No Flutter on Linux desktop(news.itsfoss.com)
news.itsfoss.com
No Flutter on Linux desktop
https://news.itsfoss.com/no-flutter-on-linux-desktop/
81 comments
Not a Flutter-fan myself, but this argument is flawed:
> As we have touched on previously, there are already enough UI toolkits available for the Linux desktop. Most notably among those are Qt and GTK
The aim for Flutter is not to just support the Linux desktop, but also Mac, Windows, iOS and Android, tablets and phones.
Something tells me you won’t find neither Qt nor GTK a popular option for delivering cross-platform applications to all those targets.
> As we have touched on previously, there are already enough UI toolkits available for the Linux desktop. Most notably among those are Qt and GTK
The aim for Flutter is not to just support the Linux desktop, but also Mac, Windows, iOS and Android, tablets and phones.
Something tells me you won’t find neither Qt nor GTK a popular option for delivering cross-platform applications to all those targets.
It's also a bit of a short sighted, if verging on arrogant viewpoint. I thought the whole ethos of FLOSS (I did check the license which is a 3-clause BSD) was to have choice? Adding another UI toolkit to just a handful existing frameworks doesn't seem like adding to an already overcrowded space.
Sure Google might discontinue their interest but it can be forked and carry on if there's enough momentum.
Sure Google might discontinue their interest but it can be forked and carry on if there's enough momentum.
>It's also a bit of a short sighted, if verging on arrogant viewpoint. I thought the whole ethos of FLOSS (I did check the license which is a 3-clause BSD) was to have choice?
The ethos of FLOSS is to be able to change software and give it to your colleagues and friends. Gnome -- one of the major competitors to Flutter for Linux desktop programs -- is famously an example of a choice-unfriendly floss project.
The ethos of FLOSS is to be able to change software and give it to your colleagues and friends. Gnome -- one of the major competitors to Flutter for Linux desktop programs -- is famously an example of a choice-unfriendly floss project.
Then pick something else. I binned Gnome after they did all that Apple like dock nonsense and went all Fisher-Price. Today, and for a few years now, I've been running LXDE. I did try KDE for a couple of years (Plasma etc) but we just didn't get on. LXDE does just enough for a quality of life desktop experience.
Also I don't think I even mentioned Gnome. Also isn't Flutter just an app UI toolkit, ala, Electron, Qt etc? Just because there's a Flutter app I want to run doesn't mean I need to change my desktop manager.
Also I don't think I even mentioned Gnome. Also isn't Flutter just an app UI toolkit, ala, Electron, Qt etc? Just because there's a Flutter app I want to run doesn't mean I need to change my desktop manager.
It's not just about choice. This article is likely a response to Ubuntu betting on Flutter for desktop apps: https://www.omgubuntu.co.uk/2021/03/ubuntu-building-apps-wit...
A bet like that is putting a large part of Linux desktop app development in the hands of Google, which as a history of killing or deprecating their projects. Sure it can be forked if there's 'enough momentum', but that would suddenly dump a huge burden on the Linux ecosystem that it doesn't need. Why not stick with established players like Qt, who actually have a proven sustainable business model (dual OSS+commercial) for decades now, or a classic like FreePascal+Lazarus, which is also established for decades?
A bet like that is putting a large part of Linux desktop app development in the hands of Google, which as a history of killing or deprecating their projects. Sure it can be forked if there's 'enough momentum', but that would suddenly dump a huge burden on the Linux ecosystem that it doesn't need. Why not stick with established players like Qt, who actually have a proven sustainable business model (dual OSS+commercial) for decades now, or a classic like FreePascal+Lazarus, which is also established for decades?
I guess they could alternatively bet on Gnome; they certainly don't have a history of arbitrarily deprecating large swaths of API and functionality while spending most of their time chasing the dream of replicating whatever "new shiny" Microsoft has come up with. /s ;P
As for Qt, they have continuously pushed their OSS branch down on their priorities over time, with the latest (that I know of) round of changes from last year being that if you want to download official Qt binaries--ones that are at all supported by them--you have to have a Qt account; offline installers and LTS access are now commercial-only.
This is way worse than Flutter; and, as much as I hate hate hate Google in general and am a very big complainer about their shut downs of everything, Flutter feels more like Android or Chrome--both of which I feel will be safe for quite a while, at least as supported products if not ones that are mostly open source--than like all the other long list of things Google has killed (which certainly has included developer tooling).
As for Qt, they have continuously pushed their OSS branch down on their priorities over time, with the latest (that I know of) round of changes from last year being that if you want to download official Qt binaries--ones that are at all supported by them--you have to have a Qt account; offline installers and LTS access are now commercial-only.
This is way worse than Flutter; and, as much as I hate hate hate Google in general and am a very big complainer about their shut downs of everything, Flutter feels more like Android or Chrome--both of which I feel will be safe for quite a while, at least as supported products if not ones that are mostly open source--than like all the other long list of things Google has killed (which certainly has included developer tooling).
> I guess they could alternatively bet on Gnome
Yeah, except Gnome looks pretty crappy on every other OS (and only marginally less crappy on Linux). Meanwhile, Lazarus Component Library looks native on every desktop.
> if you want to download official Qt binaries--ones that are at all supported by them--you have to have a Qt account; offline installers and LTS access are now commercial-only.
I was just now able to download the installer for open source usage without a Qt account from https://www.qt.io/download-open-source , so that seems very questionable. By the way, of course Qt is not going to provide production level support for free for open source software, why should they be expected to?
Yeah, except Gnome looks pretty crappy on every other OS (and only marginally less crappy on Linux). Meanwhile, Lazarus Component Library looks native on every desktop.
> if you want to download official Qt binaries--ones that are at all supported by them--you have to have a Qt account; offline installers and LTS access are now commercial-only.
I was just now able to download the installer for open source usage without a Qt account from https://www.qt.io/download-open-source , so that seems very questionable. By the way, of course Qt is not going to provide production level support for free for open source software, why should they be expected to?
> A bet like that is putting a large part of Linux desktop app development in the hands of Google
Sure, but other Linux distributions are available. I'm not trying to respond in a facetious way, but since I started using Linux in around 1993-94 (Slackware), that was how you voted, with your feet, when something really annoyed you.
