Flutter desktop shells
github.com
github.com
The alternatives for cross-platform GUI are
1. Electron
2. Frameworks like C++/Qt C/Gtk
3. Bindings to (2)
Using frameworks like Qt in native C++ sucks because C and C++ are just terrible languages for GUI development, even with the bolted on object systems and language extensions (Qt extends the C++ language).
Using bindings to those frameworks is nice in theory because you don't have to manually manage your own memory or use their type system, but in practice you still need to understand how their types bridge to your language's types and how their memory management bridges with your language's memory management (typically a GC). So while you're dealing with C types and manual memory management less frequently, when you do have to deal with it, it's more complicated than dealing with it in standard C/C++. So (3) is hardly better than (2).
Electron is probably the most cost-effective option, but it has its own problems (notably the high resource consumption that comes from bundling a whole web browser).
With those considerations in mind, learning a new high level language is the least of all problems, especially if the GUI framework is high-quality.
As someone developing GUIs with Qt bindings for Python daily, and has been for the past 8-9 years, I respectfully disagree.
Understanding of the underlying types and memory management is rarely an issue (e.g. once every 3 months), and when it is it's incredibly straightforward (e.g. maintaining reference to a window you've just created else Python garbage collects it for you).
Topics like these appear every now and then and remind me how there are still developers out there that still struggle to find the right combination of language and framework, some resorting to Electron because of assertions like this. This is a solved problem for my line of work - VFX and workflow related tools.
The styling system of qt is also quite limited. Their 'css' is a small set that does not work well on everything.
HTML+CSS allows for creating almost anything, it's the same on all platforms if the renderer is right. It's seperate from code and easily viewable and debuggable within a browser program. I have yet to see anything similiar with Qt, or a style heavy program like Discord made in Qt without becoming spaghetti.
But I also disagree that the styling is quite limited. Yeah, css in Qt is pretty limited, but creating a custom widget doesn't require all that much intimate knowledge (depending how custom we're talking). You can tweak the behavior of default widgets easily, you can draw a completely new one by painting and following the rules (for sizing, redrawing, etc). For new interaction models you'll need to be more familiar with the internals, but that's the case if you're stretching the limits of any framework.
Qt works really well for cross-platform apps that adhere to the OS' style and behavior. I think way too many apps waste time and frustrate users by reinventing the wheel. QML seems to be their approach to UIs untethered to the host OS (I don't have any experience) and seems to be their focus for the past handful of years. Their Qt (Widget) demos always show a bunch of cool features I rarely use, but I do pilfer their code for examples when I need something.
For a Discord like example; Maya, Nuke, and Houdini were all originally developed using their own UI toolkits and have migrated to Qt while maintaining each of their distinctive behaviors. They are all cross platform and require relatively low overhead and high performance with heavily customized UI behaviors.
Maya: https://damassets.autodesk.net/content/dam/autodesk/www/camp... Nuke: https://www.foundry.com/sites/default/files/styles/teaser/pu... Houdini: https://d2wvmrjymyrujw.cloudfront.net/media/uploads/products...
These apps did fork Qt5 when they moved from Qt4 to Qt5, but it was mostly Qt4 regressions and did it in a way intending to contribute their changes back into Qt mainline.
I see Electron devs being asked to match the slickness of an existing website with similar delivery time and productivity. You don't seem to be able to get close to that with native APIs, API frameworks, or higher-level tools without a lot of work. When you remove these UI and delivery requirements then almost all other frameworks seem to have a fighting chance.
From my perspective as a user, Electron apps are for companies who don’t really care much about quality and are too cheap and lazy to write native apps. YMMV.
Ironically, Microsoft was one of the early adopters of the idea of "web technologies for desktop apps":
About a year ago I began work on Label LIVE, an Electron app to interface with thermal label printers. It’s not perfect but it fills a niche. It wouldn’t exist if it wasn’t for Electron. Check out the video or download and give it a try. http://label.live
The whole reason OSes give you reusable components is so its easy to implement the conventions users are familiar with. Those are the things that are supposed to be fast and easy.
This likely violates the delivery timelines, but Maya, Nuke, and Houdini are all professional, heavily customized apps that were written in their own UI toolkits but were ported years ago to Qt and are cross platform.
Jupyter Notebook does this.
What's the reason, lack of discoverability? I'd rather run one Chrome instance than a dozen apps that use varying versions of Electron.
The dependency on Chrome is a turnoff for people who don't already use Chrome. An Electron app just depends on the operating system (apparently), so the install is much like other desktop apps.
PWA's are a sandboxed environment, where devs do not have control of browser version, cannot modify chromium, and cannot bundle necessary binary code. Even Spotify bundles in faster JSON parsers, and probably a number of other things.
An average user might not be able to connect the dots this way and realize what's going on with their machine. They'll just complain "my laptop is slow all the time and I don't know why". As developers, I feel like we have a responsibility to take better care of our users in this manner. Using a solution that will consume a GB or more of unshareable memory for something that doesn't need it (which is mostly everything) is a poor choice that puts the developer's needs (write-once-run-anywhere; faster, cheaper development time; fewer developers needed) over the user's.
So I fundamentally disagree that users only care that the app does the things they want, and nothing more.
Is this part of what Microsoft is doing with Chrome being pre-installed (via Edge)? I can't find a source, but I heard that using the existing chrome on the machine would reduce disk size for applications and possibly even allow them to share memory.
