Jeff on Eclipse

I was thinking the other day, “Why am I running for re-election in the current Eclipse Board elections?” It’s really quite a bit of work. There are monthly phone calls, quarterly face to face meetings, committees and working groups, prep for all of the above, … Oh my!

All in all I figure it takes at least 5 days a quarter just to keep up. If you want to help drive things, well, you could easily double that. When you figure there are about 50-55 work days in a quarter (depending on your location’s holiday calendar and your vacation time), that’s at least 10% of your normal work time!

So what’s up with that? We have 8 dedicated committers all willing to contribute still more of their time to help run the foundation. Why?

I can’t answer for the others but I can list some of the reasons I’m on the list.

• The other board members. I’ve mentioned this before but the set of people on the board today is frankly, superb. Diverse, talented, bright, engaged. You can’t pay for a better set of teachers.
• Committer members have an impact. We have day to day knowledge of how the committer community thinks and works, and the kinds of challenges they face. Because of that, we ground the discussions in reality and focus on the things that help committers most.
• Committers don’t know it all. Gads! But its true. Eclipse has a very vibrant commercial aspect that, to be honest, I did not know or particularly care about when I first became a director two years ago. It is extremely interesting to understand. More committers should take some time to understand how their companies and others make money with/from/… Eclipse. The board spends quite a bit of time on this.
• The board meetings. Two days, locked in a room with 30 other people, no internet access, … Wait, that was not one of the pros…

Anyway, we’d be here all day. There are more reasons and many are outlined in my vision statement. The point is, as I went through the list, I quickly realized that being on the board is a huge opportunity that excites and inspires me. That is why I am willing to donate this time.

Eclipse is many things that are good and can be many other great things. The Board of Directors’ mission is:

“to advance the creation, evolution, promotion, and support of [Eclipse]”

There is no better place to be than on the board, if you want to have a hand in shaping the larger Eclipse world. I look forward to another term on the board and more interaction with the entire Eclipse community.

This is a bit off my normal discourse but, well, it may be of service to someone out there.  I recently (yesterday) received a new Lenovo T61p laptop.  Sweet.  1920×1200 screen, fast processor, big drive, …  I’ve spent the past couple days installing stuff and getting it setup. For the most part it has gone swimmingly.  For the most part.

Last night around 1230 I was finishing up some installs and decided to go to bed.  As is my custom I suspended the laptop, got up from the desk and headed off.  There was the normal “going to sleep beep” and as I walked away, the machine beeped again.  It was waking up.  Strange.  It had been suspending fine earlier. 

This sequence was repeated again and again over the following 1.5 hours always with machine waking up.  I tried all manner of things (wake on lan settings, turning off wireless, …).  No joy.  I was starting to believe that it was the physical act of getting up and leaving that was causing the machine to wake.  Enough.  Haunted laptop?  I turned the machine OFF and went to bed incredulous and confused.

This evening the experiments continued.  In the end, after another 2 hours of suspend and resume experiements witnessed by an outside observer (wife), the problem was tracked down to a combination of a power management setting on the Mouse Device and my chair!  No kidding, the chair. 

The mouse was set to “Allow this device to wake the computer”.  Makes sense that that would be a problem but I was being very careful not to touch the machine in any way thoughout the experiments.  In fact, wiggling the mouse, banging the desk, … did NOT wake the machine.

That’s where the chair comes in. It turns out that my chair makes a bit of a thunk sound when I get up and that particular vibration was triggering a mouse button/wheel and waking the machine.  Unplug the mouse (before suspend), OK.  Get up slowly/quietly, OK.  Mouse plugged into any of the three USB ports and the machine wakes up when I stand.  Again, I kid you not.

There is probably a lesson in there somewhere about assumptions and side effects.  I prefer to just suspend the machine and walk away shaking my head at how complex things are.