> Why not stick with established players like Qt, who actually have a proven sustainable business model...etc
Well, if you're a developer then keep using Qt, you're free to do so. Isn't this how it's supposed to work?
Sure, but other Linux distributions are available. I'm not trying to respond in a facetious way, but since I started using Linux in around 1993-94 (Slackware), that was how you voted, with your feet, when something really annoyed you.
> Why not stick with established players like Qt, who actually have a proven sustainable business model...etc
Well, if you're a developer then keep using Qt, you're free to do so. Isn't this how it's supposed to work?
They may not be popular, but Qt does support them all. See Qt for iOS https://doc.qt.io/qt-5/ios.html - it even supports watches.
> you won’t find neither Qt nor GTK a popular option for delivering cross-platform applications to all those targets.
Neither is Flutter, so that's not really an argument. And in case you're arguing that they don't support those targets, here's a reference of Qt's supported platforms: https://doc.qt.io/qt-5/supported-platforms.html
Neither is Flutter, so that's not really an argument. And in case you're arguing that they don't support those targets, here's a reference of Qt's supported platforms: https://doc.qt.io/qt-5/supported-platforms.html
>Additionally, Google doesn’t have a particularly good track record when it comes to maintaining projects.
This is not a great argument when talking about a Free Software project. Flutter is 3-clause BSD licensed, so even if Google abandons it, it's not like it suddenly traps all the applications built on top of it. If the toolkit code was unusually difficult to maintain, then sure, you shouldn't use it.
The article seems to hint that Google might open-core Flutter like they did with Android, but I see no evidence for this. Hell, even Chrome, technically an open-core project, is more or less entirely usable as Chromium (the Free part of Chrome) and projects exist specifically to remove all remaining Google services and telemetry (e.g. Ungoogled Chromium). If Google made Flutter practically unusable without other proprietary components, then yes, Google abandoning it does become a risk of using Flutter. However, I don't believe this to be the case.
Applications looking out-of-place on Linux is not something Google invented. Linux does not have a common look-and-feel, nor is there a single point for theming to occur. Even concepts like the window decoration, once exclusively the property of reparenting window managers, have been totally encroached upon by the window toolkit. So adding Yet Another Toolkit to Linux app development is not any more of a harm than the current situation already may be.
This is not a great argument when talking about a Free Software project. Flutter is 3-clause BSD licensed, so even if Google abandons it, it's not like it suddenly traps all the applications built on top of it. If the toolkit code was unusually difficult to maintain, then sure, you shouldn't use it.
The article seems to hint that Google might open-core Flutter like they did with Android, but I see no evidence for this. Hell, even Chrome, technically an open-core project, is more or less entirely usable as Chromium (the Free part of Chrome) and projects exist specifically to remove all remaining Google services and telemetry (e.g. Ungoogled Chromium). If Google made Flutter practically unusable without other proprietary components, then yes, Google abandoning it does become a risk of using Flutter. However, I don't believe this to be the case.
Applications looking out-of-place on Linux is not something Google invented. Linux does not have a common look-and-feel, nor is there a single point for theming to occur. Even concepts like the window decoration, once exclusively the property of reparenting window managers, have been totally encroached upon by the window toolkit. So adding Yet Another Toolkit to Linux app development is not any more of a harm than the current situation already may be.
>Hell, even Chrome, technically an open-core project, is more or less entirely usable as Chromium (the Free part of Chrome) and projects exist specifically to remove all remaining Google services and telemetry (e.g. Ungoogled Chromium).
It has become increasingly difficult to maintain forks of Chromium due to the sheer complexity and build time. Merging patches and testing them for each new release is a huge bitch due to the time required to compile (1h40 mins on a 3700X). You can't do iterative development, since most changes require recompiling from scratch, and forget about regression testing. It's only really feasible to do serious development on Chromium with either the Google-only tools, like Goma, or building/renting your own expensive build farm.
It has become increasingly difficult to maintain forks of Chromium due to the sheer complexity and build time. Merging patches and testing them for each new release is a huge bitch due to the time required to compile (1h40 mins on a 3700X). You can't do iterative development, since most changes require recompiling from scratch, and forget about regression testing. It's only really feasible to do serious development on Chromium with either the Google-only tools, like Goma, or building/renting your own expensive build farm.
> You can't do iterative development, since most changes require recompiling from scratch...
Huh, did they break something to cause this? I see no reason why this would be true for most things people want to work on (certainly if you edit files in "base" you are going to have a bad time...).
Huh, did they break something to cause this? I see no reason why this would be true for most things people want to work on (certainly if you edit files in "base" you are going to have a bad time...).
No, you are wrong. Maintainership, and the quality thereof, is an excellent argument when talking about ANY software project, Free Software project or not.
The point of Linux is to have choice. Flutter doesn't take any options away from users, it just adds an additional one, and competition is good. Extremely strange article
>Flutter doesn’t make use of the standardized Qt and GTK widgets. This means that Flutter apps look out-of-place on Linux, especially on GTK-based desktops
the point aside that this ship sailed a long time ago when Electron apps gained in popularity and people don't really have this expectation anyway (Qt and GTK apps also don't look the same), Linux isn't macOS. There's never been some sort of informal requirement that applications need to speak the same design language set by one single standard or organisation.
Linux users use software built at Google, Facebook, by open-source communities, standalone developers or whoever else, even proprietary software if they want to and Flutter is open-source for what it's worth anyway.
>Flutter doesn’t make use of the standardized Qt and GTK widgets. This means that Flutter apps look out-of-place on Linux, especially on GTK-based desktops
the point aside that this ship sailed a long time ago when Electron apps gained in popularity and people don't really have this expectation anyway (Qt and GTK apps also don't look the same), Linux isn't macOS. There's never been some sort of informal requirement that applications need to speak the same design language set by one single standard or organisation.
Linux users use software built at Google, Facebook, by open-source communities, standalone developers or whoever else, even proprietary software if they want to and Flutter is open-source for what it's worth anyway.