Though Qt Widgets in C++ isn't too bad if you use the designer instead of insisting on doing everything in code.
[0]: https://bitbucket.org/chromiumembedded/java-cef
[1] https://github.com/MGWL/QtE5
D's team is doing an incredible job of matching C++ compiler specific ABIs, name mangling/etc -- just so D users can use C++ libraries (including templates) directly from D without loosing the features of the language.
[2] https://dlang.org/spec/cpp_interface.html
Here is also an older video of Walter Bright (2015) talking about the challenges that D is tackling
It's a thing that solved a lot of very real problems that are still halfassed by NPM and Node ecosystem - package manager, build system, tooling around it - it was all a step above the quality of TypeScript/NPM/JS. Saner object semantics, good standard library, good tooling - it just got me to solving the problems I had. DOM wrappers it had were very nice, came with their own wrappers for Observable events (from standard library). They had proper async/await implementation working waay before TS.
Dart feels like a better TypeScript and for me the learning curve was really low.
One issue I had with it was using Dartium to debug the code - Chromium build that ran Dart VM naively to avoid transpiling to JS all the time - it was buggy as hell back then. Supposedly the dev compiler for JS is fast enough now that this isn't an issue - and in the case of Flutter you're running Dart natively so it's irrelevant.
I'm looking forward to doing a new project in Flutter this weekend - I've played with a hello world app when it was in beta and I was very impressed with the dev experience - hopefully it lives up to it in practice.
That's only if you use Typescript as a class-based OOP language. If you're more used to functional programming and prototypal inheritance, then Dart will bear no familiarity.
Due to its adherence to classical OOP, it feels way more like Java than Javascript.
As someone who never did any serious programing in those classical OOP languages, I struggled a lot to find the right abstractions to implement any kind of clean business logic. It was the first time I had to actually read up about "design patterns" and unironically consider writing factory factories.
It seems that due to Dart's rigidity and adherence to classical OOP, any library which needs to deviate from this paradigm has to hook into an additional build step (e.g. built_value for immutable types, has to use build_runner). That's an acceptable compromise for lib users, but adds a huge burden to actually writing those libraries, let alone incorporating those in your app's code.
You might need other design patterns, though.
I’m with him, I want to like the idea, but I think the adoption has been slow and it doesn’t appear like it’s going to suddenly take off. I could be wrong, I just don’t see any successful indications yet.
It's apparently the primary language for Google Ads. As such, I expect it to be maintained as long as Google continues to sell ads and if they do move away from it you'll have plenty of warning.
https://itunes.apple.com/us/app/google-ads/id1037457231?mt=8
https://play.google.com/store/apps/details?id=com.google.and...
That said, Adwords used to be written in GWT. It's not impossible to migrate, but quite a big job.
That's Dart, not Flutter.
And even so, Dart transpiles to JS for the Google Ads. They could switch to plain JS, or Typescript, or even move to a whole new implementation of Google Ads, and pull the plug.
They also used to use GWT, did that hold on?
From my point of view learning Dart/Flutter isn't really a big investment so the downside of it getting abandoned soon is not as large as the payoff if it actually delivers - quality cross platform UI framework with great dev experience that I can use on my projects.
If this is true, and if the other rumors about Fuchsia replacing Android are also true, then I doubt Flutter and Dart are going anywhere.
Heck, from what you say, it already got bored with Android -- if Fucshia is to replace it in say 2022, it means that Android just lasted around 15 years as a platform.
I'm no big fan of Google, but 15 years is plenty long enough. And really, Android is definitely struggling and showing its age by this point.
I have the opposite viewpoint on this. I've worked on mobile the last 5 years or so, and dedicated native Android the last 2.5 years. The last year or two have been dramatic revolutions in the Android developer experience, from my perspective. Things like ConstraintLayout and full Kotlin support with ktx have opened up a new world for me compared to the state of the art several years ago.
Then you look at companies like Peloton using large form-factor Android devices in their core products, and Samsung DEx using Android as a seamless mobile<->desktop computer.
To me it seems like Android is just hitting it's stride. Who knows though, it's Google.
I think it makes sense that it would go back to the drawing board and try to make something better from scratch.
Time will tell.
But we also missed a lot of good stuff that we used to have.
Compare a 1980 compiler with the optimizations GCC and LLVM make.
Or v8 with the best interpreter in 1990.
Or video codecs (almost non existent then). Or cryptography code (immensely improved). Or the capabilities of InDesign compared to a 1990 DTP. Or Photoshop compared to crude editors in the 80s. Or modern GCs versus crude 80s and 90s GCs. Or a modern DAW compared to 1990s midi apps...
Unfortunately, in 2019 cross-platform compatibility for native apps has only gotten harder (good luck supporting windows, mac, linux, android, and ios), which is why people tend to default to making web apps (or hybrid solutions using frameworks like React Native) rather than native apps, and this has perhaps also created a vicious circle by reducing interest in what cross-platform options there are (QT and javaFX).
Not sure why people keep making the claim that native is dying. More and more people are doing their computing on their smartphones and the vast majority of the top 100 apps on both iOS and Android are built in Objective-C/Swift/Java/Kotlin. It seems to me like the main people pushing the whole web app/PWA renaissance are web developers who resent the fact that the web is dying.
The coolest part was distribution. Under the File menu, you had an option to build a .exe file. I would build it and put the .exe file on 1.44 MB floppy disk, and share my little VB6 programs with friends. Such wonderful memories!
