Ted Kosan's Blog

Java's Telemetry Wave Wipeout ( And A Shocking First Java Program )

In my last blog I discussed the next big wave that is washing across the Internet, which Sun's CTO Greg Papadopoulos calls the Telemetry Wave. I indicated that the Java community has a problem with respect to telemetry because a completely separate community is in control of the devices that allow the Internet to access the physical world. This community consists of the various microcontroller developer communities and almost no microcontrollers that are used to monitor and control the physical world are programmed in Java.

I also indicated that I had a shocking story to tell which is related to this problem and here it is.

Beginning Java for PIC developers: Take 1

One of the few people in the world that is a member of both the Java community and the microcontroller community is James Caska who owns a company called muvium. James is an expert in PIC microcontrollers and he has created software that allows Java to run on them. This is an amazing accomplishment considering the limited amount of resources these microcontrollers have.

A few years ago, James put together a presentation on muvium for the annual PIC developer's conference. The first part of the presentation contained an introduction to Java for beginners because almost no PIC developers know how to program in Java. The "Java for beginners" part of the presentation was very well done and I think that most Java developers who saw this version of the presentation would agree.

Even Microchip liked the presentation when James showed it to them but they indicated that most of their developers had never been exposed to object oriented programming before and that they would not understand the coding examples. James was asked if he would rework the presentation so it could be better understood by the typical PIC developer.

Beginning Java for PIC developers: Take 2

I have already indicated that what I am about to show to you is shocking. You may think that I have used the word "shocking" in order to draw more people to my blog, but I was truly shocked by the following information when I first saw it and I still am.

If you think that you have seen almost every approach to a "First Java Program" that has ever been created, I am almost certain you are wrong :-)

Here is how typical PIC microcontroller developers need to see their first Java program:

and here is a slightly more complete version from later in the presentation:

Do you think the PIC developers who attended this presentation obtained a good understanding of what Java was and how enormously powerful it could be? Beyond this, do you think that the typical embedded programmer and the typical Java programmer would have enough in common to even communicate effectively with each other?

The above code examples show why Embedded Java ( the kind that is used to monitor and control the physical world ) has never taken off and will not take off anytime in the near future. Some of us in the Embedded Java community have been trying to tell Sun and others in the overall Java community about this problem for years, but almost nobody seems to get it.

The problem has been ignored up to this point because Java has done very well on servers, cell phones and some other places. Therefore, letting the embedded market slip away has caused the overall Java community very little pain.

But now that the Telemetry Internet wave has started to build, everyone is beginning to see the huge, lucrative opportunity that is associated with it. Telemetry is the first technology which allows the Internet touch the physical world on a wide scale and, in theory, Java should own a significant portion of the Telemetry market. However, the reality of the situation is that a completely separate community controls the Internet's ability to access the physical world. Not only does this community not understand Java, most of its members do not even have enough in common with typical Java developers to effectively communicate with them.

Unless the Java community initiates strategies ASAP to overcome hurdles like this that are blocking its access to the physical world, it risks missing its change to take advantage of the lucrative market that the Telemetry wave is providing.

On a positive note, some of us Embedded Java developers have been working to find solutions to these problems for years. In my next blog, I am going to describe some solutions to these problems that I have been working on ( James has also been working on various solutions and you can visit his site to see what he has created ).

For those of you who would like to see a preview of one solution I have been working on, you can burn the .iso image that is on this site ( https://jdos.dev.java.net ) onto a CD and boot a PC with it :-)

Ted Kosan
tkosan@dev.java.net

Telemetry: The Next Big Internet Wave

In 2002, Sun's CTO ( Greg Papadopoulos ) indicated that Telemetry would be the beginning of the next Internet wave and this wave is now washing across the Internet. Greg discussed the concept of Internet waves in a talk titled "The Next Big Thing" during Sun's Analyst conference that year and here are 3 slides from that talk:

Waves of the Internet

These slides indicate that the first Internet wave consisted of an "Internet of Computers" and the second wave, which we are currently in, is an "Internet of Things that Embed Computers". The third Internet wave, which is an "Internet of Things", consists of physical objects like thermostats, switches, packages and cloths.