>Flutter doesn’t make use of the standardized Qt and GTK widgets. This means that Flutter apps look out-of-place on Linux, especially on GTK-based desktops
Isn't it a bit ironic to call this out as a downside when Qt and GTK apps generally look far out of place on operating systems that aren't Linux? It's really rich to say "this software isn't welcome because it feels out of place" when the preferred alternative feels far out of place (in the same way!) for multiple orders of magnitude more desktop computer users.
Isn't it a bit ironic to call this out as a downside when Qt and GTK apps generally look far out of place on operating systems that aren't Linux? It's really rich to say "this software isn't welcome because it feels out of place" when the preferred alternative feels far out of place (in the same way!) for multiple orders of magnitude more desktop computer users.
Are we talking about the same thing with regard to qt? My qt mac apps don’t feel quite as good as native cocoa apps, but they come pretty close. I certainly prefer it to electron.
It depends on a few things. With the "fusion" style, the basic widgets look good, but they look kind of off—definitely not native. Some widgets seem to have received little attention, like the calendar widget. The classic "rectangle with a dark to light gray gradient" button is a good giveaway. Dropdowns where the text doesn't get centered. Steppers next to text inputs that are the wrong size. Focus rings that are in the wrong place (compared to native counterparts).
Electron, at the very least, has Chromium inputs by default, which are arguably about as close to the native equivalents that you can get. Developers almost always restyle them, but you can't fault Electron for that.
Electron, at the very least, has Chromium inputs by default, which are arguably about as close to the native equivalents that you can get. Developers almost always restyle them, but you can't fault Electron for that.
what a pointless post, you didn't really make any real arguments here, you just complained that we already have stuff but didn't explain why you think it is redundant, because there are actually differences between all of these (Qt vs Electron vs Flutter vs etc) and then complained a bit about google, like why even?
Flutter is gaining serious traction because it relatively sane and *works well* in iOS, Android and (I think )Mac and Windows. You can actually make an application that looks decent and isn't a pile of shoehorned libs with an HTML engine.
I lost years waiting to have a decent way to develop desktop applications for Linux. After 20 years no one could get their shit together and both Mac and Windows ( yes, even Windows ) always had vastly superior desktop toolkits AND sane ways to package and deliver them to the mere mortals. Even when Microsoft seemed to fuck up everything they touched, Linux couldn't get their shit together.
There are many reasons for this and the other 2 had money to throw at this like it was free, BUT the Linux community never proved they can deliver a product, which in this case the customers were developers that have a job to do. A lot of lectures, morals and programing tricks, but never a product to be used by others. ( I'm talking exclusively about GUI toolkits here )
It's simply not in their nature and it's OK, but don't cry about when some evil conglomerate rolls up in town and gets all the action.
I lost years waiting to have a decent way to develop desktop applications for Linux. After 20 years no one could get their shit together and both Mac and Windows ( yes, even Windows ) always had vastly superior desktop toolkits AND sane ways to package and deliver them to the mere mortals. Even when Microsoft seemed to fuck up everything they touched, Linux couldn't get their shit together.
There are many reasons for this and the other 2 had money to throw at this like it was free, BUT the Linux community never proved they can deliver a product, which in this case the customers were developers that have a job to do. A lot of lectures, morals and programing tricks, but never a product to be used by others. ( I'm talking exclusively about GUI toolkits here )
It's simply not in their nature and it's OK, but don't cry about when some evil conglomerate rolls up in town and gets all the action.
I think you got things wrong because there is no such thing as Linux community neither a Linux community building a product for customers, in your case, developers.
What you call Linux community is actually several independent communities that build software for themselves and for others that fit in their community goals and the operation system that people usually call Linux is not a product but just a collection of independent softwares packaged by other independent communities to fit their goals and needs.
And these independent communities already proved that they are really capable to build and deliver high quality software that meets their goals and you can see their results in the most recent releases of Gnome, Plasma, Linux Mint and many many others.
If you need any software to fit your needs you should just find a community that has the same needs that you or find a company that sells what you need.
Or you can build your own sane toolkit and share it with your own community too.
What you call Linux community is actually several independent communities that build software for themselves and for others that fit in their community goals and the operation system that people usually call Linux is not a product but just a collection of independent softwares packaged by other independent communities to fit their goals and needs.
And these independent communities already proved that they are really capable to build and deliver high quality software that meets their goals and you can see their results in the most recent releases of Gnome, Plasma, Linux Mint and many many others.
If you need any software to fit your needs you should just find a community that has the same needs that you or find a company that sells what you need.
Or you can build your own sane toolkit and share it with your own community too.
But just look at Gnome, which is probably the most widely used Linux desktop. Until Gnome started Gnome 3 it was something like a nice replacement for Windows 95..XP. But then they crippled the UI, decided what the user really needs and removed all the rest. By taking pointless inspiration from touch devices and macOS they made things even worse. I really liked Gnome, but I've given up.
The other problem is that there isn't THE Linux desktop what makes building GUIs for Linux hard. Yes, I know, it is the year of the Linux desktop, but I fear it will not work out this time.
The other problem is that there isn't THE Linux desktop what makes building GUIs for Linux hard. Yes, I know, it is the year of the Linux desktop, but I fear it will not work out this time.
What you said is correct but thats what even the commenter means, he explicitly talks about the GUI side not linux community on the whole. Lets accept that GUIs built or made for linux are not the best in terms of UX.
I regularly find the linux software have the worst UI (GIMP top of the head). Unless its an electron app tho, when the UI is consistent. I also understand that its a community/solo developer effort. But, if linux wants to be a viable Desktop alternative, its time linux developers think from user perspective or bring in some UX designers to improve GUIs, or standardise a GUI toolkit/framework, which can be used to build upon by flutter or whatever. If linux wants to stick to server and Terminals then its solid.
Counter-opinion: Most programs I ever installed from my distros repository had a reasonable, good gui experience. I once set a gtk theme which I find pleasant to look at and configured qt to adapt to it as well, no idea how they do it but it works great.
As a non-professional user of gimp, never touched photoshop in my life, gimp does all I ever needed so far and never disappointed me. I edited some vacation photos, created a few humoristic images for banter with my friends, captioned some reaction pictures for online discussion. I don't understand why gimp is said to have such a bad ui, it's easy and efficient to use in my eyes.