Most people just repeat Dijkstra's headlines without having even read his text, let alone understanding it.
I think we lost something important with the shift from workgroup to client/server (especially ODBC).
I banged out a lot of dBase II, FoxBASE, MS Access (in-house) apps, back in the day. I've not felt as productive since. I miss it dearly.
Can you imagine trying to explain to a developer in 1995 what you need to use to produce an app feature equiv in 2019
Edit: Also the continued existence of null in new programming languages is a baffling choice to me.
Dart is objectively a terrible language to be productive in for app development. The abstractions provided by the language are clunky to use for window elements on a device screen. Don’t get me started on Material Design, either.
Swift, Objective-C/C++, or even Java is way easier to start building apps with, in comparison. The tooling is mediocre for Dart as well.
> functional features like sum types and pattern matching
"Functional features" means different things to different people. Dart (like most modern languages) has a lot of the core functional features: first-class functions, closures, lambdas, higher-order functions. Our built-in collection libraries are heavily oriented around functional-style transformations. You don't need an external library to map() and filter() your lists to your heart's content.
At the type system level, we also have function types, generic functions, and even first-class generic functions, which is a really unusual, powerful feature.
Sum types are a slightly different beast. Sum types are basically a functional language's answer to subtyping and runtime polymorphism. But object-oriented languages already have full subtyping and polymorphism using classes. There's a sort of zen koan here where algebraic datatypes are a poor man's subclasses and subclasses are a poor man's algebraic datatypes.
The way most multi-paradigm languages like Scala and Kotlin handle this is that sum types are just syntactic sugar for defining a little class hierarchy. Likewise, pattern-matching becomes syntax sugar for instanceof checks and field access. I like that sugar and hope we can add something similar to Dart, but I don't find it's omission to be a profound oversight. It makes some kinds of code nicer, but doesn't significantly affect the expressiveness or capability of the language.
> Dart just feels like the same sort of OO language we've been getting since Java became popular.
Yeah. The original designers of the language designed something very conservative. I think they wanted to make a VM with certain features (single dispatch, static class structure, no static initialization, etc.), and designed the safest language they could come up with to let them do that.
There is a lot of benefit to familiarity. I like classes and C-family syntax, and we see very clearly that Dart is really easy for people to learn and become productive in. We've done user studies where participants have been able to write correct Dart code without knowing what language they were using. It's hard to underestimate the value of that.
But there is also value in providing the modern tools people want in order to write clean, beautiful, correct, maintainable code. Dart has some catching up to do there. We're making a lot of progress. With Dart 2.0, we replaced the old unsound optional type system with a real, modern, expressive, sound static type system. It was a ton of work to do that while dealing with millions of lines of existing code.
We didn't get all the type system features we wanted, but we have a foundation we can build on now. The optional type system had some nice properties, but was effectively a dead end. When your types are optional, you can't hang any language semantics off them. That takes lots of features off the table: implicit conversions, extension methods, etc.
> Also the continued existence of null in new programming languages is a baffling choice to me.
I have always believed [0] that not having non-nullable types was a mistake in Dart 1.0. We are fixing it now:
https://github.com/dart-lang/language/blob/master/working/01...
There's a lot of work to do, but I'm really excited with the design. Unlike many other languages, we have something that becomes fully sound with respect to null errors. This means that once a program is fully migrated, a compiler will be able to take advantage of non-nullable types for performance optimizations. It will be quite a while before we get to the point where we can do this, but it's cool that that's on the table.
[0]: http://journal.stuffwithstuff.com/2011/10/29/a-proposal-for-...
Sum types and exhaustive pattern matching aren't about expressiveness, they're tools for aiding code comprehension by increasing the locality of code that has no business being distributed into completely different classes, and decreasing the cost of making changes by heavily reducing the amount of test code that needs to be written to make sure a closed set of options is handled appropiately; in OOP languages you can get this by using interfaces but to make use of it you'd have to use huge classes with methods pertaining to every usage of this closed set.
I would never willingly pick up another language that doesn't provide me with them after experiencing the productivity gains. I really like the direction Dart is moving towards and the steps taken demonstrate there's a team behind it that cares about correctness and productivity over being just a familiar Java-like, but this is the one hard blocker for me.
That's true for some kinds of code but not others. This is the classic Expression Problem [0]. For some things, it makes sense to keep all of the code for a single operation together. For others, it makes more sense to keep all of the code for a single datatype together. ML-style languages optimize for the former, and object-oriented languages optimize for the latter.
In practice, for the kinds of code Dart is designed for, the latter is a better fit most of the time. There's a reason OO and UI have been married together for decades.
Ideally, a language provides both styles so you can choose the one that fits your problem best. You see that now with languages like Scala. I hope we get there with Dart too.
I don't think it's fair to say that subclassing and method dispatch is objectively wrong just because it's a bad fit for some kinds of code. (Though, naturally, if it's a bad fit for the kind of code you need to write, then an OO language might be an objectively bad choice for you.) Class-based method dispatch is annoying for some things (God knows I've written enough Visitor pattern implementations over object-oriented AST class hierarchies), but it's really beautiful for others.
Being able to define a new widget class that bundles its rendering and interaction behavior together and can seamlessly extend a UI framework is something so natural that we take it for granted, but is very difficult to express in a language like SML. In fact, in order to do it, you'll probably end up doing a "design pattern" that reimplements something like v-tables at the application level.
