Head On

I am seeing signs in the sky that Sun is making a comeback.

• Textdrive is switching from FreeBSD to Solaris
• Rumors that ZFS might be ported to Mac OS X
• Tim Brays presence and statements on the RoR podcast from Canada on Rails
• Sun makes smart moves like supporting PostgreSQL

And not one of them is related to Java.

It seems that Sun is using James Gosling to spread FUD regarding any technology that competes with Java. He created Java, and he is a techie and techies have high standards and are always honest - right?

He came out slinging against Harmony a year ago, and now he is at it again with Ruby and PHP:

PHP and Ruby are perfectly fine systems, but they are scripting languages and get their power through specialization: they just generate web pages. But none of them attempt any serious breadth in the application domain and they both have really serious scaling and performance problems.

He is either grossly misinformed or he is making things up. Either way it is time to stop listening. Too bad.

(Via Loud Thinking.)

I have 16 minutes left of my prepaid wireless connection, and I don’t feel like paying more for it. I have covered a few sessions for which the amount of text is too much for the network to handle, so I’ll post it when I get home. I’m right now listening to Scott W Ambler who is really an excellent speaker. More on him later.

Mule is a very different beast to me. It seems to be a swallow anything messaging bus where you can message anything anywhere using any protocol anytime. It seems really cool - I just wish I knew what to use if for. This is not at all to say that it has no use, it is just me that has not worked with anything similar before.

BTW: Demos involving messaging are almost always cool.

Linda DeMichiel together with Mike Keith again walking us through the new EJB3 API. I must say that I like the idea of the one, true, unified O/R mapping framework. Also providing a decent set of default values id really smart - the table mapping is reminiscent of Rails. I wonder what all the app server vendors think of the fact that it can be run outside of EJB? This is another reason that you most probably can get by with Spring in most cases which really must upset the app server vendors and especially JBoss.

This was another packed session, even though according to raised hands less <10% of the audience actually uses JSF. Maybe people where there gathering arguments to not use it? The talk was very detailed and I can’t help think that JSF seems to be incredibly complex. One advice to get JSF right was to “avoid HTTP thinking”. That is the exact reason why I don’t like component based web frameworks.

I’m a little disappointed with this one as I didn’t learn that much from it. As I managed to get one of the coveted seats next to a power outlet i am still here though :).

The room was packed when Javaland superstar Rod Johnson presented the new features in Spring 2.0. The new XML tags are quite nice as they reduce the amount of configuration code.

Rod mentioned that the JPA implementation in Glassfish will basically be TopLink out of the box, making this a truly usable reference implementation.

Apart from the RAD race price thingy, Graham Hamilton of Sun presented Mustang and Dolphin, the upcoming Java releases. What cought my interest the most is that Mustang will include Rhino, and Dolphin will provide a completely new byte code format for dynamic languages. Could it be that this is also related to BEA’s future support of “languages that might take off”?

Wicket provides a really excellent programming paradigm - plain HTML templates, really nice Java and no XML what so ever. My feelings are mixed though, as it is component based with weird URL’s. Also it uses the session by default. I really have to think about this one. I’m not that big a fan of component based web frameworks as I think that they are abtract away too much of the real world of stateless requests and responses.

This talk had Jason Carreira presenting WebWork. It was good getting information regarding the WebWork/Struts merger straight from the source.

In the Q&A session, people asked a lot of questions about tool support. This surprises me - why do people find this a necessity? Are people actually like Bill Roth of BEA who said in his keynote that he can’t write XML manually? Of course, he sells tools.

XOM Design Principles with Eliotte Rusty Harold. Eliotte is a very good opionionated speaker. While I do not agree with his view on the topic on humane interfaces, he does present some really good ideas.

Quote to remember: “One out of three untested lines of code has bugs.”

EJB intro with Linda de Michiel. I remember listening to Linda presenting EJB2 at JavaOne 2000 saying that now they had fixed everything that was wrong with EJB1. I must say that is quite a deja-vu listening to her talking about how EJB3 fixes everything that was wrong with EJB2.

The first talk I attended was More Java Puzzles with Joshua Bloch and Neal Gafter. While I normally find these kind of questions way too C++-ish (”can you tell what this piece of unreadable code does?”) Joshua and Neal presents it very well. I got four out of eight right by the way.

Javapolis is held in a combined conference center / movie theater complex. This makes for very comfortable chairs, really good audio and no daytime sunlight what so ever. So far the talks I’ve been to have been really good.

It seems that my problem with wifi jas something to do with size - I can send short emails, but not long. Also, it seems that I have to keep my posts short to get them to my server. This for €20/24 hrs

The wifi access here at Javapolis is really weird. Surfing the web works like a charm, so does receiving mail. Sending mail seems to be a hard thing to do, as it works half the time. However, when refreshing NetNewsWire, I get no updates. That combined with the fact that I can’t post with MarsEdit makes me suspect that Proximus has some really funky settings for their wireless network.

If you have a Mac and want to use the Proximus wifi at the Metropolis, Safari or Firefox won’t help. I got through using Omniweb.

So I arrived yesterday to Antwerpen, and have just sat through the keynote. It was basically a shameless plug battle between Sun and BEA. The most interesting thing was Bill Roth talking about WebLogic “in the coming years” supporting “languages that might take off”. This is in line with what Jon wrote some time ago. Rails on WebLogic? Could be interesting, if not for no other reason but to gain a wider acceptance for Rails.

Whilst on the topic of aggravation:

/**
 * Returns the url for the link
 * @return the url for the link
 * @param oFooUtil the foo util
 * @param sSomething the something
 * @param iAge the age
 */
public String getRelationPath(FooUtil oFooUtil, String sSomething, int iAge) {
    ...
}

1. Do not name your methods deceptively and keep the correct information in the javadoc only
2. Do not use hungarian notation. It is extra useless in Java as it is an object oriented language which means that many, many variables will be prefixed with o.
3. Do not comment the obvious

I’ll be attending the conference part of Javapolis. Anyone else?