First off, they have the same kind of custom test suite that we do. They also have the same issue of having to test "from the outside" (as opposed to true unit testing) as we do. There are several interesting differences, though.

They do not do formal code reviews like we do. They have the same kind of review document that we do, but their developers fill it in for themselves.

In addition to the nightly and weekly builds and testing, like we have, they have their version control set up so that an update will not be accepted until it has survived a full test suite run. The gentleman with whom I was speaking said that since implementing that policy, they have not had any issues with broken builds. He also said that since implementing that policy the quality of the product has gone up remarkably, to a degree that the quality upswing has been noticed by their customers.

Just some food for thought.

One of the comments I've heard was that this workstation offer may be the key to breaking this customer's IBM shop out of the IBM stack and getting them onto NetBeans, Solaris, and Java Enterprise System. Another attendee told me that this deal is going to change the world, because developers, who need the software anyway, are going to be able to bring home this box, bootstrapping his or her family into the participation age. An important caveat to that comment was that the box must include StarOffice. I would tend to agree.

I must say, the Ultra 20 is an attractive box. The common consensus among the Sun employees with whom I've spoken is that factoring in our employee discount, there is no better hardware deal in the market, PC or otherwise.

(This may seem like something obvious, but in previous years the chaff content was very high. I had reached the point where I wrote off all non-Sun presentations as useless product pitches, and most of the Sun presentations as too high-level. This year, the best sessions I've attended were from companies other than Sun, and all of them have dived straight into useful technical content. I'm totally stoked!)

I'm also stoked about the eval forms. For two years, the JavaOne folks have been promising to start weeding out repeat speakers based on feedback. Last year there was apparently a system for evaluating sessions online. I never managed to find it, though, so I doubt it was very well utilised. This year every attendee gets handed an evaluation form when he or she walks into a session. The form asks some good questions about the speaker, the slides, the overall quality, etc. Hopefully folks are filling the forms out and returning them.

Perhaps the most exciting thing at JavaOne this year is that every other hour the JavaOne staff is holding a brainstorming session to discuss what they can do better for next year. That is huge! We're actually listening to what the attendees are saying! This is something we should have been doing from the beginning, but better late than never.

I think the increase in the geek percentage of the population is also attributable to the recent history of JavaOne. Five years ago, the conference hit its peak, with an unofficial attendance of 30k people. Back in the those days, the speaker requirements were much looser, and in my opinion, the session were much less useful. After the .Bomb exploded, the conference attendance dropped well below 10k and stayed there. This year is the first year since the crash that has had a solid attendance, and it's happening concurrent to JavaOne's new technical focus. The net effect is that we burned off the wannabes with the crash, and now we're rebuilding with a high geek factor. Long live technical content!

This year I've also noticed that there are a lot more women here than in previous years. Based on just looking around, I would say that in previous years the attendee population was around 2%-4% female. This year, it looks more like 6%-8%. Again, that's just my opinion, formed from doing a rough head count every time I'm in a crowd, but that's 2-3x increase. Now, the interesting question to me is 'why?' I could speculate, but I haven't yet come up with a defensible explanation. (By the way, I'm not the only one who has noticed the increase in the female JavaOne population.)

Something else I've noticed is an increase in the European population here. I could easily believe that since I now live in Europe I'm just more tuned in to the European community, but it could just be that there really are more Europeans here than in previous years.

During the first keynote, John Gage did his old demographics demonstration where he asks everyone to stand up and then has people sit down based on the number of years they've attended JavaOne. There were a lot of first-time attendees. Not surprisingly, though, most of the people who had been to JavaOne before, had been coming since 2000. We didn't pick up a lot of new community members during the layoff years.

Based on the 2003, 2004, and 2005 attendance, and the recent changes to JavaOne, I feel really good about what JavaOne 2006 will be. I think we're on the road to (re)building a very strong Java community around JavaOne.

Yesterday, I only managed to go to one presentation, but it was an outstanding presentation. Experiences With the 1.5 Language Features: Tips and Techniques, by Jess Garms and Tim Hanson from BEA, was a tour through the new features of J2SE 5.0, along with best and worst practices. I've been doing J2SE 5.0 development for a while now, but I learned quite a few new things in their presentation. In fact, I took notes so I can share the stuff with the guys back home. This presentation is a definite must-download when you sign up for JavaOne online.

One of my pet peeves is BoFs that don't involve the audience. BoFs should be about engaging the community and fostering discussion. All too often, BoFs end up just being bad slide shows with no interactivity.

A pitfall of an ask-the-experts-style session is that the audience may leave you hanging. In order to mitigate that risk, I had two pages of prepared questions. Much to my shock and (pleasant) surprise, though, the audience produced an endless stream of questions. In fact, in order to ask the prepared questions I did ask, I had to put audience questions on hold!

I would like to personally thank Dan Hushon (Sun), Richard Nicholson (Paremus), Mark Little (Arjuna), and Greg Pavlik (Oracle) for being on the panel and doing an outstanding job. I got lots of comments from folks afterwards saying that they learned a lot during the BoF. Thanks guys!