I remember a few weeks ago seeing a message about the Memory Analyzer project proposal. Some guys from SAP were going to open source some of their tooling to analyze Java heap dumps. I didn’t think much about it at the time other than “Hey, that’s probably a good idea”. The other day however, Andreas Buchen, the proposed project lead contacted me about being a mentor so I took a deeper look at what they are proposing.

There is a bunch of info on the SAP Wiki and Andreas has summarized it well in a newsgroup post. This is really cool! Their tool allows you to do all manner of queries and introspection of heaps.  You can look for leaks (why is that object still there?), check memory consumption (aka retained size), create your own queries, etc.

The real kicker for me was the classloader relative searching.  Under the covers, Equinox makes a classloader for every bundle.  So searches by classloader are just a step away from searches by bundle.  This puts the tooling in a much better context for Eclipse developers and, for example, lets you figure out what your bundle is doing with memory etc.

The other thing that is cool is its scalability.  They routinely work on multi-GB dumps and have a whole range of instant queries and fast approximations.

The Memory Analyzer project is quite niche and may not attract that many developers (there is lots of potential for people adding extensions) but as part of the broader suite of JDT and TPTP tooling it is a big win and will attract lots of users.  I encourage you to check out the proposal and show your support on the newsgroup.  Also, you might want to attend their talk at EclipseCon.

p.s., Don’t forget early registration for EclipseCon ends on Feb 14.  Spend your money on your loved ones not late registration for the conference.

What the heck? “Eclipse” and “runtime” in the same sentence? Most people think of Eclipse as a tooling thing. Or perhaps an RCP thing. Well, take a look at the EclipseCon program and you will see that Eclipse is that and a whole lot more. I thought selecting talks on the program committe was hard but choosing which talks to actually attend is going to be even harder!

To see what I mean, go to the EclipseCon site and pick any of the days on the left. Then scroll down to the track selector and check only Eclipse as a Platform, OSGi DevCon and Rich Client Platform. Now look at the program. It doesn’t matter what day you pick, you will be double or triple booked pretty much the whole time. It is going to be a very busy week.

To make matters harder, I found many talks related to runtimes and platforms that did not fit nicely into the canned tracks. Here are a few that caught my eye:

• Eclipse Riena Project
• eRCP vs Android et al
• Swordfish
• Eclipse Command Language
• Jazz Ajax Framework
• RAP vs GWT
• API Tooling

Ok the last one is tooling and I am biased by involvement but how could a runtime guy not jump with joy about tooling that helps makes your modules more modular and backward compatibility a reality?

Anyway, all up I counted about 25 long talks (including those above) and about 40 short talks related to Equinox, Eclipse in various runtime/platform guises, OSGi, etc. Hope a lot of them are being recorded…

Don’t miss the fun, go register before the Valentine’s Day deadline and spend the money you saved on something for that someone special. See you there.

Oh, and for you committers out there, don’t miss the Committer Hackathon on all week.  Get your team together and hack like crazy.

It is that time of year again — the elections for your committer and add-in representatives to the Eclipse Foundation Board of Directors is here. This year the list of candidates is most impressive. Last year the committer rep candidate list was unfortunately dominated by IBMers (including me!). This year less than half the candidates are from IBM and Mik and me are from small Eclipse-related startups (yes, I left IBM). The better breadth of experience and focus is greater this year as well. There are candidates from the embedded, web and tooling platform areas of Eclipse. It would be good to have more than me waving the runtime flag but that will come as Eclipse broadens this year.

Regardless of what part of Eclipse you focus on, please check out the various mission statements and see which fit best with where you think Eclipse is and where it should be going. The coming year is going to be very exciting with lots of new projects, directions and challenges. Voting in the election is one way that you can help shape those changes. Voting starts on February 25th and goes until March 7th.

Ottawa is unusually quiet today. People are burrowing in for a winter storm that is coming up from the US. A quote from Environment Canada

A low pressure area over Arkansas is getting organized and setting its sights on southern Ontario

I would have thought that the Arkansans would be too busy organizing Super Tuesday to bother sending weather up here!

