I hereby resolve to blog more next year! Hopefully, the new blogging app I’m using on my phone (testing it right now) will make it easier to fit blogging into spare moments.

Anyway – let me wish you a very happy holiday season! My holiday is Christmas, so it seems a good idea to test this app’s photo posting function with a festive tree…

OK – this post is just for the JavaFX developer crowd.  It shouldn’t be taken as a benchmark of JavaFX current or future performance…

Last night, I tweeted about the results of a quick test I did as preparation for building what I hope will be an interesting JavaFX demo application.  The reason I tweeted the results was because I found them a little surprising; and I thought people would be interested.

Now, a couple of people found the numbers hard to believe… So, here’s the code…  It’s very simple: it just inserts a number of dynamically created Line nodes (with random co-ordinates) into the scene graph:

/*

 * A minimal JavaFX program to test on-the-fly creation of

 * new Line nodes, and their insertion into the scene graph.

 *

 * You can play with different values of "n" in the run function.

 * The program will print out to the console the time in seconds

 * that it takes to create the nodes.

 */

package minimalinsert;

import javafx.stage.Stage;

import javafx.scene.Scene;

import javafx.scene.paint.Color;

import javafx.scene.shape.Line;

import javafx.util.Math;

package def stage: Stage = Stage {

 title: "Minimal Dynamic Line Node Creation"

 width: 640

 height: 480

 scene: Scene {}

}

function run(args:String[]) {

 var i = 0;

 /*

 * Set n to be the number of new Line nodes you want to add to the

 * scene graph dynamically.  Warning - when n gets more than a

 * a few hundred, this code takes a *LONG* time to run

 *

 */

 var n = 100;

 var t1: Double;

 var t2: Double;

 var time: Double;

 t1 = java.lang.System.currentTimeMillis();

 while (i < n) {

 addNewLineDynamically();

 i++;

 }

 t2 = java.lang.System.currentTimeMillis();

 time = (t2 - t1)/1000.0;

 println(time);

}

function addNewLineDynamically(): Void {

 var x1: Number;

 var x2: Number;

 var y1: Number;

 var y2: Number;

 x1 = rand(0,640);

 x2 = rand(0,640);

 y1 = rand(0,480);

 y2 = rand(0,480);

 insert Line {

 startX: x1, startY: y1

 endX: x2, endY: y2

 strokeWidth: 2

 stroke: Color.BLACK

 } into Main.stage.scene.content;

}

function rand(min:Number, max:Number): Number {

 return min + ( Math.random() * (max - min) );

}

I see the following performance with different values of n:

• n=10 , 0.02 seconds
• n=50, 0.11 seconds
• n=100, 0.36 seconds
• n=200, 1.8 seconds
• n=300, 5.3 seconds
• n=400, 11.8 seconds
• n=500 22.5 seconds
• n=600 37.7 seconds

I assume other will get similar performance when they run the code…. All of which begs the question: what am I doing wrong? As I say – this is one for JavaFX developers…

Why Applets Are Important – Augmenting Browser Functionality

Before considering JavaFX applets in detail, it’s worth a moment to step back and ask the question: why are applets important? Clearly, their reason for existing is to augment the functionality of a web browser. For example: to play audio or video; or to allow sophisticated graphics for video games or other apps with complex graphics needs.

Now, many people will argue that the JavaScript/HTML/CSS is on a trajectory such there will soon be nothing the browser won’t be able to do natively, and thus the reason for Flash, Silverlight and JavaFX applets will disappear. They may be right: not having to rely on browser plugins is an highly attractive idea.

However, life is rarely that simple, and the future might not be black and white. The truth is: future JavaScript/HTML/CSS represents strong competition for Flash, Silverlight and JavaFX. What the future ends up looking like, though, will depend on how the various technologies compete. There are battles to be had in: developer productivity for creating compelling applications; and application deployment etc.

Results Of The Test

Start-up Time – Observations

First the good news – if you’re lucky with your hardware/OS platform, and you have an up-to-date install of JavaFX, JavaFX applets can start in around one second.  That’s a great result, and a huge improvement on where the Java platform has been historically.