Me: Hi, honey. How was your day.
Wife: Read my blog.
<long pause>
Me: Wow. That's a bummer.
Wife: How was your day?
Me: Read my blog.
<long pause>
Wife: Well, it could have been worse.
Me: Read the post two posts previous.
<long pause>
Wife: Ah, it was worse.
Me: What should we have for dinner?
Wife: Read my blog.
...

At some point I think we're going to lose our language skills altogether, but we're going to be really good typers.

Perhaps I'm easily flattered, but that did it for me. First off, he's obviously seen me talk somewhere before because he recognised me. (The names on the JavaOne badges this year are very small and not readable from more than a couple of feet away.) Second, he cared about my lab! One of the difficult thing about being a lab author this year is that there's no way to know if anyone ever looked at the lab. Not only did he want to do the lab here at JavaOne, he had downloaded the previous three drafts versions of it from the Grid Engine web site!

It's just very cool to have actual evidence of my "message" being heard. I often feel like I'm preaching to the geese.

I am a back end guy. I consider GUI work to be on the same order as Chinese water torture. I will gladly do back-end work, from servlets, to grids, to socket-based, custom protocols, but I would rather have voluntary dental work than write a view component.

An important part of any demo is a pretty GUI. Since I don't do that, we had another developer on the team pull together the GUI. Unfortunately, the GUI wasn't built to be used as a demo; it was built for the prototype. It was functional, but not especially pretty, and not especially fast or stable. (In fact, it's a web page with seven separate servlets on it!)

Well, needless to say, the reaction to my new super demo system was less enthusiastic than I had hoped, which goes to show you that regardless of how wiz-bang your system is, it's nothing without an equally impressive GUI.

First off, he says that NetBeans has the best tools for web development that he's seen. Even though he's an Eclipse user, he fires up NetBeans whenever he needs to debug a JSP or do any serious web development. He also says that Eclipse is a memory pig. He says that on a system with a large number of plugins the IDE becomes quickly unusable. He said that for projects that are very large, that have a lot of classes, NetBeans consumes dramatically less memory than Eclipse.

On IBM in general, he says that the lack of quality is plainly evident. In an IBM mobile portal install, a team of IBM people took two weeks just to get a sample app working. They still haven't managed to get the production environment working. He said that his IT team is afraid to deploy Java EE^TM apps without a developer standing by, because the deployment tools very often break leaving a partial or broken deployment.

About the keynote this morning, he said that the most interesting part was the SDN subscription with an Ultra 20 for $29.95 a month. The reason is that they currently do their development on Eclipse on Windows and then deploy on WebSphere on Solaris 9. With the Ultra 20, NetBeans, and Sun Java Enterprise System Application Server, he can do the development on the same architecture as the production environment, making it possible to discover certain bugs before deployment that he cannot now.

Last thing he said is that in order to maintain adherence to the Java EE spec, he uses JBoss and Sun Java Enterprise System Application Server, Platform Edition to test deploy his apps. That way, if the developers use something WebSphere-specific, he finds it immediately. He says it also helps detect bugs he wouldn't otherwise find until production.

Very interesting. If his story is typical, our stock is very undervalued.

I got in last Wednesday, hit the ground, and crashed. Nine timezones and a 12-hour flight is enough to wipe me out completely, especially since I didn't sleep at all the night before the flight. (I didn't even go to bed; too much to do!) Thursday, Friday, and Saturday was the Software & Java^TM Ambassadors Conference, an internal Sun conference to share our product direction with a select group of engineers from the field. This year we had a very high density of information. I can't give you any details, but suffice it to say that we've got some great stuff coming up in the next year. We're announcing (and have announced) some of the stuff here at JavaOne, but a lot of it is still coming in FY06 or later.

I spent all day Sunday in the hands-on lab making sure that my Building Grid-enabled Applications lab works in the lab environment. There were quite a lot of minor tweaks and changes that had to be done, including updating all of the screen shots for NetBeans 4.1. After I got everything cleaned up and ready to go, I made the mistake of calling my co-presenter for the Building J2EE^TM Grid-enabled Applications Technical session (Thursday at 2:30!). We are have a demo that we are using for the session, and it was down. I spent two hours hanging from a pay phone while trying to talk Eric through finding and fixing what was wrong. (By the way, we failed.)

Today, I got to the Moscone center a little after 7:30. When I got there, there was already a line for the general session. It was only about 20 people, though, so I went on about my business, which was doing some speaker and lab preparation and getting some coffee. I was rather impressed to see the line in front of the breakfast room. It was rather long. By the time I got my coffee and got back upstairs, the line for the general session had reached around the corner. And kept going! I finally found the end of the line at the next street corner, in front of the Metreon theater. That was about 8:15. By 8:30, the line had reached around almost all the way to the next steet corner! I haven't seen a line at JavaOne like that in years! I felt like Chistopher Eccleston as Doctor Who, looking at the exceedingly long line that I was going to have to stand it, smiling broadly, and quietly muttering, "Awesome."