[0]: http://journal.stuffwithstuff.com/2010/10/01/solving-the-exp...
I do however disagree on subclassing being the best fit for UI code. "The Elm architecture" as well as the model presented by React functional components with hooks offer sets of tradeoffs that I at least have found are better for most of the development I find myself doing when writing (and particularly when maintaining) regular business frontends.
People repeat this a lot, that OO goes with UI, but I don't think it's actually true, and I think e.g. React Hooks and immediate mode GUI in general demonstrate that it's not true. OO has been coupled with UI out of inertia, not because OO has unique strengths when applied to UI.
Sum types are a primitive language feature, akin to product types, which absolutely no one rejects the value of. The sum/product analogy to arithmetic is compelling to me: they are the building blocks of complex types and deserve recognition. I think that "zen koan" you mentioned earlier is being very generous to subclassing. You want ADTs most of the time!
It is evidentially true. Thousands of successful applications and a billions of lines of UI code have been written in object oriented languages. It does work and it can't be that bad if that's continuing to happen even after the emergence of other alternatives.
Whether there are better ways is a good question, but I think it's pretty clear that you can ship good apps using OOP for your UI.
> React Hooks and immediate mode GUI in general demonstrate that it's not true.
I'm far enough into my career now — I've been doing UI programming of one form or another since the 90s — to have seen that pendulum swing several times. If there is a silver bullet, we haven't found it. It's probably not immediate mode GUIs because if it was, I wouldn't have seen game teams tear them out to replace them with something more retained several times in the 2000s.
What I think actually happens is that we forget the problems lurking in the solution we are not currently using. The grass over there gets greener and greener until we hop the fence and the cycle starts over. Incremental progress does happen. (I am not keen to revisit MFC any time soon.) But if a given concept (1) has been around a long time (2) has not already supplanted the alternatives, it's pretty unlikely that it is now an amazing solution today. The only time when that isn't true is when the surrounding technology context has changed since then.
For example, neural nets weren't a good solution for AI problems in the 80s because compute was too expensive and we didn't have a lot of data. Now that CPUs are cheap and everyone puts their entire life on the Internet, machine learning is here.
I haven't seen anything around UIs that to me looks like a significantly changed context, so I think we're still orbiting around retained-mode and immediate-mode as both having their own trade-offs and neither being a slam dunk.
> I think that "zen koan" you mentioned earlier is being very generous to subclassing. You want ADTs most of the time!
I really don't think that's true. Just look out there in the world. More code is written in languages doing subclassing every day than in languages with sum types. Despite the fact that sum types have been around since the 70s. You have to have a very uncharitable opinion of all of your fellow programmers to believe they've all been getting this wrong for decades. Heck, the software you are using right now to read this comment is sitting on a stack of several layers of subclass-based architectures! You've got JS running on top of the DOM inside a browser written in C++.
Sum types are really nice. But open-ended subclassing is too.
Compare that to Rust where compilation speed is almost as bad as C++, RLS is slower than e.g. Qt Creator's clang lints, and has auto-completions so innaccurate that they are almost worse than nothing.
Buuut.... I wanted to make a deep copy of an object (a map of maps) in Dart, and one of the suggestions on Stackoverflow is to serialise it to JSON and then deserialise it. Eek. In C++ you just do `auto a = b;`. In Rust `let a = b.clone();`. How do they leave out such basic functionality?
Sure, but does it do what you want? :) If that map contains pointers or other types with "interesting" assignment semantics, then your idea of a deep copy and that author of that type's idea may not be the same. Cloning is a surprisingly hard problem.
The two languages you compare to don't have GCs and prefer value semantics. But in most GC languages, everything is by reference and "copying the bits" isn't as meaningful of an operation.
I looked up how to deep clone maps in some other GC languages:
https://stackoverflow.com/questions/4157399/how-do-i-copy-a-...
Ruby: Top suggestion is to marhsall/unmarshall it.
https://stackoverflow.com/questions/5105517/deep-copy-of-a-d...
Python: Import a separate "copy" module. May have to implement some custom methods if you use user-defined objects in the map. The module isn't thread-safe. A comment recommends converting to JSON and back.
https://stackoverflow.com/questions/28288546/how-to-copy-has...
Java: No built in solution. Have to traverse the map yourself and do element-wise copies.
Definitely true that they may have done something weird, but I think the idea of a deep copy is pretty easy to define: It should not be possible to modify the original object using only the deep copy. (Note that this doesn't preclude copy-on-write, etc.)
I think it is quite weird that so few languages (even GC'd ones) provide a proper solution for this. It's clearly a thing people want to do.
https://medium.com/@hinchman_amanda/null-pointer-references-...
https://www.infoq.com/presentations/Null-References-The-Bill...
https://www.quora.com/Why-was-the-Null-Pointer-Exception-in-...
https://www.lucidchart.com/techblog/2015/08/31/the-worst-mis...
entity = Entity.fetch(id)
if (!entity) doSomething()
In Kotlin, for example, you mark something with a question mark to say that it can be null, which forces you to check for null before using it:
val entityOrPossiblyNull: Entity? = Entity.fetch(id)
if (entityOrPossiblyNull == null) {
doSomething()
}
else {
// The compiler knows that the variable is not null in this branch,
// so this assignment is OK.
val entityForSure: Entity = entityOrPossiblyNull
doSomethingWithEntity(entityForSure)
}Different languages solve this problem in different ways. Many languages get rid of null entirely, and use option types in its place. Other languages, like Kotlin, fix it in the type system, by differentiating between e.g. String and nullable String (spelled String? in Kotlin) .