However, what the test results really show is that the above case of super-fast start-up is very much about the future potential of the technology.  It’s clearly not where the technology stands today. In fact, the frequently observed start-up times in the test ranged from 1 second to c. 45 seconds, with start-up taking even longer than this for some people.

The applet used in the test is a small applet, about 30KB in total. Thus download of the applet itself should not be a significant contribution to observed start-up times. So, what is the wide range of start-up times due to?

It turns out that there are two major determinants of the start-up time of a JavaFX applet: 1) The hardware/OS platform of the user – which determines which version of Java and the Java Plugin they can use; and 2) whether or not up-to-date jars for the JavaFX run-time are installed, or whether they have to be updated on-the-fly during applet start-up.

For users running MacOS X, the start-up time for the Java Plug-in is slow (perhaps 8 to 10 seconds), even on new hardware. On Windows hardware, the corresponding start-up times are between 1 and 5 seconds.

For users that don’t have the latest JavaFX jars installed, depending on broadband speed e.g. 2Mbps, 5Mbps, it will typically take between 15 and 30 seconds for the 14MB or so of data to be downloaded.

Clearly, those end-users of apps that are worst off are people using Mac OS X and that don’t have the latest JavaFX jars.

Rendering  & Other Bugs

Certainly on both Windows and Mac OS X, there can be terrible rendering bugs in JavaFX applets when the web page in which the applet is embedded is scrolled. On Windows, sometimes the web-page stops rendering immediately after the embedded applet, so all content below the applet is lost. JavaFX applets don’t work at all in the Opera browser or in Safari 4 on Windows.

What Role Can JavaFX Applets Play Today?

The rendering bugs in JavaFX applets are sufficiently bad as to make applets embedded in web pages that are likely to be scrolled an unacceptable choice for developers. That rules out applications like fancy embedded graphs and charts or embedded video players.

The variation in likely start-up time for a JavaFX applet – anywhere from 1 second to 45 seconds – without a good way of communicating what’s going on to the user – mean that only categories of app where the user is prepared to wait 30 seconds or more for their application to start are going to be possible choices.

The above suggests that the sweet-spot for JavaFX applets today is for building creative, browser-based video games. The elegant  JavaFX programming model should make it productive to develop interesting  games on short timescales.   As an example, here’s a preview of a neat JavaFX game called Pastello which is heavily inspired by the popular indie game, Crayon Physics (I really recommend clicking through to Pastello, by the way – it’s an impressive piece of work).

Driving next stage of JavaFX adoption

Casual games are, of course, an interesting and substantial category.  However, if that’s all that JavaFX applets can achieve, I think we have to say they’ve been a failure.  That’s because it will mean JavaFX applets wouldn’t have moved the Java platform forward – the truth is, regular Java applets have always been pretty widely used for browser-based video games.

How then, to drive the next stage of adoption for JavaFX applets?

Rendering bugs on scroll must be fixed in order for JavaFX applets to take off for visualization applications. The variation in start-up time means that users expectations must be managed. As a first step, it will be important to let the user know it might take up to 45 seconds for the app to start up if they’re on a Broadband connection, and even longer if they’re on dial-up. However, that’s very much a workaround.  What are really needed are progress bars that can be shown embedded in the web-page (and styled by the developer) to indicate download progress.  Indeterminate progress indicators – like the spinning animated gifs developers can use now – are useful; but they are a really poor second choice to determinate progress indicators for long JavaFX start-up times that will have to be accommodated for some time yet.

Conclusion

Hopefully, the JavaFX and Java SE teams have been aware of all the issues brought to the forefront by this test for some time; and they already have solutions in the works. If not, I hope they take on board the clear findings from this test, and the likely consequences for JavaFX adoption: that is, rendering bugs and long start-up times with no effective progress-indicator-style feedback will put a massive brake on developer adoption for JavaFX applets until these issues are addressed.

With the release of version 1.2, the JavaFX platform is starting to become a viable choice for developing real-world, browser-based rich internet applications (RIAs). However, there are questions to be asked about deployment. In particular, it’s important to know what are the chances that people will actually be able to use a JavaFX app without having problems getting it to run.

