When Will JavaFX Get Real Community Acceptance?

APIs for Java are needed. The JavaFX functionality should be accessible from the Java platform, that means from Java, Groovy, Scala... I hate it, to be forced to use a yet another language for the GUI portion of a desktop application. It is a nice language and who wants, should use it. But Sun please understand, most people don't really need a freaking new declarative scripting language. There are already far too many scripting languages around.

JavaFX functionality should be easily integrated in Swing and therefore in the Netbeans application platform. It should be possible to create professional looking applications and use some of the the cool new features such as HTML rendering, PDF display, media playback and so on. For professional GUIs I do not want to have a complete Flash-like interface as in the JavaFX Script applications I have seen so far.

Forget about interactive TV. TV is a totally passive medium. People switch their brain off when looking at a TV screen and let things happen, but do not want to make decisions.

I have a hard time seeing JavaFX going anywhere, and would generally be much more excited if it was not this isolated applet/webstart technolgy but rather a full replacement for Java. The language is interesting and succinct, but it has to compete with stacks and languages already accepted. So for me the only thing that would breathe some life into it, would be if it could be used on Android.

As usual in 99% of the articles about "will JavaFX survive?", I have do make this correction: it's not two years old (as of jan-2010); v1.0 was released in dec-08 so that's just 13 months old. It's kinda wrong and unfair to round that up to "2 years". Yes, Sun has been hyping FX much longer than that, I think since JavaOne 2006, but that was a gross mistake. Yes, early-access releases were made available a few months before dec-08 but these were really early; v1.0 had massive fixes and breaking changes even compared to the latest preview release. Even that v1.0 is now widely considered to be just "developer release", it didn't have enough functionality or quality for any real-world deployment; but even not considering this, JavaFX is just one year old (plus one month). JavaFX Mobile is much more recent, the FCS release shipped less than 2 months ago and still just for WinMob. JavaFX TV has yet to ship. The design tools have yet to ship (the NetBeans FX Composer is in beta - I'd expect the FCS to come in 1-2 months, perhaps in some NB 6.8.1 update, perhaps paired with the FX 1.3 release). The JavaStore, that is a big part in Sun's plan to promote and also monetize on FX (plus all other client-side investment like JDK 6u10+), has yet to hit FCS status and quality (I guess this depends from JDK 6u18 and FX 1.3). The Design Tool has yet to be released publicly even in EA form, all we get is a continuous stream of presentations in conferences.

In summary, it's come a long way - and very quickly - in these 13 months since 1.0, but there are still important missing pieces. The next 6 months will be extremely critical for FX's sucess; in this timeframe we will see:

a) JDK 6u18 (end jan / early feb) with more work in plugin, JAWS, core AWT/2D support, deployment, startup etc. (as well as general performance - like HotSpot v16 backported from JDK7)

b) The FX 1.3 release (late feb), with tons of important stuff (compiled bind, many new controls although not every one we want) and hopefully more Mobile ports (MWC happens in mid-feb)

c) FX Composer FCS; Design Tool at least beta/R. (Now the missing piece will be a decent (free) visual editor for Eclipse; the market advantage of Eclipse is a great barrier even if the tools Sun is providing will totally rock)

d) Early JavaFX TV

e) The JavaStore really finished and open worldwide. If all items above are well delivered, I see no reason why the Store can't make decent success. Notice that (different from the Mobile market), in the desktop Sun is not catching up - I don't see any app store from Adobe or Microsoft in my Windows desktop, not yet.

f) Finalization of the Oracle/Sun buyout, so everyone gets to know which Sun projects and products Oracle is really pushing forward (and not just put in life support for a few years, if not kill immediately). (I've just seen a FX guy tweeting announcements of open positions - apparently for experts in Saffron and other font engines - so the may know something we don't; people surely don't hire more engineers this week if the project may be closed next week after the EU finally clears the deal and Oracle starts rolling the dice).

No one "wanted" Flash or Silverlight (that also contain new languages, XML-based DSLs, runtimes etc.), either. Now it's possible that Sun doesn't have what it takes to compete with Adobe and MS, but that's another story. Users (i.e. developers) never "want" stuff that doesn't still exist. Innovation never comes without risk; some people will always complain that the same features could be delivered as simple libraries or perhaps incremental changes to existing languages/frameworks.

Don't underestimate the difficulty to extend a system that's as successful and mature as the Java language and its core APIs (including AWT/Java2D/Swing). I suggest you to track the whole discussion over JDK 7 lambdas (lambda-dev and closures-dev lists), and that's a single language feature- although a very important one, but something that's considered almost trivial in languages that were designed with that stuff since v1.0. But not in Java; we still risk having no closures at all in Java 7, or perhaps a half-baked design with lots of excuses (backwards compatibility yadda yadda difficulty for average programmers yadda yadda migration of legacy APIs yadda yadda...). Now, you will see that JavaFX Script contains not one, but many language innovations - BTW it does support closures, but that feature barely catches people's attention, in the shadow of less-trivial stuff like powerful properties, binding, sequences, object trees and other declarative aspects, trait-based OO model etc. The chance of having all these goodies in a reasonably backwards-compatible, Java-derived language - and decently implemented wrt features, good syntax, orthogonal behavior with existing language features etc. - is zero.