Also yes, the ui of electron browser based "apps" is "consistent", but it fails to integrate with everything that uses the gtk theme that I like, disregards what my window manager tells it to do window-decoration wise and is sluggish when used with picom. In addition to that it makes my laptops fan spin loudly. Programs using electron provide a user experience much worse than even a badly done curses ui. Given that those programs are often provided by businesses where "ux-developers" may work, it's probably for the best if those people don't influence the gnu/linux ecosystem in the future.
You see, "linux developers" view their programs from a user perspective all the time. You may think they don't because you have some expectancy that the user has to be your average mom, dad or ten-year old child, but thats wrong. The programs are developed foremost with the developers interest in mind and may satisfy others who have similar intentions. "linux" doesn't want anything. The parent comment already explained that.
"Linux" left server and terminal exclusivity a long time ago.
As a non-professional user of gimp, never touched photoshop in my life, gimp does all I ever needed so far and never disappointed me. I edited some vacation photos, created a few humoristic images for banter with my friends, captioned some reaction pictures for online discussion. I don't understand why gimp is said to have such a bad ui, it's easy and efficient to use in my eyes.
Also yes, the ui of electron browser based "apps" is "consistent", but it fails to integrate with everything that uses the gtk theme that I like, disregards what my window manager tells it to do window-decoration wise and is sluggish when used with picom. In addition to that it makes my laptops fan spin loudly. Programs using electron provide a user experience much worse than even a badly done curses ui. Given that those programs are often provided by businesses where "ux-developers" may work, it's probably for the best if those people don't influence the gnu/linux ecosystem in the future.
You see, "linux developers" view their programs from a user perspective all the time. You may think they don't because you have some expectancy that the user has to be your average mom, dad or ten-year old child, but thats wrong. The programs are developed foremost with the developers interest in mind and may satisfy others who have similar intentions. "linux" doesn't want anything. The parent comment already explained that.
"Linux" left server and terminal exclusivity a long time ago.
I don’t think so. I feel that GTK is pretty boring and won’t last long. Sorry if you got offended.
Also, with the development of WASM most of the stuff we do on our desktop will move to browsers and we can use the best UI ever with CSS on browsers.
Also, with the development of WASM most of the stuff we do on our desktop will move to browsers and we can use the best UI ever with CSS on browsers.
> I feel that GTK is pretty boring
And for a software library, that's a good thing
> and won’t last long
Don't worry, gtk is 23 years old. It will last long enough for whatever you're trying to do, certainly longer than the javascript framework of the week.
> Also, with the development of WASM most of the stuff we do on our desktop will move to browsers
Nah. Webassembly can do some impressive stuff, but when you compile stuff to a binary blob you might as well do it for native code. Will run faster and feel better
> and we can use the best UI ever with CSS on browsers
My whole post was about why the browser ui is inferior to native applications. You should at least try to justify your position when writing up an answer in a discussion.
And for a software library, that's a good thing
> and won’t last long
Don't worry, gtk is 23 years old. It will last long enough for whatever you're trying to do, certainly longer than the javascript framework of the week.
> Also, with the development of WASM most of the stuff we do on our desktop will move to browsers
Nah. Webassembly can do some impressive stuff, but when you compile stuff to a binary blob you might as well do it for native code. Will run faster and feel better
> and we can use the best UI ever with CSS on browsers
My whole post was about why the browser ui is inferior to native applications. You should at least try to justify your position when writing up an answer in a discussion.
You give the tautological answer. All you say is true.
Parent was saying that the Linux model has failed spectacularly to deliver the software they needed.
Possibly there are categories of software illsuited for development by donated time from loosely coordinated volunteers.
Parent was saying that the Linux model has failed spectacularly to deliver the software they needed.
Possibly there are categories of software illsuited for development by donated time from loosely coordinated volunteers.
OP's comments apply just as well to the Linux communitIES.
> and works well in iOS, Android
from what I know, Flutter is still unusable on iOS because they haven't found a way to fix the extremely janky first load (after install, after upgrade, after reboot) animations.
This problem appeared more than a year ago when Flutter made the switch from OpenGL to Metal on iOS.
from what I know, Flutter is still unusable on iOS because they haven't found a way to fix the extremely janky first load (after install, after upgrade, after reboot) animations.
This problem appeared more than a year ago when Flutter made the switch from OpenGL to Metal on iOS.
> Flutter is still unusable on iOS
Flutter is actively working on the bug you're referring to, and I think you're valid in bringing this critique of the jankyness of animations on first run, but the large number of Flutter apps currently deployed clearly proves it is very usable on iOS. You have to workaround that jank, which is extremely frustrating, but I wouldn't write off the entire platform for it. It will be fixed.
Flutter is actively working on the bug you're referring to, and I think you're valid in bringing this critique of the jankyness of animations on first run, but the large number of Flutter apps currently deployed clearly proves it is very usable on iOS. You have to workaround that jank, which is extremely frustrating, but I wouldn't write off the entire platform for it. It will be fixed.
There are also tons of Cordova applications currently deployed and I bet noone over here feels in love with them.
If you're going to say that there is no "decent way to develop desktop applications for Linux", I'd at least expect a discussion about why you don't think Qt is such a way. My experience with it is that its an excellent cross-platform toolkit, has libraries already installed on many/most Linux desktops, and can run on Windows / Mac to boot. Qt is also not "a pile of shoehorned libs with an HTML engine".
Additionally, Qt has bindings most major languages. I have found the Python bindings (PyQt) to be quite good.
Additionally, Qt has bindings most major languages. I have found the Python bindings (PyQt) to be quite good.
>don't cry about when some evil conglomerate rolls up in town and gets all the action.
ok but also don't cry if they drop support for it after 18 months and your app dies.
ok but also don't cry if they drop support for it after 18 months and your app dies.
Exactly, and with all that risk Flutter still manages to win, that's how screwed up desktop Linux has been.
It's a FOSS framework built on top a FOSS language ( Dart). Even if Google "drops support" ( as if they support Flutter or Dart), there will probably be a fork, and even if there isn't, that won't stop all the already developed and deployed apps.