After presenting these three Internet waves, the slides progressively show what effect the waves have on clients, functions, protocols and organization. Overall, I think that these slides represent some of Sun's best thinking on how the Internet is going to evolve and I am surprised that they are not displayed more prominently on Sun's website.

In the "Functions" timeline, I have circled the word "Telemetry" in red and labeled it "The next big Internet wave" because that is what I want to discuss in this blog. I think that the Java community is well positioned to take advantage of the opportunities that the Telemetry Wave will bring, but it has yet to put good strategies in place for doing so. One of the main barriers associated with telemetry that the Java community needs to overcome is that a completely separate community is in control of the devices that allow the Internet to access the physical world.

Of course, these devices are based on microcontrollers like PICs, AVRs and ARMs. The community that is in charge of these devices consists of the smaller developer communities that have formed around each of the popular microcontroller families. The problem that the Java community has is that almost none of these microcontrollers are programmed in Java when they are used in situations where they are monitoring and controlling the physical world.

This may seem like a small problem, but it turns out to be a whopping huge one that the Java community has not even begun to address yet. If this problem is not addressed soon, it is my opinion that the Java community is going to completely miss its chance to take advantage of the opportunities that the Telemetry Wave is making available.

In a future blog, I am going to relate a story which shows what happens when typical microcontroller developers are exposed to Java for the first time. This story shocked me when I first heard it ( in fact, it still does! ) and I think it will shock you too :-)

Ted Kosan
tkosan@dev.java.net

One Small Change for a Page, One Giant Leap for J2SE

If you have not visited the http://java.sun.com/j2se page recently, I suggest you go there and take a look. Hmmmm, "Core Java"... yes, we all know about that. "Desktop Java"... yes, wonderful technology, many of us use it every day. "J2SE Embedded"... what!? J2SE Embedded? Where did that come from?

While much of the Java community has been busy battling in the J2EE/.NET wars, the Law of Accelerating Returns has been relentlessly pushing microelectronics into the microcosm. All of a sudden, J2SE 5.0-capable embedded systems about the size of a dollar bill are about to hit the market and I think this has the potential to cause a significant paradigm-shift in the overall Java community.

J2SE-based embedded systems are about to open a huge and lucrative "sensor to boardroom" market for pumping process data into J2EE systems. I think that enterprise developers who jump on this trend early, and start expanding their skill-set to accommodate it, should be well positioned to take full advantage of the opportunities it is going to generate.

How can early mover J2EE developers prepare themselves to take advantage of these new embedded-space opportunities, you ask? Luckily, the ability to facilitate inter-community communications is one of the many areas where our wonderful java.net community really shines :-) Development efforts related to this new class of J2SE-based embedded system are currently occurring inside of the java.net Embedded Java community and we would be more than happy to have J2EE developers drop by and discuss this further with us.

For anyone who is interested in learning more about the new J2SE-based embedded systems and the opportunities they will be making available, feel free to subscribe to the new embjava email list and start asking questions :-)

Embedded Java - the Mouse that is about to Roar

Embedded Java Needs to be Leveraged Very Differently than Most Other Java Technologies.

One of the main reasons that Embedded Java has not enjoyed wide-spread success to date is that it is a significantly different technology than almost every other technology in the Java ecosystem. Due to this difference it did not benefit very much from the strategies used to market Java in general. Beyond this, the high growth rates that the other areas of Java experienced since Java's inception made the task of figuring out how to market Embedded Java a low-priority item.

A precise way to explain why Embedded Java did not grow like most other areas of Java is that Embedded Java lives within a different fitscape than they do. As Max Goff explains in his new book "Fitscapes and Fallacies", a fitscape is a "fitness landscape" within which a technology "lives" and it consists of all the constantly changing environmental factors that a technology needs to adapt to in order to be marketable.

The Embedded Systems and Enterprise Systems Spaces - Oil and Water.

When it became evident around 2000-2001 that Embedded Java was not experiencing the growth rates that it should, various members of the Embedded Java community started researching the cause of this and the following diagram illustrates what was discovered:

The Embedded Java Community is a Bridging Community

This diagram is deceptively simple and yet extremely important because it not only explains one of the main reasons why Embedded Java has not been very successful to date, it also shows how Embedded Java can give Enterprise Systems companies access to a huge, new lucrative market.