The general session was OK. Because of the SJA Conference, the announcements had already been spoiled for me. The level of enthusiasm was overall a little low. Last year was great because everyone was stoked about the new features of Tiger. This year, Jonathon and JonnyL didn't really have much exciting to say. (See the next paragraph.) Even the t-shirt hurling contest entry was a bit of a dud. The highlight of the session was when McNeally jumped up on stage and heralded the entry of Duke followed by a jazz band and a birthday cake. Even cooler, Gosling brought up on stage the original members of the Green development team. That was the only rousing sound of applause in the entire session.

Among the things announced was Project GlassFish, the open-sourcing of the platform edition of our app server. I thought GlassFish was huge, but the audience reaction was non-existent. No applause. Nothing. I talked to another conference attendee about it over lunch, and he said that with JBoss out there, it just wasn't an exciting announcement. It's nice to now have a choice, but he didn't see it as particularly thrilling. Interesting. The gentleman sitting beside me right now has a very interesting perspective. He says that from a product perspective, the announcement doesn't make any difference; companies don't want free software running in production. He says, however, what is going to fundamentally change the app server landscape is that Sun has just declared that they're willing to listen to the community about what a "good" app server really is. Interesting.

I'm off now to the Sun bloggers meeting. This will be my first year to make it. I'm very curious to see how many people show up.

Before I go, though, let me plug my BoF. I have a BoF panel on grid computing tonight at 9:30pm in the Golden Gate A3 room at the Marriott. I have some CTO's (and one CEO) there to talk about what grid computing means to the future of Java computing and computing in general. Come on over!

To correct this problem, I've started doing two things. First, every time I play someone, I look their opening up on the web. I can then read about it, learn how best to play against it, and potentially use it myself in the future. A friend showed me a neat trick for finding openings on the web. Go to Google, and enter the opening's notation as the search text. For example, if you search for "1. e4 e5 2. Nf3 d5," you will discover this opening is called the elephant gambit, and that it's not used very often because it doesn't ofter enough in return for the pawn.

The second thing I started doing is reading chess advice on the Internet. There are a large number of chess sites out there, many of which have useful advice sections. About.com's chess site has a neat feature where you can play through an opening, and it will tell you what it is. Unfortunately, though, their library is pretty narrowly focused on the mainstream moves.

The best site I've found so far, though, is Chess Kids. Yep. I'm learning chess from a site designed to teach 8-year-olds. There are 9 classes, with each class divided into 6-8 lessons. The first lesson is what the pieces are called and how they each move. Not exactly useful for someone who's been playing chess for 20 years. Fortunately, though, the lessons ramp up quickly. I'm currently in class 8 learning about the Sicilian Defense and the French Defense. In class 7, I learned about how to win or draw an end game with just kings and pawns. If you can get past being talked to like you're 8, this site provides a really good primer/refresher in the study of chess.

In a previous post, I mentioned that I was busting my butt on an important project with a short deadline. As luck would have it, this project involved setting up a complex Grid Engine cluster on a rack of Linux servers. Part of the complexity came from a series of scripts which Grid Engine was supposed to run in response to events in the cluster. For the purpose of this rant, the only two that are interesting are suspend.sh and resume.sh. By the nature of what I was doing, these scripts already had a low trust quotient. I had just written them, and they both kicked off some rather opaque processes, so I was anything but certain that they were doing what they were supposed to, much less in response to the proper events.

After I got everything put where it was supposed to be, I tested the system. Much to my surprise, the suspend.sh script actually did what it was supposed to do, when it was supposed to do it. The resume.sh script, however, didn't appear to do anything. Fine. I added a line to resume.sh that touched a file in /tmp so I could see that it was at least being called. Nothing. So I went through the Grid Engine configuration, looking for a reason why it wasn't working. Nothing. I tried running resume.sh by hand. Worked fine.

Then I started getting creative. I changed the configuration to use suspend.sh on both events (suspend and resume), and suspend.sh worked in both cases. Back to a problem with the resume.sh script. I ran it by hand again, just to be sure. Worked fine. I changed the configuration to run a third script (instead of resume.sh) on resume. Worked fine. I copied the contents of resume.sh into the third script. Worked fine! I removed resume.sh and made it a hard link to the third script, and switched the configuration back to using resume.sh. Nothing! I switched the configuration back to the third script. Worked fine!

In the end, the best guess I have as to what wasted 4 hours of my time on an extremely time critical project was that the NFS daemon on either the client or server somehow decided that the name, resume.sh, was bad and should be ignored. Gah!

I freely admit that if I had trusted the scripts and the Grid Engine configuration more, I would have found the problem much sooner. I spent a lot of time misguidedly trying to debug both of them. Nevertheless, the fact that I even had to deal with a ridiculous problem like that makes me never want to log in to a Linux box again!

Anyway, on to the point. Since we're moving, and since we're doing it from about 6000 miles and 9 time zones away, we're going to need to work with a competent real estate agent who's willing to work with us remotely. Anyone have any recommendations? We're looking in the area around Menlo Park. Anyone have anything good or bad to say about any particular neighborhoods? Any spurious advice of any sort?