I try out JavaFX pretty much every 3 months or so. Have been doing so for almost a year now and i can finally said that JavaFX is now good. They recently finally released a update that enables it for Mobiles and they have been seriously tuning it for desktop use with the last two updates of Java.
The last time i tried out JavaFX it was still a trainwreck. Â Even if i always have noted that JavaFX may well be a lot more powerful than Flex and Silverlight. In fact i am pretty much sure that it is more powerful.
However they got HUGE disadvantages against Flex and Silverlight and that Sun, now Oracle got vs Adobe and Microsoft. This continues to be the case. JavaFX even that it now has a mobile release, a devices release and the desktop reflects maturity. It will not be enough. Or it is not enough yet.
The Good
3 months ago or so, JavaFX had the highest CPU usage rate from all the RIA solutions. Now it seems to be lowest one and looks to get the best CPU performance rate i have seen. I tried this out in a Old PC, a Netbook and a Average Desktop PC with all JavaFX samples and the same amount of samples for Flex and Silverlight so i came very sure with what i was going to say here. So passing from worst to best in that crucial area,  In that time frame is nothing short of incredible and my props to the JavaFX team. It is a incredible evolution.
The bad
Last time JavaFX was excruciatingly slow to load. Any sample or any decent application was slower to load than any similar in scope  Flex or Silverlight alternative.
It is still the worst but by very very little. In some cases it maybe about even with Flex and Silverlight
The Ugly
1.-Authorization
Given that JavaFX is still essentially java at the base, That means that authorization and certificates are still a issue. Â So every single thing done in JavaFX must be properly identified so it can load.
When it comes to “Load To Desktop†webstart apps it is not a big issue since the behavior and install experience is pretty much identical to .NET ClickOnce and Adobe AIR so it is to be expected.
But “Load In Browser†apps are another thing. Since usually you only need to identify them once and every time after they will load fine. This is based on who made it and how you authorized the maker. So if you encounter several apps from the same maker and you authorized the first app you encountered from said maker as a trusted party. All apps from that maker will just load every single time you encounter them in the future. If you didn’t marked it as trusted maker or the apps comes from a maker you have not authorized or encountered. Then each app will prompt you before it can load.
Complicated?, Not really. Annoying?, Not for me. Â But for those users that could not stop whining about UAC in Vista having the nerve to prompt them once a day when it needed to?. Â Oh you bet they are going to love (NOT) how JavaFX works. So this is something that needs to be seriously thought out and resolved.
My proposed solution is for Developers to be able to ask for “Clearance Keys†at Oracle so that the prompt only comes from not registered JavaFX apps or widgets. Well, it is just an idea that Oracle could steal from how oAuth works.
2.-Installer
The greatest virtue and flaw from JavaFX comes from being tied up with the rest of the Java Runtime.
It gives it ubiquity and all the power in the world, but it makes it a big download, gives it a big installation footprint and makes it a power hungry install when compared with Adobe Solutions that comes in 5 separate plugins and Silverlight that is a single plugin install.
But this is a issue that i don’t find Oracle being able to develop a workaround that satisfies them, given what they would lose if they unbundled JavaFX to be on its own.
3.-RAM Consume
The worse for last. RAM consume continues to be off the charts for JavaFX. It even makes the worst Flex, Flash or AIR apps look lean and light in RAM consume if compared to JavaFX. So it is dead last in this one.
Were else you could see a simple tiny widget (in-browser) consuming over 100mb of RAM on its own or a simple desktop app consuming as much as 500mb of RAM?. Â Only with JavaFX.
So i guess that such dramatic gain on CPU efficiency had to be loaded somewhere. In this case to the GPU and to the RAM. Â But even as bad as JavaFX RAM consume is. It is logical they decided to go this way since any laptop and desktop now comes with at least 3gb of RAM and some form of GPU chip or card a lot better than what came as default two years ago when RIA were materializing.
Conclusion
I don’t know how much the Start Load, Authorization and RAM consume issues can be improved but if they manage to improve them even a 50%. That alone would make JavaFX a true contender for 2nd place in the RIA space. A 1st place is something i find almost impossible for they to get and that will be battled out on and off by Adobe and Microsoft.  A 2nd place is possible even if quite hard to see it now. But a very close 3rd place is what Oracle should aim for now,  given they are still in a very very very distant 3rd place. But maybe in 3 months i will have to change my mind.  So lets see.




JS-KIT Acquires Sez Who To Consolidate Online Commenting
March 17th, 2009It had to happen. do i am surprised?. no because i was the one talking on and on that Sez Who only out was to be acquired by JS-KIT or Disqus. back when IntenseDebate was not part of Automattic it was also the way to end any chance of any other competitor to ever touch them. They now have to worry about Automattic, Facebook and more than surely Google coming with something in this space now that Facebook has got into this too.
Fortunately for JS-KIT this acquisition and the accumulation of deals and talent puts them into the position to be able to fight Google, Facebook and Automattic and hold on the space since they are the true experts of the space with lots of history on the market.
So far in terms of acquisition JS-KIT acquired the old Haloscan and Sez Who. the first served trackbacks and alternative comment systems and Sez Who serves commenter reputation and Identification. both are good complements but i think that the missing link now is actual comment tracking and harvesting tech. that sounds like Cocomment. they already got a deal with them but with the recent fallout of Cocomment last competitors (co.mments and Commentful) as one has gone off and the other has stagnated and its also dying. Cocoment is a great opportunity to expand the tech and talent in order to anticipate possible threats like backtype that is a new kind of Comment tracker and harvester. the reason is because disqus could be looking into that too since it is pretty much their only chance they have to not be overwhelmed by the now very strong competition competition in this space. i see it hard for them to survive on their own unless they have one last ace to show.
The Sez Who acquisition surely will help JS-KIT to compete but it will still have to fight for the first place. i think a Cocomment acquisition would make sure they get the first spot as the combination of talent and tech should allow them to evolve their services faster and grow at a faster pace.
But now i also wonder about a possible brand name change and image revamp as they are not longer just JS-KIT’s anymore.
JS-KIT
Via JS-KIT and Techcrunch
1 comment »
Posted in Tools, Web 2.0, Widget Distribution
Tags: comments JS-KIT opinion Sez Who tracking