly happy with it but can't be assed to take the time to fill out your survey. As the saying goes "you can't improve what you don't measure." This should have continued something like "so be sure to consider what you can't measure". You can't force everyone to take your survey.
Sure in part that puts the blame on the users who don't bother to fill out the survey. But that shouldn't also give the right to ignore them and what they might want. Perhaps a non-vote should be counted as a vote for no change instead of simply being ignored. It's pretty well known that the squeaky wheel gets the grease but that shouldn't mean the entire UI should be slathered in the grease the squeaky wheel wants but no one else asked for.
Since IntellIJ is available for many languages / platforms (Java, Android, Python, PHP, Rust, etc.) I'm seriously interested from which user group this is coming. Because in this context "dated" seems to mean: "does not look like VSCode". Since most Java devs I know don't even use VSCode I doubt it's coming from that user group.
Great analogy.
GUIs have no other purpose than providing ways to interact with the software and communicating those controls via our eyes. Buttons are spread across the window because that turns out to be very efficient at both showing users what common features are available, and giving them a way to quickly use those features without having to dig through menus.
You can replace the entire menu bar and a few other things with learning the shortcut to search Actions and the omni search shortcut for code & more. So long as those are retained in the new UI, I don't see the issue here.
Tangent: This to me is the same as the obsession with multiple monitors. You don't need it. Virtual desktops/workspaces do the same job in less physical space and let you actually organise your software.
It's like saying you don't need toilets when you can dig a hole in the ground and use leaves to wipe instead of toilet paper.
I also use a mouse on my PC, so I'm fairly precise in case I want to use the icons. There's no need to blow them up, so my fat fingers wouldn't misclick on a tablet.
I like to code with two windows side by side, usually IDE + browser for docs, sometimes on a small screen. I'd rather have the extra pixels for the docs instead of random useless space around the IDE's widgets.
Now I’m scared …. Is JetBrains changing the Pycharm UI? Why? For fashion?
Please no, this is the tool I use every day all day.
JetBrains if you’re listening… don’t fuck up the UI. I do not want a redesign. I do not want your IDE to be like VSCode. It’s a GOOD thing that JetBrains does not look and feel like VSCode.
I hope you are listening to your users.
I pay money for Pycharm … don’t change it.
A proficient user will have learned the keyboard commands for actions and navigation and rarely if ever use the visual elements. This holds for any tool.
Those who have yet to take this step, but are unhappy that the UI update is coming, should take this opportunity to step up their professionalism and master the tools of their trade. It will be well worth the investment.
Negative space is perhaps the one thing I might worry about, since not all devices come with giant or high resolution screens.
Otherwise I welcome the update, and understand that Jetbrains need to continue attracting new users, who are used to different UI then those of us who are older.
The problems with such UI overhauls is usually that the changes overshoot the target by a mile and the designers seem to focus more on visual design than usability. Animations are almost always too long, because their purpose - for example signalling that it has registered the input - is forgotten.
At this point at least web apps should just offer options to completely configure the UI layout with a JSON or HTML file and CSS.
I'm really excited about the new Jetbrains IDE and am looking forward to it maturing.
> Haven't seen them but it will be [...]. The whole interface is going to [...]. Mark my words
Oh really? Sounds like you've spent time using it and forming and informed & balanced evidence-driven opinion alright.
> Rounding of corners. Religion at this point. Homogenization. JetBrains wants to be like VSCode
Author has never used VSCode - not a rounded corner in sight.
But... I have seen a few commenters here on HN that seem more informed than the author and do prefer the old UI. If you're in that category, great. But I wonder are you in a niche minority. I'm a VSCode user but need to use IntelliJ & PyCharm in work as I support some custom IDE plugins & integrations. I hate the UI. I've tried switching to it wholesale on multiple occasions (company covers licences so no cost difference to me) but got quickly frustrated with the experience - especially that the discoverable (mouseclick) UI is always so many more clicks compared to workflows in other apps & that needs to be down for new users during the keyboard/command learning curve.
That's a subjective opinion on my part as a VSCode user, but I get the impression that it's shared by a lot of long-time (& probably now former-) JetBrains users. If they're bleeding users to VSCode, it stands to reason that the userbase that state behind are going to be the subset that prefer the old UI. So pushback is understandable.
I don't have a particularly strong opinion about the new UI, other than that it looks way too much like VSCode and way too little like the current one.