I'm not sure I'd learn Dart apart from Flutter, but if you just treat it as part of learning the framework (think learning ruby along with rails, or Vue's random HTML-esque DSL, which people do), the overall learning curve is not high.
Dart is open source: https://github.com/dart-lang
It took me two minutes to find that using googles real proprietary technology..their search.
But this is true of so many widely used open source projects.
Literally nothing
Forks are nice for small patches or development teams with a plan, no one outside Google cares enough about Dart fwik.
I just started coding as if I was working with a version of Typescript.
I only Google from time to time for the standard library docs.
That was in fact the most pleasant things about Flutter for me, the language was a non-barrier.
... is this a comment from 1997 ? Qt is licensed under both GPLv3 and LGPLv3 - and Trolltech as a company hasn't existed for more than a decade...
And trolltech is very alive and publicly traded in fact. It has just been renamed to Qt company: https://en.wikipedia.org/wiki/The_Qt_Company
GTK+ and wxWidgets are other viable and popular choices for cross-platform desktop GUIs.
While Flutter doesn't work yet for desktop or web, ReactXP already works for these environments, and is used in the new Skype for Web.
Windows support is done and targets W10, Xbox, and Windows Mixed Reality. MacOS support is experimental, but mostly working. Linux currently needs an Electron wrapper.
There's not much of an ecosystem for multiplatform yet for desktop projects at all. For macOS you have a way of "making a framework you can use" and that's about it.
There is a lot of promise, though, in the mobile space: https://touchlab.co/touchlab-square-kotlin-multiplatform-col...
My sense is that as the mobile multiplatform space gets ironed out, the desktop space will benefit.
Though generally speaking, Kotlin/Native has a _much_ lower level of "institutional" risk, since the Kotlin approach doesn't subvert the control Apple or Microsoft have over their platforms. Flutter, on the other hand, seems to abstract it away.
Its also pretty slow compared with other native solutions.
If you want a debugger you need CLion, though. And the compiler is really, really slow – but they seem to be working on benchmarking things recently, so presumably it will be improved.
Check the IDE: (https://www.lazarus-ide.org/) and the cross-platform apps built with it: http://wiki.freepascal.org/Lazarus_Application_Gallery
Besides Ad Words, there is very little being sold by Dart.
Plus how much money would it cost to replace all the existing software stack?
I can relate to this. While that's always been part of my motivation, there were other things in play too, like wanting to do something "better" than before. Or using new methods. Or applying X to Y just because. Now, as I get older, I don't want to deal with overly complicated shit, excessive dependencies, or unnecessary bling.
As an example in the GUI space I keep wondering why GTK 4 is getting a scene graph. I've done GFX programming for years and scene graphs are almost never the right answer. Why does a GUI toolkit even need one? No GUI should be that complex. No GUI toolkit should provide one for applications - that's not their job. Don't get me wrong, GTK is my preferred toolkit (in spite of gnome), I just think people keep extending things for the wrong reason. If it works and isn't broke, don't change anything. If the world changed and wants multi-touch, fine, add that. But don't go around doing your science experiments in projects that lots of other people use as infrastructure or building blocks.
Change is constant. Real progress is slow.
Interesting, why would you say that ? I saw a bunch of projects go from "more-or-less smooth at 1080p" to "buttery 120fps smooth at 4k" by migrating to UI toolkits with a scenegraph architecture.
If you have the time to spend to build an app, learning Dart to do it will be the less of your concerns.
Sure, maybe in production you want to use JSX and/or TypeScript, but learning a new language and a new framework together is overwhelming, period. Pedagogically, it makes it hard to conceptually separate what is a language feature and what is a framework feature. (To this day, I still have trouble separating what parts are Ruby and what parts are Rails.)
I love working with it more than any other languages I've worked with.
Python, Java, Go, Javascript, Lisp...
I think there's some combination between it being typesafe, amazingly expressive, dynamic and also productive because you're on the web platform.
Generics might be a bit much for some Javascript people because it's like a whole language in and of itself but it's worth it IMO.
Flutter is the best-designed framework I've encountered (whether mobile, front-end, back-end or machine learning).
The only downside is Dart, and while it's annoying, it's also very minor.
Few exceptions do exist, where several companies support out of there need and importance.
I think Fuchsia and Flutter are Google's primary likely plans for the future of their consumer platforms. I don't really know why, it's just a feeling.
A lot of it comes from Google planning 10+ years out a lot of the time. Reader was a mistake to kill, but you could see how Google management thought it would not be a widely useful product in the long-term future (though let's repeat, they were wrong).
I think Flutter+Fuchsia has that 10+ year plan to it. I'm on board with Dart because it seems likely to be a major ease on development along with a likely promising future.
As such, I'll pass since this feels like a dead-end just like polymer was. It felt like polymer had the same hype cycle as flutter - was going to make development so much easier/faster/less error prone etc. It is now a semi-abandoned legacy hulk of a project that people hate working with (anecdote I know, but that has been my experience) because its totally different and alien to how people are used to working, for no/little obvious benefit apart from "this is from google! its got to be good!"
I wish Google would just accept defeat and support existing common tools rather than suffering from "not-invented-here" syndrome and create something totally different and non-compatible, rather than contributing to the existing industry de facto standards to make them better. There may still be a few "google-scale problems", but UI frameworks and HTML libraries are not them.