In order to begin to answer this, I’d really appreciate your help in running a little test. Even a few seconds of your time would be valuable here (and if you have a few minutes, that’s even better), so I hope you can help.

Assuming you’re reading this blog on a desktop computer (as opposed to a mobile phone), below is – or should be(!) – some JavaFX running. If it’s running properly (note, it may take a few seconds to start) you should see a black rectangle with some colored triangles bouncing around inside the black rectangle. If you don’t have JavaFX installed, you should be prompted to install it.

Here’s what I’d like you to do:

• If you can see the colored bouncing triangles above, please leave a comment saying – “Works for me.”
• If you can see the colored bouncing triangles above, but you had to install JavaFX first, please leave a comment saying, “Works, but only after installing JavaFX.”
• If you can’t see the colored bouncing triangles, please leave a quick comment saying “Doesn’t work for me.”

If you can spare a little more time…

1. Please can you also include in your comment: a) what OS you’re running (e.g. Windows XP; Vista; Windows 7, Mac OS X Leopard, Mac OS X Snow Leopard; Ubuntu Linux; OpenSolaris etc); and b) what browser you’re running e.g. IE, Firefox, Safari, Chrome.

2. Please give an estimate of how long for the applet to start e.g. 2 seconds, 5 seconds, 30 seconds, 2 minutes.

3. If the JavaFX didn’t work for you right away, didn’t work at all, or there was some other issue (e.g. problems with the way the applet repaints when you scroll the web-page) please give a short description of your problems.

Thanks so much in advance for your help with this…

JavaFX

First, watch for important new developments in the JavaFX platform.  I expect there’ll be more than a few announcements relating to JavaFX.   Up until now, JavaFX has really been of interest only to innovators and early adopters.   It hasn’t been suitable for the so-called “early majority” – a big group of developers that will only begin to adopt new technologies when they’re close to being ready for prime time.  That’s because JavaFX has been missing key features – such as off-the-shelf GUI components, and good visual tool support – that make it easy and productive to develop applications.   Expect big developments on these fronts to be released and/or demo’ed at this year’s JavaOne.   The first real JavaFX Mobile handsets should also be demo’ed too – so people will, for the first time,  get a really good idea of how well these will work.

I’m also hoping for a couple of JavaFX surprises too.  For example – it would be great to see more new partners for JavaFX Mobile announced; and great to learn about the road map for the new integrated JavaFX 2-D/3-D scene graph, that will see JavaFX gain capabilities for high-performance 3-D graphics.

Project Vector aka The New Java App Store

Second, watch out for a brand new offering from Sun called “Project Vector”. Details of this one are sketchy.  What is known is that it’s an app store that leverages the distribution power of the Java installer on Windows; which means that apps might have the potential to be placed in front of hundreds of millions of possible users every year.   Clearly, the devil is in the detail on this one… How will it work?  How will apps be targeted to customers?  How smooth will the installation process of apps be?  How much will it cost?  How will free apps be handled?  How will payments be handled? Will it work for mobile Java ME and JavaFX Mobile apps?  What about the Mac platform?  And Linux and Solaris?  Expect the announcements and demos at JavaOne to answer at least some of the more important questions for this new, potentially rather large app store.

All in all then, this promises to be a rather interesting JavaOne.  It runs from 2 June through 5 June, with CommunityOne starting earlier, running 1 June through 3 June.

Twitter was built using software tools like Ruby on Rails, which is provided as part of Mac OS X

However, Apple forgot to say that for the meat of its systems, Twitter is actually moving increasingly to Scala – a language built on the Java platform.  Why? Because the Twitter folks believe that the Ruby platform isn’t up to the job.  So much for Java being some “heavy-weight ball and chain that no-one uses” (as Steve Jobs once famously said)

Vision 1 – The Boring Option

Vistion 1 is very much “Steady As She Goes”, or what I’d call, “The Boring Option”. The vision here for the new combined organization is essentially – “Oracle as it was before,  but with incrementally  increased revenues and profits from making Sun profitable”

For example, increased revenues would be achieved by selling Sun hardware running Oracle/Sun software; and incremental increases in profits would be achieved by running the Sun business unit profitably through immediate, substantial cost-savings.  There might also be a goal set to increase revenues and profits from Java e.g. from increased license fees. You can dial in the level of instant profitability (leaving out re-structuring charges) here by defining the cost-savings required, and thus the number of jobs that need to be lost.

