A few month ago, i wrote a post about the situation between JavaFX and HTML5 and their positions in the fight for the next leading GUI-technology in the Java world. I said, JavaFX is only interessting to improve existing Swing application but not the right thing for the developing new applications.

JAXenter.de wrote an article about with a simalar conclusion: https://jaxenter.de/ruhe-in-frieden-javafx-31013

I can fully agree to this point of view. There are greater problems than the technological difficults and the few bugs and problems with the implementation of some JavaFX parts. The greatest problem is the missing support by Oracle. JavaFX for mobile devices is only supported through 3rd party solutions like RoboVM. Scene Builder is not very stable or bugfree. With only simple changes in a FXML, you can cause troubles that Scene Builder is not able to ... it crashes or only shows an empty screen.
You have to wait month for new versions or simple bugfixes.

Oracles own applications use Swing or are webbased. The change from Swing to JavaFX was often proclaimed, but never came to the surface or was visible in any other way. Swing has an extensive repository of mature and feature-rich components. It isn't easy to replace them in a short time range. A simple migration isn't that simple and would took a lot of work. You know that you can open JavaFX scenes in a SWT application. I used it, but it isn'T realy a good integration. It is only a canvas, where the JavaFX scenes is drawn on. No real interactions between the two worlds. It like the WebView in JavaFX. Yes you could write an application in HTML and JavaScript an use JavaFX as an wrapper.. but.. why? An IE11 should be a better solution than this.

The usage of Swing disappeares slowly, but there no new JavaFX applications to repalce them. There are a few JavaFX fans, but i think they only exist, because Swing is deceasing. Since i begin to programm in Java i wrote a few Swing applications and the first contact with JavaFX was more than disillusioning. Netbeans + Swing working great and every thing works very well. Scaling of the GUI works good. In JavaFX you can implement it too, but it isn't as intuitive as with Netbeans. The CSS-implementation in JavaFX is not comparable with the CSS we know from HTML.

Why JavaFX 2 wasn't accepted well? In my oppion, the reputation was bad, the support was insecure in version 1.0.
When version 2.0 arrives, most developers found a other or better solution with better support and more promising future. So the support from the developer-side was missing and without this every thing only works on a minimal level to save the resources for more promising projects.

JSF 1.x wasn'T that good, but there was no doubt, that it will be supported an new version will be developed. I think that Apache supported JSF, was a big advantage. And the newest versions are great. Maybe JavaFX can go the same way with more support.

So i support the conclusion: If oracle will invest more resources in JavaFX more developers show interest and so it could be a future with JavaFX.

I can't understand the fear that without JavaFX the JEE market could be damaged or it can lose it's impact in the developer-environment. Node.js could be a threat, more than php, but in the last years the JEE-market grow with microservices, middleware-services for android-apps, reactive-development and frameworks like node.js. REST-services, web-portale-solutions, small containers, websockets und push-services in the Tomcat servlet-container. Very nice and the movement doesn't stop till now. The big Topics of the last years.. Desktop-GUIs wasn't a topic. And with the great Java webframeworks, HTML5 and AngularJS the client-side won't be losed. JEE and Web-Clints work great together.
The lose of a few desktop application won't hurt, when you get new shiny web-application based on Java technology.