The main reason I use JetBrains stuff is that I'm familiar with the interface, it's messy but I know where to find all I need. VSCode is unusable without writing json config files and dropping into a terminal.
I always wonder why the keyboard jockeys are using IDEs at all instead of Vim. Why even use the Vim plugin instead of Vim? If optimal keyboard usage is your goal, Vim is all you need and want, it is the king.
Me? I like being able to open the IDE preferences and see the configurations for all the testing, virtual environment, or whatever crap I need to have set up to do my work. Typing text is easy, dealing with bullshit configs is the annoying part.
They've come into my messy table and threw all the papers into the ground. If I need to sort them all again I might as well not pay them for the privilege.
If you’re looking for a reason, you’d have found one regardless of anything else.
Of course I'd rather use the free alternative that nearly every language provides official support for. It's the job of JetBrains, who are selling a competition paid product, to make it worthwhile for me to keep paying.
As the value of their product decreases, and the quality of the free offering increases, at some point a small increase in productivity is not worth the added cost. It has to be sufficiently better.
It turns out, though, there is a lot of value in homogenizing your application skin. Users can recognize each other's applications, and help each other out. Whole categories of problems are avoided, like poorly written custom components that aren't PLAF compatible. I don't have the answer, but I find it interesting to consider the forces pushing for and against putting greater UI control in user hands.
This is pretty well documented in the UX world, it’s called Jacob’s Law[1] and it states that your users spend most of their time using other products, and they generally prefer yours to work like others they’ve already trained themselves to use.
I get that change always comes with resistance, but I will never understand the knee jerk reaction of “they’re morons and they have no idea what they’re doing” that is so common with change.
What's there to understand? In most cases they are morons that don't know what they're doing (and are not doing the change to improve users' lives, but to drive internal promotions for product people).
- Friend who was a UX researcher on Windows in the Windows 7 days
So, this makes me wonder about the timing for this and the technical architecture. Are they eating their own dogfood here? Or is this a separate thing?
To OP, I recommend downloading the version you like and sticking with it. You'll be good for years. Use this time to occasionally research alternatives.
Now I'm finding the key mappings (ex: Multi-line select) and functionalities (setting up your workspace a module) are drifting across the JetBrains products.
Attention is needed with these UIs to prevent drift and preserve homogeneity across applications.
The community will do the rest.
Even more flexibility would be cool as well in more specialized software, but I think FF does enough while not overwhelming the user with possibilities.
> The new UI will remain customizable. It is, and will be, possible to change many aspects of the look and feel, such as icons, colors, font sizes, and spacing between controls through theme plugins. We’ll add options as needed to accommodate user preferences, like the ability to customize the toolbar.
But on the other hand, it puts a bunch of things I regularly use into nested submenus for seemingly no reason (the top bar where they used to be is now mostly empty).
Caveat: I only used it for a short time, as neither Webstorm nor Rider work properly with the current EAP + new UI.
JetBrains is obviously becoming more like VS Code here so that they aren't edged out of the market. "Copy your competitors". Complaining as if we are paying customers is understandable because UI changes are annoying. However we have to see things from JetBrain's perspective. They're trying to stay alive in a market where their market share is rapidly decreasing being taken by VS Code.
The only thing that I can think of is that vscode is free and “good enough”. The integration for JavaScript and golang seems to be more “natural” in VSCode, somehow.
If you are using the free IntelliJ tools you get a crappy js editor when using the rust or Java specific IDE. Now I need to install another IntelliJ app.
Also I like being able to run vscode with a dev container. It allows new team members to instantly start working on a project and grt most things done.
Looking around me, I see many people who have to work on crappy corporate PCs where management figures VS Code is good enough, and it's free, so why pay for IntelliJ? And since the devs themselves don't seem to complain, they'll just use VS Code.
Rider is also stuck between a free VS Community version and VS Code which some people successfully use with the C# plugin (although I wouldn't use the latter for large code bases).
Remote development extension that makes development from remote machines, VMs, docker containers seem like it's your local machine.
I switched from CLion to VSCode for this, even though CLion is generally a nicer IDE for C++. VSCode is also better for memory usage.
- People don't know better. That is, they literally have no idea what a proper IDE is and what it does. Hence all the "VS Code is as good as IDEA" comments
- It's free. People feel entitled to have even the most complex software apps for free.
VScode being free and comparable feature wise surely is the main challenge.
