Log in

No account? Create an account
18 May 2006 @ 05:56 pm
More JavaOne  
  • The Apple Java team closed a major bug last night, that has been outstanding since 2002 and had several hundred people CCed in the trouble ticket system. The drunken developer involved stumbled up to the page, spent a few minutes in front of the corwd looking for the right form fields, wrote out a note saying that he had fixed the bug and was totally plastered, then sent it to all interested parties. Of course, he failed to mark the bug "fixed", so he had to do a follow-up posting, also CCed to several hundred people.
  • DDsteps looks nice. I'll probably give that a try.
  • A lot of people want to eable inlining XML fragments into Java source. All the iterations I've seen look Really Ugly.
  • Lots of scripting panels, but there seems to be much less emphasis on Aspect-orianted stuff this year. It seems the "Aspect people" have either decided to make do with Annotations or move on to full-fledged scripting languages. I was talking with someone who called this the "Bruce Tate pattern". While there are a few Ruby, Rails, and PHP books in the bookstore, they are NOT carrying "Beyond Java".
  • Thirsty Bear brewery and spanish restaraunt, less than a block away. Where the afore-mentioned Apple guys do most of their drinking.
  • It still takes work to avoid the sessions that amount to hour-long commercials.
  • A few ways to do AJAX:
    • The Rails way: invoke a URL and replace the contents of an element on the current page with the HTML response.
    • The Sun way: Hand code a bunch of complicated API stuff, but abstract it into a JSF component for reuse.
    • The Google way: abstract the AJAX stuff into an asynchronous RPC structure, with object serialization.
  • Blu-Ray = Bastards. They held a panel on creating content for Blu-Ray in Java, but the development platforms cost 5 figures. I am NOT getting another mortgage in order to get my grubby hands on a windows-only kit.
  • The new invokedynamic bytecode, if added, will be the first new Java bytecode ever. If it gets into Dolphin, we should see it in 2008, which means it will be used by Groovy in 2007, by JRuby in 2008, and Jython in 2011. Jython 2.1 has been a few months in the future fr about 18 months now - not as bad as Vista, yet.
  • The Groovy development team has a bad case of ADD, grafting on neat feature after neat feature without any sense of prioritization or consistency.
  • Still hearing a lot of buzz about the Google Web Toolkit thing. Even though the downloads are only for linux and windows, I have it working on my Mac.
  • Two of the frameworks blatantly "inspired by" rails are giving presentations tonight: Grails and Trails.
  • Party tonight, as described by James Gosling: " The Mythbusters will be there, mostly doing robotic stuff (sorry, no exploding cement trucks in public places). And we're doing the t-shirt hurling contest. Fun with cannons and chainsaws.... And the Brazilians doing performance art..."
Ryansynx on May 19th, 2006 05:33 pm (UTC)
8 months
It took me 8 months to become disillusioned by java. My position now is "why bother learning java - you're catching up to the past, why not skip ahead a generation or two and work on the next big thing". Of course I don't know what that is yet, but I have my eye on functional languages.

Java was great for it's time, but its too complex for new users. Enterprise development doesn't have to be complex - it's a myth that you need huge things to crank out that billing system. I guess people just forget KISS after a while.
St. Sean the Amusedseanb on May 19th, 2006 07:23 pm (UTC)
It took you 8 months?
Agreed. Java was a mediocre leap forward in '95, but it feels old and crufty now - it's now older than C++ was when Java came out.

Add to that Sun's tendancy to write APIs that only IBM wil lwant to use and you have a complex, ugly, backwards mess. Sun knows they have a problem, and they are trying to improve - simplifying the APIs, embedding scripting engines, and trying to give nice solutions to the 80% simple problems instead of just for the nifty edge cases that expert groups and lead engineers love to dwell upon. Java is improving, but it doesn't make me happy yet ... and I doubt it ever will.

Still, it's better than working in C++ or .NET, and the pay is nice.

I would love to be doing more of my work in something more dynamic, like Python or Ruby. Heck, working on a Haskell-based project would be enough of an enjoyable shock to my brain that I would probably be entertained for years.

Unfortunately, I think that improving development environment (and, thereby, job satisfaction) like this will require switching employers. It may require getting out of medical software entirely.

A lot of people I've talked too have the feeling that it's time for The Next Big Thing - not just the people that read Beyond Java.


I started taking Java jobs because Python jobs were hard to find. 5 years later, I still feel like I'm settling for Java.