Furthermore, Google is using Flutter internally and some of their apps are getting rewrites in it ( like Google Pay); it's unlikely they'll drop it.
Furthermore, Google is using Flutter internally and some of their apps are getting rewrites in it ( like Google Pay); it's unlikely they'll drop it.
Blender is pretty successful product. But I agree, all UI toolkits that are available right now suck. Flutter may turn things around, not fan of it needing a new language however. There's reason why react and react-native are popular as well.
> works well in iOS
Many people would disagree with you.
Many people would disagree with you.
Hard to assess how much of this is Google's fault, or anything that can be dealt with by anyone.
Do folks have some examples of non-native iOS platforms that do have flawless no-lag 120Hz display?
Do folks have some examples of non-native iOS platforms that do have flawless no-lag 120Hz display?
It shouldn't be non-native, that's the problem. The best apps are always native ones. They make the most of the available APIs, fit in the OS the best both graphically and technically, and are the fastest.
If it has to be cross platform, something like React Native which leverages system controls would be a better approach than flutter which draws its own controls. But even there you will have to deal with a lot of translation due to it being JavaScript.
If it has to be cross platform, something like React Native which leverages system controls would be a better approach than flutter which draws its own controls. But even there you will have to deal with a lot of translation due to it being JavaScript.
Are you proposing that iOS might be so broken it's impossible to flawlessly copy a handful of screen-size rendering buffers at 120Hz? Because if you can do that, then it's possible to make a flawless no-lag non-native 120Hz UI.
I don't know if you can even access the screen at such a low level on iOS, I think for sure you'll have to go through a middleman to get pixels on the screen.
Sure, you have to go through a middleman. Does the middleman stop you from copying a texture to the screen?
Of course, there is much more to a good application than just how it looks.
There isn't any reason why an application can't look good _and_ have good functionality, reliability etc. As much as developers may like to assume that looks are irrelevant, they are for quite a lot of non-technical consumers/customers.
I know it's sacrilege, but you could build it in Electron. It is an HTML engine, but it is certainly not a pile of shoehorned libs.
> You can actually make an application that looks decent and isn't a pile of shoehorned libs with an HTML engine.
As if desktop OSes aren't full of "shoehorned libs" and non-native rendering engines among other cruft...right down to standard i/o interfaces that all processes use, which are based on the idea of a Teletypewriter.
> I lost years waiting to have a decent way to develop desktop applications for Linux. After 20 years no one could get their shit together...
That's funny because after using an Arch based desktop system (Manjaro) for a few years, I'd never go back to the Windows/Mac way of doing things on my main workstations. Imagine having to search the web to find a download page and/or get walled into someone's idea of a garden in order to find apps? Yuck!
And what apps exactly am I missing? Photoshop? Because when I moved to Linux every app I needed from Windows or its equivalent was already here - Chrome, VS Code, Beyond Compare, Android Studio, Gimp, Pinta, LibreOffice, Postman, Thunderbird, Slack, Spotify, Discord, OBS, FreeRDP, VLC Media Player and the list goes on... Barrier, a decent Notepad replacement (Mousepad - which is better than Notepad), MySQL workbench, Azure Data Studio (for SQL Server), etc. etc. etc... All that and my OS stopped getting in the way of me doing things like Docker, Node/NPM, bash scripts and many other such things that are bolted on and poorly supported in Windows/Mac.
I guess when you argue with people who think the Mac desktop toolkit is any good...well I don't expect to get anywhere here. I mean, the fact that you pretty much have to use Objective-C or Swift to even interface with them is already a huge detriment. Once you get past that you only have to deal with the normal missing features and anti-developer stuff that Apple throws at you. I'll take Win32 and C# any day over that garbage, but beyond that - I love HTML for desktop apps. The rendering engine behind Chromium is the most advanced in the world and it can do things that desktop toolkits couldn't dream of offering.
> BUT the Linux community never proved they can deliver a product.
The only reason a quarter of developers are on a Mac is because Linux laptop drivers aren't always great. It has absolutely nothing to do with GUI toolkits or any perceived lack of apps. It's so easy to build apps for every platform now with Electron - the vast majority of people don't care what their desktop apps are built on and that's why Electron is pretty much eating the world, not Flutter.
> It's simply not in their nature and it's OK, but don't cry about when some evil conglomerate rolls up in town and gets all the action.
Yeah, I'm sure people will complain that there's a new way to build apps especially if it's perceived as "better" than Electron (which I doubt it will be), but I guess we'll see. Currently I there aren't any popular Flutter apps and none of the desktop software that I see people use at work were built with Flutter.
As if desktop OSes aren't full of "shoehorned libs" and non-native rendering engines among other cruft...right down to standard i/o interfaces that all processes use, which are based on the idea of a Teletypewriter.
> I lost years waiting to have a decent way to develop desktop applications for Linux. After 20 years no one could get their shit together...
That's funny because after using an Arch based desktop system (Manjaro) for a few years, I'd never go back to the Windows/Mac way of doing things on my main workstations. Imagine having to search the web to find a download page and/or get walled into someone's idea of a garden in order to find apps? Yuck!
And what apps exactly am I missing? Photoshop? Because when I moved to Linux every app I needed from Windows or its equivalent was already here - Chrome, VS Code, Beyond Compare, Android Studio, Gimp, Pinta, LibreOffice, Postman, Thunderbird, Slack, Spotify, Discord, OBS, FreeRDP, VLC Media Player and the list goes on... Barrier, a decent Notepad replacement (Mousepad - which is better than Notepad), MySQL workbench, Azure Data Studio (for SQL Server), etc. etc. etc... All that and my OS stopped getting in the way of me doing things like Docker, Node/NPM, bash scripts and many other such things that are bolted on and poorly supported in Windows/Mac.
I guess when you argue with people who think the Mac desktop toolkit is any good...well I don't expect to get anywhere here. I mean, the fact that you pretty much have to use Objective-C or Swift to even interface with them is already a huge detriment. Once you get past that you only have to deal with the normal missing features and anti-developer stuff that Apple throws at you. I'll take Win32 and C# any day over that garbage, but beyond that - I love HTML for desktop apps. The rendering engine behind Chromium is the most advanced in the world and it can do things that desktop toolkits couldn't dream of offering.