Critical Pieces of Information that are Indicated by the Diagram

1) The Embedded Systems community is almost completely blind to the contents of the Enterprise space and the Enterprise Java community is equally blind to the contents of the Embedded System's space.

The main reason why Embedded Java has not found a strong market to date is that Java is a product of the Computer Science community and since 90%+ of the Java community has a Computer Science-only background, they are barely even aware that the Embedded Systems space exists. Computer Science does not give the background needed to see that unique strategies are required for marketing Embedded Java in the embedded systems space. For a more detailed explanation of the differences between these two spaces see:

The Embedded Java Community is a Bridging Community ( more detailed version )

2) A huge, lucrative sensor-to-boardroom market is currently being created that very few of the companies that live exclusively in either the Embedded Systems or Enterprise Systems spaces are in a position to pursue.

For decades the Embedded Systems and Computer Science spaces have evolved almost completely separately from each other. Internet technologies, however, are finally forcing the communities that live within these two areas to start learning about each other in order to be able to solve the "sensor to boardroom" imperative that is beginning to pressure companies at all levels.

The rising levels of global competition created by the Internet is relentlessly driving all organizations to operate increasingly more efficiently and this is forcing companies to automate the monitoring and control of all of their operations. Due to the orthogonal nature of the Embedded Systems and Enterprise Systems areas, however, sensor-to-boardroom technologies are extremely difficult to develop because these technologies span both of them. This scenario is aptly described by the saying that "oil and water don't mix".

The Enterprise Systems community is currently in the process of building out the sensor-to-boardroom vision but they are almost completely blocked from accessing almost any aspect of the physical world for monitoring and control purposes due to the Computer Science-only background of most of its members.

Conversely, the Embedded Systems community can easily access almost any aspect of the physical world for monitoring and control purposes but it has almost no understanding of object-oriented Enterprise Computing technologies. Furthermore, this community has little understanding of business as a whole, the skills needed to turn a profit or the nature of the increasing levels of global competition that the Internet is creating. In short, the Embedded Systems community is almost completely unaware that a sensor-to-boardroom imperative even exists.

3) The Embedded Java community is small but potentially very important because its members have an intimate knowledge of both areas.

By definition, an Embedded Java developer not only has a solid understanding of the Computer Science concepts upon which Enterprise Systems are built, they also have a strong background in Embedded Systems. The Embedded Java community is in a unique position to act as a bridge between the Embedded Systems and Enterprise Systems communities thereby enabling the creation of sensor-to-boardroom systems which are required to span both spaces.

4) The initial companies in the Enterprise Systems space that have the vision to realize the unique nature of the Embedded Java community, and to leverage it, should be able to capture a significant portion of the sensor-to-boardroom market.

Of all the companies which live in the Enterprise Systems space, Sun Microsystems is probably in the best position to leverage the Embedded Java community for the purpose of pursuing the sensor-to-boardroom market. Even though the Embedded Java community has remained mostly unnoticed by Sun to date, members of the Embedded Java community have paid close attention to Sun's vision of how network distributed computing should develop over the next 10 years.

If, however, the larger Enterprise Space companies like Sun do not start moving into the sensor-to-boardroom market soon, the smaller more agile companies in this space, like JBoss Inc., are likely to make the first move and capture the initial market.

Stay Tuned...

Various members of the new java.net Embedded Java community are in the process of developing a completely new kind of Embedded Java system called "Enterprise Outpost" which is specifically designed to help companies crack open the sensor-to-boardroom market. This new embedded system is built almost entirely with open source and open hardware technologies and it is based on J2SE 5.0

Over the next couple of months I will be using this blog to describe the Enterprise Outpost concept and how it can help companies solve the sensor-to-boardroom problem :-)

On The Edge

Believe it or not, this blog account was set up in April of 2004 but I am just now making my first entry! The reason for this is that I wanted the central theme of this blog to be related to Embedded Java but there was not an Embedded Java community on java.net to associate it with... but now there is :-)

On Monday January 31st a new Embedded Java community will make its appearance on java.net and the community will be co-lead by Dan Johnson and myself. Due to the java.net scheduled maintenance on Feb. 2nd, however, we decided to delay officially announcing the community until after the site was back up. Until then, I will mostly be making experimental blog entries in order to become better acquainted with the java.net blogging system.