One group of Ottawa inhabitants that is apparently not going to experience a storm is the Foundation IP team. Last year 154 non-EPL Contribution Questionnaires (CQs) were submitted to IPzilla in the month of January. 122 of those were submitted on the 30th or 31st! While past performance does not necessarily represent the future gains, the team has been stocking up on water, Tums and crates of patience. It may be all for naught. As of a few minutes ago only 75 CQs had been submitted for the entire month and the queue has only 58 pending.

This is great if its for real. Its bad if there is a storm brewing just out of radar/satellite range. The day is not over until midnight but it seems unlikely that committers around the world are poised to submit an onslaught of CQs. More likely is that committers and project leaders have forgotten about the deadline. Also possible is that we have simply identified and cleared all the interesting third-party code out there.  Uhhh, no.

Seriously folks, if you have not submitted CQs for the things that you think you will need for Ganymede, please do so now. The IP team really needs to know what they are dealing with over the next few months so they can schedule and prioritize accordingly.

I tried to come up with some jokes and humorous stuff to put in this post but nothing seemed to work. So here it is, straight out…

After a lot of thought I have decided to leave IBM. I’m leaving the team after more than two decades of enjoyable, challenging and exciting work first at OTI then, after the acquisition, at IBM. As you can imagine this was not an easy decision. Many of the people in the world I continue to call OTI are friends from university and we literally grew up together. From covert beginnings in a townhouse (the only one on the block with 10Base5 networking) just across from where I live now to chocolate mousse at Montebello to acquisition and assimilation by Big Blue. We have been though a lot together. I will certainly miss them.

But wait, I am not leaving entirely; I am not leaving Eclipse. As one of the people who developed the first prototype before it was Eclipse, I am deeply committed to the technology and the community. So much so that I am putting my money where my mouth is and starting a company around enabling success with Eclipse, Equinox and Eclipse in the runtime space. To that end, my plan is to stay involved in Eclipse both as a committer and my current project leadership roles.

One of my goals is to drive still more participation in Eclipse whether it is me personally, my company or vicariously through others. As with the evolution of Eclipse in the RCP space, we are now entering the runtime space and Equinox is particularly well-positioned to make a real impact.

Oh, and in case you are wondering, yes, I am planning to run for re-election to the Eclipse Board of Directors. My work there over the past two years has been extremely rewarding. I learn something with every board interaction and believe I am having a real impact on the board and the quality of life at Eclipse.

So, exciting times ahead. Look for me at EclipseCon or out on the boat with my kids.

Recently I was reviewing the Equinox project, the direction and the community around it.  One of the things that stood out for me is that Tom Watson has been quietly leading vast swaths of the activity in Equinox from the framework and service implementations to the OSGi spec work and areas in between.  Not only does Tom produce great code and designs but he has a great sense of the community and repeatedly performs heroic feats to meet their needs.

It is with great pleasure then that the Eclipse PMC invited Tom to co-lead the Equinox project and he accepted.  Congratulations Tom and thanks for all the great work!

I was pretty excited to be on the EclipseCon program committee again this year. It’s interesting to see all the different submissions and talk with the other PC members, getting their insights as to what is important and interesting. This year I was completely blown away with the volume and high signal-to-noise ratio of the talk submissions.

An initial set of acceptance messages has been sent out and we are hoping to get some more space so some additional talks can be accepted. The first draft of the program is available as a list of talks (scheduling has not been done).  You can click on the track names to get this list of talks in a particular area (e.g., OSGi DevCon, RCP or Eclipse as a Platform). Early registration is open and presents a huge cost savings (hint hint).

See you there.

The other day RedMonk analyst Michael Coté and I chatted about Equinox, OSGi, components and what it means to use components to building modular systems. We had a great time and ended up with a two part videocast. The first part is a discussion about the technologies and issues while the second is a demo of how to structure component based code using a real-world example from p2, the new provisioning work in Equinox.

Since these are screencasts you might want to view it in full-screen mode. A wide range of formats including MP4 and audio only are available on blip.tv for both part 1 and part 2.