This is an easy enough way to do M&A; and it can work well enough with the right kind of assets.  It has the benefit that it’s easy for auditors, shareholders, and integration planning teams to understand.  However, whether this is the right way to proceed with Sun’s assets is an open question.  There’s a risk here that “boring”  might become “terminally dull”, with “terminally” being the operative word.

Vision 2 – The Exciting Option

Vision 2 is transformational for Oracle.  It’s what I call “The Exciting Option”. What is it?  Well – ask yourself the question: what does the future of computer systems look like in a ten-year time horizon?   I don’t think there’s any doubt.  It looks like amazing software applications, both “thin” and “rich”, running on all types of client hardware – cell phones, TVs, personal computers – connecting to amazing services running in the cloud.

It’s a huge opportunity, just huge.  The interesting thing is – the combined Oracle / Sun has all the pieces it needs, along with the required resources, to go after it.  So, the exciting vision is something like, “Oracle is the biggest, most profitable computing company in the world, defining the future of how people use computer systems running on any device, for business and pleasure.”

Here, the instruction to the teams is to take a step back, and really try to figure out what can be done with all the combined company’s existing assets today – both in the data center, and on the client; what further investment is needed to accelerate the company towards achieving its new goals; and what barriers to success need to be broken down.  This is not about incremental increases in revenue and profits.  Rather it’s about going after ten times increases.

I know which vision I’d rather pursue.  If history is anything to go by, however, Oracle’s leadership will take a different, rather more conservative view.  You never know, though.  Sun represents a rather different kind of opportunity compared to previous Oracle acquisitions, which only ever had the possibility to deliver incremental benefits; but I’m not holding my breath…

By the way, as a footnote,  if by some chance, Oracle’s leadership is up for a bit of revolution, rather than the usual simple Darwinian evolution, Larry Ellison might even be able to get some kudos from developers by smoothing out some of the smaller bumps in the road himself… like convincing his close friend, Steve Jobs, to allow a fabulous version Java/JavaFX on iPhone

According to Larry Ellison, the key drivers for Oracle wanting to buy Sun were based on Oracle’s core strategy of purchasing industry-leading technologies. In this case, that’s: Solaris – which Oracle rates as substantially the best Unix platform in the world; and Java – the most important software asset Oracle has ever bought, deployed on billions of devices.   Buy owning hardware as well as software, Oracle aims to to deliver complete systems (as opposed to components) engineered to provide superior reliability and security, at no additional cost to the customer.

Oracle says it plans to increase investment in Java.  There are some obvious questions about Oracle’s commitment to the Java platform on the client (at its heart, Oracle is a server-side company – database and middleware).  However, Oracle has been developing client Java applications for a long time so hopefully it understands: the economies that come from allowing the same technologies to run both client and server side; and the value of having one of the largest developer communities in the world, building systems that run on billions of devices.   Hopefully, Oracle also understands that Sun has a deeply impressive software stack (Netbeans IDE, Java/JavaFX, Glassfish, MySQL, VirtualBox etc), favored by some of the smartest developers around the world.  With increased investment, this stack has the potential to dominate the industry.

That’s the story.  It makes some sense at a high-level.  The devil is in the detail when it comes to these things, however – so only time will tell if this is good for customers or not. It will be interesting to see if IBM is now content to sit on the sidelines, or whether it believes there is room to  offer more to Sun’s shareholders, in the form of  hostile bid, than Oracle’s offer, which has the full support of Sun’s Board (Oracle is offering to pay $5.6B for Sun net of debt and cash).