> BUT the Linux community never proved they can deliver a product.
The only reason a quarter of developers are on a Mac is because Linux laptop drivers aren't always great. It has absolutely nothing to do with GUI toolkits or any perceived lack of apps. It's so easy to build apps for every platform now with Electron - the vast majority of people don't care what their desktop apps are built on and that's why Electron is pretty much eating the world, not Flutter.
> It's simply not in their nature and it's OK, but don't cry about when some evil conglomerate rolls up in town and gets all the action.
Yeah, I'm sure people will complain that there's a new way to build apps especially if it's perceived as "better" than Electron (which I doubt it will be), but I guess we'll see. Currently I there aren't any popular Flutter apps and none of the desktop software that I see people use at work were built with Flutter.
Photoshop and Gimp are worlds apart.
If you're not a power designer/user, GIMP is horrible to be part of your workflow.
Depending on the user( mostly developers). Linux in its current state is a NO for the people who want AAA Games and Adobe/Affinity Toolset. And many of the apps you mentioned are Electron Apps which are what commenter mentions about. Dont get me wrong but usability/readability of linux way apps dont come anywhere near other OSs because they're primarily designed by developers not design teams. I dont blame them. But to be a mass desktop alternative, linux has some climbing to do.
> Dont get me wrong but usability/readability of linux way apps dont come anywhere near other OSs because they're primarily designed by developers not design teams.
I think this claim harms your overall point (which I agree with). I can't think of any Linux programs (other than Calibre) which I have serious problems using. Furthermore, they tend to be extremely fast because they're very lightweight. The entire VLC package from my distribution is only 13 MB, and VLC is a rather large program by Linux desktop standards. Even Gimp is extremely simple and intuitive to use (to anyone who isn't a complete noob with graphics concepts), it's just extremely lacking in features and polish compared to Photoshop.
If you stick with the point that AAA games and Adobe are still a "no", you have a fine point. I've found that 99% of the games I want to play are actually available for Linux, these days, but that probably speaks more to my tastes than anything else.
I think this claim harms your overall point (which I agree with). I can't think of any Linux programs (other than Calibre) which I have serious problems using. Furthermore, they tend to be extremely fast because they're very lightweight. The entire VLC package from my distribution is only 13 MB, and VLC is a rather large program by Linux desktop standards. Even Gimp is extremely simple and intuitive to use (to anyone who isn't a complete noob with graphics concepts), it's just extremely lacking in features and polish compared to Photoshop.
If you stick with the point that AAA games and Adobe are still a "no", you have a fine point. I've found that 99% of the games I want to play are actually available for Linux, these days, but that probably speaks more to my tastes than anything else.
Gimp is certainly not the best, which is why I use Pinta for simple workflows. If there's some feature missing from Pinta, I use Gimp.
> ...usability/readability of linux way apps dont come anywhere near other OSs...
As a web dev - almost all of the apps that I needed from Windows were on Linux and the usability is exactly the same.
Personally, I don't care if Linux is ever used by the masses though. I'd probably prefer if it wasn't. I also don't care if AAA games ever come to Linux. I keep Windows machines around for that. I also have Macs for doing stuff with Apple iThings.
There is no obsession within me to do everything on the same machine. I don't need my phone to do stuff that my desktop does. My laptop is used for meetings and has almost no software installed on it. My Windows machines are used for entertainment. My Linux machines are for work. For devs, if you really, really need Photoshop you can run it in a VM. I don't see the big deal with that - it was a selling point for Apple's Intel Macs for a long time.
So, outside of AAA games (which devs largely don't need for work) and Photoshop - what am I missing?
> ...usability/readability of linux way apps dont come anywhere near other OSs...
As a web dev - almost all of the apps that I needed from Windows were on Linux and the usability is exactly the same.
Personally, I don't care if Linux is ever used by the masses though. I'd probably prefer if it wasn't. I also don't care if AAA games ever come to Linux. I keep Windows machines around for that. I also have Macs for doing stuff with Apple iThings.
There is no obsession within me to do everything on the same machine. I don't need my phone to do stuff that my desktop does. My laptop is used for meetings and has almost no software installed on it. My Windows machines are used for entertainment. My Linux machines are for work. For devs, if you really, really need Photoshop you can run it in a VM. I don't see the big deal with that - it was a selling point for Apple's Intel Macs for a long time.
So, outside of AAA games (which devs largely don't need for work) and Photoshop - what am I missing?
I mildly disagree with the article. “there are already enough UI toolkits available for the Linux desktop” - that is not my view of the Linux ecosystem.
I am not a UI developer, but a year ago I did evaluate Flutter and SwiftUI. I liked both, but it was so much easier using Flutter, and cross platform is not much more work.
I am not a UI developer, but a year ago I did evaluate Flutter and SwiftUI. I liked both, but it was so much easier using Flutter, and cross platform is not much more work.
Unless we want Electron for desktop apps, we need to give these a chance. We had others but they didn't take off. Java FX has gotten lost in the shuffle. Widget toolkits that are based on C ABI's are too low-level for most apps and we need something with a smoother cross-platform development experience.
I hate (sic) that some people attempt to dramatise what are purely technical decisions, like choosing the OS or the framework.
IMO Flutter competes with Electron, not necessarily Qt. Flutter’s key advantage is that it compiles natively and has less runtime bloat than Electron.
It does not offer platform-native UI widgets like Qt, nor does it have Qt’s breadth of features. But not every app needs that.
It does not offer platform-native UI widgets like Qt, nor does it have Qt’s breadth of features. But not every app needs that.
I’ll wait for someone to figure out react-native for linux instead
I think part of the problem here is that Linux doesn't really have anything 'native' when it comes to GUI. Even the most common options are split into GTK and Qt derivatives (of various versions too). Unlike pretty much every other OS that has one native GUI.
But it would be amazing if a react-native app would adapt to KDE or Gnome/GTK style widgets based on the DE used.
But it would be amazing if a react-native app would adapt to KDE or Gnome/GTK style widgets based on the DE used.