[1] https://github.com/flutter/flutter/wiki/Desktop-shells#c-wra...
Not to mention that Google might be a bit sensitive to anything JVM / Oracle right now, given the lawsuit they lost.
Had the chance to do some Flutter work, and really enjoyed it. Could knock out the UI and later modify it with ease. Spent most of my time focusing on app logic, and not tons of UI work.
Happy Flutter is spreading elsewhere, I will try this out and spend some time looking into Hummingbird also.
The other app I tried was some journaling app. Both of them used unconventional UI. Are there any good examples of native performing and looking Flutter apps?
EDIT: Apparently somebody downvoted this? What’s up with that?
https://github.com/flutter/flutter/issues/813
https://github.com/flutter/flutter/issues/28113
I wish the team would prioritize this more, but I guess that's Google for you.
FWIW I doubt it’s that Google doesn’t care about iOS.. it’s kind of the raison d’être for Flutter, no? Probably has more to do with it being really hard to reinvent a drawing layer from scratch on an OS you don’t control or whatever it is they’re doing. They could be trying their best and just be hitting a wall. Often happens with cross platform frameworks.
Realistically, I'd imagine that Google prioritizes less there because a lot of preference for iOS dev is to just go full native. I do wonder how much Flutter is used by Google for iOS apps though then.
There are a number of us dedicated to this on the team internally at Google. It's definitely a priority. Each platform has its own challenges though, and there's a very wide variety of issues that can arise that are platform specific. The more feedback we can get about what people want to do that isn't working, the more we can address it :)
I would really think that this "problem" would have a great solution by now. But I think most of the developer attention is on web for obvious reasons (more ease, more conducive to mistakes/novices/etc).
However I think there should be way more desktop apps. I hate using services and having my data all on some remote service w/ little options for privacy and safety/backup, etc. I also like the general speed of desktop apps (though I know the web is trying to pull on more desktop/os resources) and I like the way desktop apps can interact with the host OS (creative windowing/sidebar/docking/etc and utilization of system tray/menu bar/etc.)
I would much rather see my needs addressed by suites of standalone/private desktop apps than by a bunch of URLs and Saas's.
Interestingly, the HN zeitgeist seems to recently be turning its awares back to this topic (it seesm to me from recent HN threads). Is it? And if so, I wonder why? What's in the air now?
So much so that you literally have platform-specific widget implementations to go along with the platform-specific themes. Which sounds worse than it is, but defining new widgets is super easy.
Not a write-once-run-everywhere but more like a write-1.2x-to-support-2-platforms type of thing. It risks being uncanny-valley at times, and implementation can drift from platform needing re-alignment, but it should let you do those really platform-specific stuff in sane enough ways.
Chris Sells csells@google.com
My (perhaps idiosyncratic?) priorities in order:
- a ui programming model that is thoroughly thought-through. to be honest i think javafx fits this bill, it took long- hard-learned lessons from Swing but rebuilt things up from that learning. the scene graph api model imo is so ideal and then the way they build the widget/controls on top of that...creates a lot of flexibility yet remains simple
- actually works reliably (ie, is production-quality) on all platforms * this is probably where 100s of these xplat desktop frameworks fall down. many are just so far from reliability
- full support for building binaries (+ installers would make you really stand out! +, heck, streamlined app store deployment would be wow) for all major systems (mac, win, linux) ... ideally producing these all from one same build machine/os
- solid set of optimized core widgets (all the stuff you'd expect in any widget toolkit...and performant...eg, the table widget should be optimized for unlimited records etc.. again javafx seemed to get it right w/ their tableview)
- polyglot (again java/fx ... you can build in clojure, jruby, kotlin, etc etc)
Where I depart from perhaps the majority is that I don't care about native LaF... and I suspect striving for native LaF is sort of at odds anyway with the xplat goal. So in my mind, I'm okay w/ non-native.. of course I do want ability to pixel push and have things look nice and I want things generally fast. (javafx again is doing something cool w/ its ability to style things in its 'css' support...)
I can think of a lot of apps that don't use the native OS LaF and yet look great.. Spotify, for one, and there are many productivity-oriented apps in this list... Another.. I use Jetbrains IDE for developing in Java, Clojure, Ruby, Elm, Javascript, etc... and that app - afaik - is Swing-based and yet they have made a killer user-experience across the board imo..
I guess I just talked myself in to using JavaFX
Some issues (it's been some time so memory is a bit hazy):
- Lack of good documentation, both official and alternative (ultimately found a good reference book).
- Lack of customization on some widget, even compared to Swing (remember being really disappointed at how hard it was to modify the built-in tab-view).
- I really don't like the built not-quite CSS. Remember having trouble finding documentation for it. Also why a different language and not just code — like for the rest.
- Was really put off by some parts of the API design, even compared to Swing. Something that I remember is that there was a half-baked support for "theming" that isn't really usable nor used.
- This is much more debatable, and I won't wish ill on people who dismiss it, but it feels like it's completely out of the zeitgeist. I don't know if people doing unsexy jobs use JavaFX, but very people talk/write about it over the internet. Compared to Swing, the dearth of resources is staggering. It also created a lot of FUD that you can easily find if you google a bit. Don't use Swing cause it might be deprecated in favor of JavaFX (unlikely?). JavaFX hasn't really been updated nor hyped since Oracle took over, it's just rotting. IIRC it's not even bundled with the VM anymore, which is more headaches for distribution.