It doesn't help that VScode is catching up and in many cases surpassing their engine for some languages. Looking at their poor typescript performance and longstanding python type hinting bugs. I've also heard VScode has much better docker dev container support.
I get the it's not necessarily the same resources working the UI and these things but still. If the UI drastically change to be more like VSCode it's just one more reason for me as an existing customer to change.
I am a paying customer because the UI is so much better than vs code. For my daily work vs code actually covers all functionality that I use, but the UI is shit so I use JetBrains. They will loose the customers that are already using it, betting that they will get more new one - a bad bed imo.
But I am a paying customer! I've been subscribed to the 'all you can eat' option for several years now.
And as a paying customer, I'm not thrilled about the UI change but I'll get used to it.
That doesn't make any sense. If they start looking like VS Code I would switch to the (free) VS Code as there is no point in paying for (almost) the same experience.
Why cant someone who prefers no change just use the old one and do no updates.
That being said, I really agree with the text that the design to be chosen should be based on A/B testing and whether the design improvement leads to an improvement in productivity.
Why are JetBrain engineers re-working on what was a pretty good UI and not actually working on broken internal things?
Just vim :-)
This trend of crying at every pixel change is honestly laughable.
Also you can install themes on Jetbrains' IDEs.
Except for the terminal window, it looks like an entirely different IDE.
It amounts to "Here are just some examples of the parties we might share your information with / reasons we might share your information but don't consider this a definitive list."
Also their autocomplete sucks. As in always inserts the wrong thing in the wrong place.
When you get a PM/Designer who insists on following the design principles they learned on YouTube, they tend to throw out the UI along with all the implicit knowledge accumulated within it. This is related to Chesterton's Fence https://fs.blog/chestertons-fence/
Also Fleet isn't replacing the other IDEs
The "Electron bloat world" editors have been much more snappy in my experience.
Typically the "java world" editors will lag out to the point where tapping on the "File" menu will take 3-5 seconds on occasion. Also they seem to rely on "indexing" and while the "indexing" operating is taking place, the entire editor just slows to a complete crawl.
I've seen this in Android Studio, IntelliJ, Tizen Studio, Eclipse - pretty much anything that's "java world". Apologies for the imprecise nomenclature, but this has been my experience and after so many years it has reached a stage of just low expectations from me whenever I see anything related to these IDEs.
Sublime is the only non-ancient editor I've used that is actually snappy, I can't perceive any lag. It feels like native software used to feel, yet looks half decent.
Vscode is slower, with both input and general UI lag. It helps to turn off hardware acceleration, but only a little. Luckily, you get used to it. Unless you go back to Sublime/Vim/Nano for a while, you stop noticing it.
Then there's the "you should be using"-ones...
* Visual Studio. Syntax highlighting takes a second or two when opening "large" files, input latency is bad, sometimes it takes longer to save a file than it does to compile your project, which now runs without your changes.
* XCode. Syntax highlighting works on Tuesdays, input latency is horrible. Randomly need to restart when certain features just stop working.
* Android Studio. Input and UI latency, needed 5 out of my then 8 gigabytes of RAM to load an empty project, had to close most other programs when working on a medium sized app, including my browser (Slack, Email, Teams etc).
* Goland. I've tried using this on both Windows and Linux, but I've never gotten anything to compile consistently. Indexing takes minutes, CPU usage randomly spikes, memory consumption constantly climbs... it just seems to eat resources doing nothing.
Sublime is the only (non-Vim/Emacs) editor I've used that takes my time, room/lap temperature and ergonomics seriously.
Are you still using a hard disk drive? I think I was last time I saw a delay that took that long on IntelliJ IDEA.
Actual professionals use keyboard shortcuts to easily navigate or bring to the forefront whatever they need in Intellij (the JB product I use)
If you are clicking through menus all day or are otherwise negatively affected to such a degree by a visual change like this, you are an amateur
So really, JetBrains is bringing consumer-grade UI practices to the world of... amateurs
This world for amateurs exists exists, as a crutch, within the larger world of the professional product.
However, the world of professional usage of Intellij remains essentially unchanged (granted, the new icons and extra negative space are annoying, they just are not significant)
TLDR: git gud, use a keyboard
Sprinkle in one or two for e.g. git branches and you get most of the speed you need.
I don't need to look at a splatter of information on my screen - I need code and a select few other pieces.