• Do you think that smart Solaris developers will entertain the idea of migrating to AIX?
  • The answer is that they won’t.  They will prefer either Linux, or Windows, or Mac OS X.
• Do you think that smart Swing and JavaFX desktop and mobile developers will entertain the idea of migrating to SWT?
  • The answer is that they won’t.  They will prefer to focus their time on JavaScript web apps, Flex/Flash apps, iPhone apps and .NET apps instead
• Do you think that smart Glassfish developers will entertain the idea of migrating to Websphere?
  • The answer is that they won’t.  They will prefer to focus their efforts on JBoss instead.
• Do you think that smart MySQL developers will entertain the idea of migrating to DB2?
  • The answer is that they won’t.  They would prefer to focus their efforts on PostgreSQL.

I won’t go on.  You get the idea. The point is this:  one of the things that IBM will get when it buys Sun is a bunch of technologies that smart developers really like.   The Sun technology stack of Solaris, MySQL, Glassfish, Java Swing/JavaFX, and the Netbeans IDE is incredible.   Sure, IBM can kill as much of this as it wants.  However, if it isn’t rather careful, it won’t benefit from doing so. Rather, it will push dollars into the hands of others such as Microsoft, Dell, Adobe, Apple, and RedHat; and cost IBM folks their jobs down the line.

I will end this entry with a message to the folks at IBM that are about to wield their axes.  It’s the same message I once gave to someone who was about to make a big mistake on a Friday afternoon during a rationalization exercise I was part of years ago, and it goes like this… The people you fire will leave the company and get on with their lives. You, however, are staying – so this is all your responsibility now;  and you are just about to screw everything up and destroy the opportunity.  When people understand what you have done, it’s your job that will go next.  So, if I were you, I’d take the weekend to think very carefully about whether you really think you’re doing the right thing, or just satisfying a misguided personal agenda.  It’s not too late -  there’s still time to fix this…

In that case, I got a phone call first thing Monday morning to say that, on reflection, they’d changed their mind and decided to fall into line with my views.   However, I’m not optimistic this blog will have an analogous affect on IBM, given I have zero influence there   But hey, it was worth a try…

Assuming the deal goes through, there would be a major review of the combined company’s key products, software, technologies and services.  The aim of the review would be to decide which of these to keep, and which to kill.

Here are some what you might call “detail” questions that Sun customers will want to know the answers to:

• There’s a lot of BS thrown around on this topic, but the truth is that Sun is much more committed to making software freely available to developers than is IBM.   Will IBM kill the Sun strategy here?
• There are many Sun products and technologies that compete with offerings from IBM.  Will IBM kill these or cripple these; or will IBM contemplate killing IBM products while keeping Sun ones?  For example, what would IBM do with software products and technologies like: Solaris; Netbeans; Glassfish; MySQL
• There are many aspects of the Java platform that don’t compete with IBM offerings directly; but are these of any interest to IBM?  For example, is IBM interested in JavaFX?  Or does the company actually believe that SWT is the way forward?

For customers, the problems with the above are that: such decisions are often not made rationally in M&A integration processes; and, in large organizations such as this, customer views and needs will simply not be taken into account.   Rather, at the top, the process will be designed such that the outcome is pre-defined to be “the answer they first thought of”; and all the detailed decisions will be taken by people with personal agendas, who are focused on keeping their jobs and building their careers in the new, combined organization.    What’s good for the customer won’t even be on anyone’s radar.

The Really Big Question

All of the above is detail, though. There’s a much bigger question that needs to be answered. It’s this.  For IBM, would it be worth $6.5B to make the whole of Sun essentially “go away”? I haven’t done the arithmetic to figure out the answer, but I suspect that it’s very possibly the case that IBM could be interested in buying Sun, simply to shut it down.

Don’t look to IBM to answer that one directly, by the way.  CEOs of public companies almost always lie to their shareholders about the real reasons for M&A, particularly when it’s a big acquisition.  Because of the consequences of doing that, they almost always lie to everyone inside the company too, spinning them the same story that they spin shareholders.  So, we’ll only find out the answer after the acquisition has gone through.

No matter what the answer, though, in event of IBM buying Sun, I’d expect massive consolidation with many great Sun offerings being canned.   I’d also expect the prices of Sun offerings that survive to increase.   As I said at the start of this blog, it’s difficult to see how this is good news for customers.   The problem is that the alternative – of IBM not buying Sun – might not be good news for customers either…