Only exception: https://tornadofx.io/ has some love in the Kotlin community.
- You have to link to different JavaFX jars for different platform, which makes any kind of Maven (etc) based dependency management a nightmare.
The premise is very similar (non-native UI on OpenGL/DirectX surface, AOT compiled language, C++ based runtime, near native performance) I would go so far to say Flutter is heavily based on AIR.
When I ran the current Flutter-desktop demo it appeared to have the stability and feature-set of AIR 1.0 beta which was released in 2008 (It is now v32.0 as of 2019). It's a great jumping off point but I don't necessarily understand the hype when I am seeing something I saw 11 years ago.
Features I would expect Flutter Desktop to provide to match AIR.
- Hardware accelerated Video
For multimedia apps
- WebView
This may seem counter-intuitive but a chromium/webkit based webview component is a must.
- read/write/delete access to files in the app's sandbox.
- saveAs, filePicker dialogues
- Multi window support
- Toast/Notification Support
- Extensibility
AIR has a C lib and header for passing vars between the AIR context and the native plugin. This means the conversion has limited overhead. We can pass bytearrays, bitmapdata and even call methods directly on the AIR vars from within native code.
Being limited to using JSON across the bridge is not ideal, as there is conversion overhead and also you lose type safety.
I have added bindings for Swift, C#, Kotlin in recent years for building plugins with these more modern languages. It also allows building extensions in .NET and not being restricted to C.
References
https://github.com/tuarua/WebViewANE
Not all platforms have the same user interfaces/paradigms. What is the Windows equivalent of the OSX dock? What is Linux's equivalent?
Most cross platform frameworks solve that kind of problem by letting you write platform specific code. The problem there of course is that you're just writing multiple apps again, often in a non-native or at least second class, language. (By second class I mean it's not what the OS authors intended their apps to be written in.)
Isn't this throwing the baby out with the bathwater? Only a very small portion of your code should be platform-specific (10-20%?), which is still a huge win over maintaining multiple native apps. Of course, in practice, there are many other disadvantages of using cross-platform frameworks that may outweigh the benefits, such as the inevitable impedance mismatch between languages, APIs, concepts, etc.
https://www.reddit.com/r/rust/comments/9bapwt/thoughts_on_wh...
I think the solution is to split in 2: A UI back-end with logic, layout, etc and a UI front-end renders in each UI native toolkit.
So when write Button(title="hello") is NOT a widget, but a struct with data. It must be pushed to UI.Coccoa.render(button) to show up. But not only that, is possible to say:
Form
when(toolkit=UIKit) Field(UILabel="UserName", UiTextBox(PlaceHolder="Write your user")
Field(Label="UserName", Edit(PlaceHolder="Write your user")
Also sounds like what wxWidgets already does.
You must buy the whole Package with react and others. You don't code normal UI on Swift and still be part of the react world.
Think is like build a JSON API. My idea is that. This mean that on the client, is full native (with some minimal coupling to the how transfer data).
ReactXP (https://microsoft.github.io/reactxp/) uses TypeScript and targets mobile and desktop platforms using native controls, not emulated controls. While Flutter doesn't work yet for desktop or web, ReactXP already works for these environments, and is used in the new Skype for Web.
Windows support is done and targets W10, Xbox, and Windows Mixed Reality. MacOS support is experimental, but mostly working. Linux currently needs an Electron wrapper.
How would non-tech users handle backups?
In this way you'd have the data storage decoupled from the app so the choice would be there.
Ah, but you see, the current crop of developers loves these things because you can make a lot of money when people have to pay you rent. They might talk a good game about open source and empowerment, but they're behavior says they'd rather make money.
Not trying to be harsh and I bet this project has cool plans.
But what worries me is that there are 100s of options in this space and _all_ of them are short.
Why can't the passionate community rally behind one effort, and fully solve this problem?
A solution doesn't seem to need to be programming language-specific, either, so that doesn't seem to be a good reason to keep 'forking' this problem.
I have looked at multiple different solutions how to hack flutter to run on desktop but didn't want to invest my time in them as they would probably crash in weird scenarios. This will make things so much easier! modern tooling + no electron = win for both developers and users.
Can't say much about this desktop targets, but on the mobile side, in terms of size and performance, they are much closer to native than react.
It brings React's view=function(state) model to native applications, along with CSS box model and Flexbox. It has the developer ergonomics of Electron, but without the need for its heavy runtime.
It is still under development and I'm looking forward to playing with it next time I try writing a GUI app.
If it's not using, for example, Cocoa controls on macOS, then how is it native GUI?
- Native code - does the GUI framework compile to machine code?
- Native widgets - does the GUI framework use platform widgets, or a platform agnostic layer?
In the context of Revery, we used "native" to mean "native code", ie, "compiles to machine code" - and further clarify in the README. It's important for us to make this differentiation - because often people look at the framework and assume its JavaScript.
Do you have a suggestion on how to clarify this?
Note that, if you're interested in using the platform UI - there's a project called Brisk (https://github.com/briskml/brisk) that is very similar to Revery, but uses platform UI frameworks (Cocoa/Win32/etc).
It's probably just Mac-heads that are particularly bothered by this topic, given that MS seems to create and sunset their native GUI frameworks fairly regular, and Linux has no single officially blessed one that everyone is attached to.