In general, developing gui is still too hard with too many trade-offs. Qt is so very large and requires C++ knowledge; or you can develop in python but python applications are hard to distribute. For other languages, the library bindings are often unmaintained because it is a lot of work or there are multiple choices and it is hard to know which one to use. QML is interesting but I tried to develop a csv viewer with it and found it didn't have all of the widgets that Qtwidgets has.
There are lots of other options, almost too many: Electron apps are good for web developers but they are also inefficient and use more memory than I think is necessary. Pascal and Lazarus seems useful but I don't really want to learn Pascal as it's not a popular language now. Vala could have been good but it seems to have lost momentum. There's Swift but it still seems to be mostly an IOS thing. There are also wxwidgets, gtk, fltk, ... None seem to really convince me that they make gui development straight forward.
Many other languages never really have the manpower to develop gui libraries or even polish their bindings to libraries like Qt and GTK, or there'll be 3 or 4 choices that are all half finished. What would be nice, would be to be able to easily develop gui in a modern compiled language that creates compact executables for distribution that look good on the linux desktop. I'm thinking about languages like go, rust or nim.
There are lots of other options, almost too many: Electron apps are good for web developers but they are also inefficient and use more memory than I think is necessary. Pascal and Lazarus seems useful but I don't really want to learn Pascal as it's not a popular language now. Vala could have been good but it seems to have lost momentum. There's Swift but it still seems to be mostly an IOS thing. There are also wxwidgets, gtk, fltk, ... None seem to really convince me that they make gui development straight forward.
Many other languages never really have the manpower to develop gui libraries or even polish their bindings to libraries like Qt and GTK, or there'll be 3 or 4 choices that are all half finished. What would be nice, would be to be able to easily develop gui in a modern compiled language that creates compact executables for distribution that look good on the linux desktop. I'm thinking about languages like go, rust or nim.
I like what Syncthing are doing: They just pop open a web browser window for the UI. It's like making an Electron app, except without having to ship an entire browser with your app. I have no idea why it's not more common.
Because that way you as a dev have zero control over the browser used, which can cause a variety of issues ( e.g. IE or Safari or an obsolete version of X being used, a browser extension causing problems) and an "app" has some advantages ( can run in background, can have toolbar icon, can be launched on startup, etc.)
> can run in background,
Syncthing runs fine in the background.
> can be launched on startup,
You can launch an app on startup regardless of what GUI toolkit it's using, or whether it's using one at all. Are you saying that you can only run a program at startup if it's using Electron?
> can have toolbar icon,
I think there's a misunderstanding here. I was talking about using the browser for the GUI of your app via an embedded HTTP-server, like Syncthing does; not replacing your entire program with a HTML-document with some JavaScript, which would clearly have some limitations.
> you as a dev have zero control over the browser used,
Well, that's also not true. You can check that the browser uses a certain user agent string, or you can add a specific version of a specific browser as a package dependency.
Syncthing runs fine in the background.
> can be launched on startup,
You can launch an app on startup regardless of what GUI toolkit it's using, or whether it's using one at all. Are you saying that you can only run a program at startup if it's using Electron?
> can have toolbar icon,
I think there's a misunderstanding here. I was talking about using the browser for the GUI of your app via an embedded HTTP-server, like Syncthing does; not replacing your entire program with a HTML-document with some JavaScript, which would clearly have some limitations.
> you as a dev have zero control over the browser used,
Well, that's also not true. You can check that the browser uses a certain user agent string, or you can add a specific version of a specific browser as a package dependency.
This is what I do for my home-grown apps. I write as much as I can in JS, and then provide a localhost API to access native operations like notifications or file management. I have one always-on background server that lets me execute arbitrary SQL on one SQLite file and use that as a shared database between all of my small apps and it works great.
Yeah if it absolutely has to be web-driven, I prefer to use my own browser and just have it as a pinned tab in Firefox. That way I can leverage its privacy protection too.
Why not use React Native which utilises a more commonly adopted language than Dart?
How's desktop support?
I dont want Flutter on Linux desktop because it will not integrate well. Flutter apps are almost certainly not going to use DBus, or other Freedesktop standards to fit in well, in the vast majority of cases. This isn't an inherent limitation, it's just that cross-platform apps generally tend to be lowest-common-denominator in nature. (It'll be interesting to see if Canonical/Ubuntu, in their adoption, do buck that trend & integrate well with Freedesktop systems.)
I don't want Flutter apps because it's highly unlikely Flutter apps will visually integrate with my desktop. I can pick GTK and KDE themes, and a good number of apps will fit the look I have picked, but with Flutter, the power & control seems to lie in the app designers hands, not the users.
So I have arguments against Flutter. I also don't like this article's arguments against Flutter. They seem shallow, petty, emotional, & highly non-technical. They seem like pointless griping. I worry the article is giving more credence to Flutter with it's shallow damning than it is dissuading anyone.
They argue that Linux doesn't need another toolkit. This status-quo-ism feels very pro-Cathedral-ism, anti-Bazaar-ism. Free software should be hunting for better ways to do things. There being existing ways to do things is not a good argument against trying for other.
The author argues Google does not have a good reputation maintaining software. They cite Google Glass and Google Play Music as examples. There's a lot of examples of Google ending support for things, true. Glass, I believe, is still sold to enterprises, and had a new hardware edition come out 2 years ago. This is hardly not supported. Google Play Music was migrated to a newer alternative application, and as a user of both, it's been, mostly, pretty decent a transition. In terms of the open source libraries they release on Github, generally they seem to be well supported & long lived. Google Guava has had it's 30th release. There's a vast number of projects that have been well supported for a really long time.
I don't want Flutter apps because it's highly unlikely Flutter apps will visually integrate with my desktop. I can pick GTK and KDE themes, and a good number of apps will fit the look I have picked, but with Flutter, the power & control seems to lie in the app designers hands, not the users.
So I have arguments against Flutter. I also don't like this article's arguments against Flutter. They seem shallow, petty, emotional, & highly non-technical. They seem like pointless griping. I worry the article is giving more credence to Flutter with it's shallow damning than it is dissuading anyone.
They argue that Linux doesn't need another toolkit. This status-quo-ism feels very pro-Cathedral-ism, anti-Bazaar-ism. Free software should be hunting for better ways to do things. There being existing ways to do things is not a good argument against trying for other.