P.S.: One bright point of JavaFX for me is that I don't have to use a JavaScript[-derived] language. Or to "program" in XML. Call me a static-typing/class-model bigot, but I will work all day long with those languages to build real-world complex software, the day I start smoking crack and lose 90% of my brain cells. To each his own; IMHO, JavaFX Script has lots of appeal, enough to bring back to the client-side some hackers who have spent the last decade writing only back-end stuff, in disgust with the load of duck-typed spaghetti that is JavaScript, HTML, etc. (Yet I want the standards-based, HTML5 web to succeed big time, too... but I'll let other people write the code. Or use something like GWT or RAP.)

IMO, we won't ever see a (significantly) "better Java". Face it, the Java language and standard libraries are too widely used to be messed with at this point. Also, as Osvaldo said, from a language design point of view it's not easy to add radically different features to an already mature language.

For me, the future on the JVM is JavaFX. A whole new language (which is NOT a "scripting" language) and a whole new set of standard libraries; at last, we will have a replacement for AWT and Swing.

As an OO & Functional language, I consider JavaFX Script more well designed than Groovy or Scala. Anyway, given that those other two don't really add much in the way of new APIs, I doubt they really stand a chance to become mainstream.

We will have to wait and see...

Ah, good point. I stand corrected. Thanks for enlightening me on this topic.

I still like the jQuery ecosystem (insane amount of plugins you can use without having to wrap them in any way), but for large apps you are correct.

Either way, I'd still much faster choose jQuery or GWT over JavaFX (going back to the original discussion) :-)

Yes, for a new business web app GWT is a better choice today, as even JavaFX 1.3 won't be complete/mature enough.

On the other hand, JavaFX really takes UIs to a whole new level of "richness". The stuff (effects, animations, transitions) described in the excellent Filthy Rich Clients book is considerably easier to do with JavaFX than with Swing or with GWT/GWTCanvas.
I believe these new RIA tools (JavaFX, Silverlight, Flex), as well as browsers supporting HTML 5 + CSS 3 (see this article, for example) will change the way UIs for typical applications are developed. This is one of the reasons that will make the "old" JDK less and less attractive for new projects.

It could be possible that JavaFX is the next big thing and not Scala or Clojure?.

Im a Flex fan, I sit down to discuss with my team for a new project, we brainstormed and I decided to use Flex because it is the one I had experience but my team have not experience before with eather Flex or JavaFX at the end we decide to use JavaFX because they told me JavaFX is more intuitive the syntax and the language, it have performance and it can reuse Java Libs or integrate with Java tech more easily. I was dissapointed at that time but I gave a shot with JavaFX, It was a great experience, I think JavaFX have a lot of potentional.

I will not touch Silverlight, Anything come from M$ ruins me as always.

Also Oracle will put lots of cash to JavaFX so I think JavaFX have a nice future. HTML5 or Ajax(Jquery or GWT) It is also very nice and I want it also succed but as Rogerio said JavaFX brings more richness in graphics and animations and thats what the end users wants the eyecandy. Why IPhone is popular because the eyecandy that sells it is marketting almost sells by their self and that JavaFX have.

So I hope Java7 gets closures but also I think JavaFX will begin to have a success and more with the next release 1.3

Well, I looked at code written with Scala, Groovy, and JavaFX Script and to me the JavaFX code looks more elegant.

One such example can be found here, with Scala, Groovy and JavaFX Script versions of the same small GUI program.

For an enterprise app I would still use Java (with GWT on the client side).
For a rich desktop app, though, JavaFX seems a much, much better option. It's what it was created for in the first place, after all (between other things, such as rich mobile UIs).
I have not yet really started using JavaFX, as I am waiting for the new UI controls in JavaFX 1.3.

Now, I bet the "bind" keyword and "on replace" triggers in JavaFX Script are more useful language features than currying or partial functions. Also, JavaFX Script is a younger language, not yet fully developed. It doesn't yet include generics, annotations, enums, or direct support for maps, but these (and more) will eventually be added. So, I agree that right now Groovy and Scala are more complete languages. I just don't believe they have a great future (time will tell...).

I have looked at Griffon. It's nice, but in the end it's only a wrapper on top of Swing and Java2D. The JavaFX GUI APIs, on the other hand, are completely new technology which uses a different rendering model: "retained mode", while Java2D is "immediate mode" only.

Just for the record, it seems JavaFX Script actually supports currying and partial functions (I am not 100% sure at this point, though). The "Pro Java FX Platform" book has a relevant example in page 181, although the word "currying" is not mentioned.