The goal of the project is to allow developers efficiently create fast, native apps for all major platforms. It's the best of both worlds: speed and portability, simplicity and great APIs."
Your apps are more predictable to developers between platforms, but less predictable to users.
A user expects all apps on their computer to behave the same. If you bring your own widgets, you violate that convention.
Most people use Electron because most native GUI's suck as they're antiquated APIs are unproductive with slow dev/iteration times. Flutter lets you build beautiful Material Apps using a refreshingly modern Reactive UI with built-in hot reloading.
Dart is like a cleaner Typed version of JavaScript, a bit more static and restrictive in someways but still really easy to pick up and learn if you have any 'C-like' language experience.
This is the kind of thing that leads to an "uncanny valley" app, where things feel slightly off, or they look good until the OS changes something and Flutter doesn't keep up.
Disclaimer: I wrote that line :(
I don't trust Google to emulate a perfect UI among all platforms. I feel like they're generating too big of a workload for no reason, it's a big red flag. I think eventually they'll give up on emulating native controls and choose a unified appearance for all platforms (forcing Material design for everyone) as well as giving the option to customize. I think their endgame is to push Flutter as much as possible so they have Fuchsia developers, then let the other target platforms go unmaintained. Their language, their OS, their devices.
<rant> Hover probably suits my needs, but I think the Flutter engine could be separate from the app, on the desktop, as the Dart jitted VM seems to have enough performance to do so. That could fully remove the need for a toolchain for the majority of apps. Just feed it Dart, in source form.
In the case of Go, a library could enable communication with such a Flutter VM through channels. The resulting executable could be much simpler, and would merely focus on extending the VM with platform features. Concretely, it shouldn't be necessary to recompile the Flutter engine, just to add serial port access, for example. Likewise, the Flutter engine could be upgraded by simply replacing the binary.
The thing is, your language of choice, any, could be used to extend such an engine. Fundamentally I don't believe in monolithic apps which compile in the kitchen sink, as they don't allow me to use the best language for the task. Most languages try to be everything to everybody. Why not communicate better, between processes? Some of these processes may run on the other side of the world. That kind of stuff should be easy, so no DLL's please, instead a kind of broker. </rant>
I work in a municipality and we simply can’t keep up with different frontend stacks for web, mobile and desktops or even the pace of which modern JS framework versions become obsolete.
OMG I wait for this so much... Finally a modern, high-level, really cross-platform JavaScript-free GUI framework...
It really feels like Typescript won the war to be something of a JS replacement a long time ago - I like the look of Flutter, but I really don't want to learn another language that I won't use for anything else.
BTW, I realise Dart is likely similar to Typescript in many respects, and I wouldn't expect a big learning curve - but I already have enough languages under my belt and don't want anything else to add to the occasional syntax confusion I get.
You might be thinking that Flutter underneath is using JavaScript - to which Dart compiles. However this is not how Flutter works, there is no JavaScript involved anywhere in the stack.
Dart is its own language - and it can run natively, unlike TypeScript - which is just a type system over JavaScript and needs to be compiled to JavaScript to actually run.
So no, you can't use TypeScript with Flutter.
And is there a place where more details are available?
There is bytecode, but the story here is kinda complicated - there are actually 2 different versions: one is called DBC and is used during development of Flutter applications on iOS, there is another called KBC - that one is more like a classical bytecode, it is currently in development.
> And is there a place where more details are available?
You can get some high level information at https://mrale.ph/dartvm
Let me know if you want to know something special.
I haven't heard of this. Why another bytecode version? Increased performance?
KBC is an experiment in this space where you have a more classical interpreter which tiers up into native JIT. In KBC mode bytecodes are also given to VM rather than generated internally inside the VM like DBC does.
Theoretically KBC should have better startup latency than just JIT (because you can start executing code immediately, rather than spending time to generate native code) and peak performance that matches JITs - because it tiers up into JIT.
https://github.com/dart-lang/sdk/issues/34343
Not sure if this will use AOT or JIT though.
You can do either. You can compile a snapshot and run that on the JIT VM. This basically gives you faster startup because it doesn't have to do all the source parsing, resolution, type-checking, etc. but gives you the general runtime performance of a JIT VM (think V8 or the JVM).
Or you can AoT compile your app all the way to native code. It's similar to how a shipped Go or OCaml works where it compiles your code and the language's runtime together into a single executable.
By the way, is there a good performance breakdown between Dart's AOT vs JIT? I've read some tweets/comments that JIT is still generally faster. Not including warm-up of course.
For me, React, ReactNative, Electron, even Vue, is enough for now, I can not trust Google on yet another web effort anymore, life is too short to trust a giant who behaves unpredictably sometimes.
Does anyone know of something that will take you from zero mobile knowledge to building some simple apps?
I use it as my daily driver, but its performance pales in comparison to something like Sublime.
I'd definitely wait until the Flutter team has replaced the OpenGL dependency with Apple's Metal. (I really hope they will do it.)
But deprecated just means "start moving off of this please" not "it's already broken"
Who's to say Google's plan isn't to make this become big, then purposely give iphone a shittier experience just a few ms of more jank to make android look better? Think long, long term. I wouldn't trust this.
I don't get why you are downvoted..
But it does work on all platforms.
One more issue is usability, since os desktop environments, doesn't support web-apps. One needs to pack it with electron or something similar to it.
Which increases your application size and memory usage.
For the rest React-Native is probably the best choice right now. Using mobile optimized native UI elements saves quite a lot of battery.