Skip to main content

Java One Day 0

Posted by cayhorstmann on May 7, 2007 at 11:44 PM PDT

??? style="float: right; margin-left: 1em;" />This year, I got a genuine press
pass from a kindly soul at Sun, which entitles me to priority seating at
the keynote, the opportunity to ask hard-hitting questions in the press
conferences and unlimited free booze in the press lounge (I hope).


??? src=""
style="float: left; margin-right: 1em;" />The day started with an 8am
conference call for Java
, an august group of Java evangelists. Bob Brewin was
supposed to tell us all about href="">F3, erm, JavaFX,
except he got cold feet at the last minute, and he only gave out vague
generalities about the resurgence of web-delivered desktop apps. I
practiced my “reporter asking hard-hitting questions” persona
and asked him why applet/Web Start/JVM versioning on the desktop was in
such sorry state, and why it didn't get enough engineering attention at
Sun. He agreed that this was a problem. href="">This
InfoWorld article has the actual JavaFX news.


Last year, there was a free NetBeans day at the Argent, erm, Westin San
Francisco Market Street Hotel. It was very crowded, so I was glad to see
it moved to Moscone this year. There were several other tracks too, on
GlassFish, Linux/Solaris, and Web 2.0. The NetBeans track. had a slick
presentation covering GUI builder improvements in NetBeans 6 and a nifty
demo involving href="">JXMapViewer.
This isn't your grandparents' NetBeans.

Another NetBeans presentation showed the Ruby support. I must admit
that I have never quite caught the Ruby/Rails train. Ruby looks nice
enough, but it doesn't really solve a problem that I have. People told me
that duck typing would bring me freedom. I tried using Groovy for some
data manipulation work, figuring it would be easier to pick up than Ruby.
I ended up going back to Java. Sure, Groovy has a prettier loop for
iterating through the lines of a file, but in the end I preferred
autocompletion in the IDE and compile-time error checking.

Rails addresses a pain point (Java web applications), but I was never
able to get past the example that shows

some Ruby loop structure 

JSF may be no walk in the park, but at least I don't have to worry
about HTML table tags—I just use a h:dataTable and bind it
to a list of objects. What would I do in Rails if I had to stripe every
second row with a background color? Write nested loops???

Sure enough, the same code showed up on today's slides.

There was a demonstration of a Rails debugger that lets you set
breakpoints in a mess of HTML+Ruby code. But only if you use the right
kind of debugger. That looks like work in progress, but overall it is
pretty amazing that you can do so much Ruby work in NetBeans.

In the last talk of the day, a fellow who had been programming Java for
all of seven months showed off his href="">SunSPOT powered app that rewarded
his dog with food pellets. What did he use for his web UI? The NetBeans
wizard for JSF.


The No Fluff folks
hosted a presentation on Groovy and Grails at the trendy W hotel. I was
glad I went. Much better food than the awful Moscone fare. (Before today,
I had never had soggy pretzels.) Nice presentations by the Groovy and
Grails gurus, Guillaume Laforge and Graeme Rocher.

I really want to like Groovy. It looks like Java and it is tightly
integrated into the JVM. My problem is that I don't understand it. When I
first saw it a couple of years ago, I found out to my horror that there
was no grammar. The only way to check the meaning of an expression was to
feed it to the interpreter du jour. They fixed that now. A few weeks ago,
I used Groovy as an href="">example for
metaprogramming in my graduate programming languages course. I had a
nifty lab where I added a method to java.awt.Point. Trouble was,
the method was not inherited. I spent a few hours looking at href="">scarce
documentation and a complex implementation. I was unable to find a
spec for the MOP. Being the hard-hitting reporter, I cornered Guillaume
who cheerfully agreed that no such spec existed. The MOP is whatever it is
in the implementation du jour. These things make me appreciate the Java
culture of “one specification, multiple implementations”.