The author argues Google does not have a good reputation maintaining software. They cite Google Glass and Google Play Music as examples. There's a lot of examples of Google ending support for things, true. Glass, I believe, is still sold to enterprises, and had a new hardware edition come out 2 years ago. This is hardly not supported. Google Play Music was migrated to a newer alternative application, and as a user of both, it's been, mostly, pretty decent a transition. In terms of the open source libraries they release on Github, generally they seem to be well supported & long lived. Google Guava has had it's 30th release. There's a vast number of projects that have been well supported for a really long time.
"I dont want Flutter on Linux desktop because it will not integrate well. Flutter apps are almost certainly not going to use DBus, or other Freedesktop standards to fit in well, in the vast majority of cases. "
On the plus side -removing the Linux-centric dependencies could potentially level the playing field for the BSDs. This would be a good thing because it would prevent Linux from becoming a Windows-style monoculture.
Meaning if Linux continues to go off into the weeds, then the BSD having a viable desktop that doesn't depend on the freedesktop foundation infrastructure will insure that desktop users aren't all left hanging (relying on windows or macos).
On the plus side -removing the Linux-centric dependencies could potentially level the playing field for the BSDs. This would be a good thing because it would prevent Linux from becoming a Windows-style monoculture.
Meaning if Linux continues to go off into the weeds, then the BSD having a viable desktop that doesn't depend on the freedesktop foundation infrastructure will insure that desktop users aren't all left hanging (relying on windows or macos).
https://github.com/search?q=org%3Acanonical+dart&type=
First result is a dbus library.
First result is a dbus library.
This is great to see.
It will be interesting to see if anyone besides Canonical takes steps to provide Linux hooks for their Flutter apps. It'd be very interesting to see Canonical also try to release some of their apps on other platforms too! That could be a thing!
Thanks for the link.
It will be interesting to see if anyone besides Canonical takes steps to provide Linux hooks for their Flutter apps. It'd be very interesting to see Canonical also try to release some of their apps on other platforms too! That could be a thing!
Thanks for the link.
Flutter most definitely deserves a place on the Linux desktop, if in fact you ARE in favor of choice. Concerns about it overtaking GTK and Qt are well-founded, because they are lousy to work with (and in the case of Qt, non-free).
If, as a developer, I build cross-platform applications that deploy across desktop and mobile platforms, I want sane tooling. Dart and Flutter are a step in the right direction, as is Go, another language birthed at Google.
If, as a developer, I build cross-platform applications that deploy across desktop and mobile platforms, I want sane tooling. Dart and Flutter are a step in the right direction, as is Go, another language birthed at Google.
Ironic how they are criticizing Google on an AMP page
I didn't realize that the Linux community was ruled over by a committee who can dictate such things. Where and when do I vote on illegalizing the use of Flutter in Linux apps?
QT, GTK, and all these GUI "toolkits" (love that 90s OOP fetishist term btw) are outdated technology. The future of apps is layout engines, not C/C++ libraries with community mantained bindings to high level languages.
All these XP era look alike apps like Clementine are going to be left for the minority of developers willing to mantain these old projects. Everyone knows that true Unix is CLI, and anything user friendly or consumer oriented will either be on the browser, or run in some render engine like Flutter. To me that's great, the future looks bright and cross platform.
All these XP era look alike apps like Clementine are going to be left for the minority of developers willing to mantain these old projects. Everyone knows that true Unix is CLI, and anything user friendly or consumer oriented will either be on the browser, or run in some render engine like Flutter. To me that's great, the future looks bright and cross platform.
> As we have touched on previously, there are already enough UI toolkits available for the Linux desktop.
What exactly is enough? I think this author clearly does not understand what problems Flutter is trying to solve or why someone might want to use it
> Additionally, Google doesn’t have a particularly good track record when it comes to maintaining projects. Just take a look at Google Glass and Play Music
Those are not projects, those are services. This is open source. If needed it can be forked or moved to different owners. Golang? Kubernetes anyone? Both are huge and started at Google
> Ever since the creation of Linux in 1991, major corporations have mostly left it alone.
Nonsense. Corps make a huge part of commits to the kennel and other projects
Unclear what world this author lives in
What exactly is enough? I think this author clearly does not understand what problems Flutter is trying to solve or why someone might want to use it
> Additionally, Google doesn’t have a particularly good track record when it comes to maintaining projects. Just take a look at Google Glass and Play Music
Those are not projects, those are services. This is open source. If needed it can be forked or moved to different owners. Golang? Kubernetes anyone? Both are huge and started at Google
> Ever since the creation of Linux in 1991, major corporations have mostly left it alone.
Nonsense. Corps make a huge part of commits to the kennel and other projects
Unclear what world this author lives in
Flutter on Linux has nothing to do with Google, cross platform development or choice.
Technically, Flutter is just another way of shipping apps, just like Electron and many other Electron-like alternatives.
Flutter on Linux just got attention because it was adopted by Canonical to build Ubuntu apps.
But it's just an illusion to think that Canonical chose Flutter because it is technically superior to GTK or QT.
Canonical relies heavily in GTK and Gnome for its products but has almost no influence over GTK and Gnome development when comparison with its main competitor, Red Hat. So they decided to use Flutter to escape from competitor influence or to have a technical advantage or both.
Technically, Flutter is just another way of shipping apps, just like Electron and many other Electron-like alternatives.
Flutter on Linux just got attention because it was adopted by Canonical to build Ubuntu apps.
But it's just an illusion to think that Canonical chose Flutter because it is technically superior to GTK or QT.
Canonical relies heavily in GTK and Gnome for its products but has almost no influence over GTK and Gnome development when comparison with its main competitor, Red Hat. So they decided to use Flutter to escape from competitor influence or to have a technical advantage or both.
> Warning: Undefined array key "apester_tags" in /srv/users/itsfoss/apps/newitsfoss/public/wp-content/plugins/apester-interactive-content/inc/qmerce-tag-composer.php on line 36
I dont think verbose logs must be printing on a web page when it's public. Or is it fine?
I dont think verbose logs must be printing on a web page when it's public. Or is it fine?