Raible Designs

AppFuse Roadmap updated for 2.0

I've updated the AppFuse Roadmap for the upcoming 2.0 version. The current plan is to have a few milestone releases before the final 2.0 release. I can't promise any specific release date because of developer's schedules and current commitments. However, my hope is we can have it done by the end of this year.

Milestone 1

• Maven 2 integration
• Migrate from CVS to Subversion
• Remove XDoclet, putting generated files into source tree
• Get all flavors (WARs) deployed and running on Tomcat and Jetty Maven Plugin
• Spring 2.0: Upgrade XML to XSDs, annotations where appropriate
• Upgrade WebWork to Struts 2
• Documentation - move to new Confluence or DocBook-based system for 2.0 documentation

Milestone 2

• JDK 5, Annotations and JSP 2.0
• Archetypes for creating new AppFuse projects
• Documentation, including screencasts for tutorials

Milestone 3

• Integrate XFire by default (for existing classes and generated classes)
• Code Generation/AppGen - hopefully with the AppFuse Maven Plugin
• Documentation

2.0 Nice to haves

• TestNG replacing JUnit
• Hibernate Validation Annotations (hard part is hooking it into the web frameworks)
• Tomcat Plugin for Maven (that operates like Jetty's plugin)
• JPA Support
• Support for JDK 1.4 using something like Retroweaver
• Support for using Ant to build/test/deploy AppFuse-based projects
• Someone to do all the documentation ;-)

Let me know if there's anything that seems to be missing from this list.

If you can't wait until the end of this year, you may take a look at Patrick Lightbody's newly announced Project Able. He has a lot of good ideas in there that I hope to learn from.

A device-free weekend

This weekend, I did something I haven't done in a while (if ever). I left my laptop at work, and I promised myself to not touch a keyboard all weekend. When my phone's battery died on Saturday morning, and I realized I'd left my charger at work, I decided to go no-devices all the way. Of course, I still watched TV and enjoyed my car stereo, but being computer-free felt liberating. It was an awesome weekend, and I'm thinking about doing it from now on.

Saturday, Julie and I had a great time watching Superman, IMAX, in 3D. It was killer. The whole movie wasn't 3D, but the 4 scenes in 3D were incredible. I highly recommend it. On Sunday, we had kickass seats to The International thanks to Virtuas. We sat on the 18th hole and watched golf while getting free cocktails served to us for several hours. It didn't suck by any means.

Tutorial: How to Create an AJAX-based file upload progress bar

Helge Tesgaard has written a nice tutorial on How to Create an AJAX-based file upload progress bar with AppFuse 1.9.3. Check it out if you have a need for such a thing. I especially like Helge's comment on productivity towards the end.

How accurate are java.net's top project statistics?

Yesterday, java.net updated their Top Ranked Projects page to reflect the top projects for July 2006. I've always thought these were fairly accurate. However, a couple of weeks ago, I did a bit of investigating (since GlassFish always seems to top all categories) and I found some interesting information. First of all, when last month's stats were first published, I noticed that DWR was on top of the "mailing list traffic" category. I did some investigating, and found the following:

DWR: 62 + 522 + 1 + 3 + 1 = 589
AppFuse: 65 + 134 + 542 + 21 = 762

So I thought, am I missing something? Why does AppFuse rank lower. I posted a question to the java.net forums, and got the following response:

I have a typo in my spreadsheet where I ran the numbers. Let me recheck and fix.

Thanks to Helen for her honesty, but isn't this something that could easily be automated? This would eliminate any calculation errors? Especially since there seems to be issues with the previous month's statistics.

For June 2006 (excluding AppFuse's issues and cvs lists):

GlassFish: 184 + 204 + 28 + 80 + 2 + 36 + 4 + 11 + 17 = 566
AppFuse: 684 + 19 = 703

AppFuse had more mailing list traffic, yet GlassFish is #1 and AppFuse is #3. Is java.net trying to make GlassFish look more popular than it is?

Denver JUG Tonight: Echo and Wicket

Looks to be a good show at tonight's DJUG: Justin Lee will be talking about Wicket and Tom Poindexter is presenting on Echo2.

BTW, did you know that Wicket may become an Apache project? I consider the main web frameworks in Java to be: JSF, Spring MVC, Struts 2 and Tapestry. RIFE, Stripes and Wicket follow close behind, but are generally very unheard of in the Java community at large. It's kinda wierd to see most of the Java web frameworks (save Spring MVC) end up at Apache, isn't it?

Update: This was a good meeting, mostly because both speakers were excellent. Relaxed, comfortable and good humored. Justin has a recap (and downloadable presentation) on his blog.

7 simple reasons to use AppFuse

IBM developerWorks published my "Seven simple reasons to use AppFuse" article today. Here's a summary:

Getting started with open source tools for the Java™ platform such as Spring, Hibernate, or MySQL can be difficult. Throw in Ant or Maven, a little Ajax with DWR, and a Web framework -- say, JSF -- and you're up to your eyeballs just trying to configure your application. AppFuse removes the pain of integrating open source projects. It also makes testing a first-class citizen, allows you to generate your entire UI from database tables, and supports Web services with XFire. Furthermore, AppFuse's community is healthy and happy -- and one of the few places where users of different Web frameworks actually get along.

While you're there, you might be interested in reading the "Introduction to Spring 2 and JPA" tutorial. I don't know if we'll get JPA support into AppFuse 2.0, but it's certainly a possibility.

As far as AppFuse 2.0, here's the current structure I've started on for Maven 2:

appfuse
    - data
        - hibernate
        - ibatis
    - service
        - pom.xml
        - src
    - web
        - jsf
        - spring
        - struts
        - tapestry

After code is moved into the directory structure above (or completely re-written), I'd like to move to working on creating single module archetypes and multi-project archetypes (data, service, web) with Maven 2.

After getting the Maven 2 structure checked in, hopefully we can start looking at replacing AppGen. Scott Ryan has done a fair amount of work on this so far with his AppFuse Maven Plugin.

I plan on documenting the plan of attack and milestone features for 2.0 sometime this week.

Update: I started working on the Maven 2 conversion last night. The above structure has changed slightly. Now there's a project in data and web (notice the pom.xml and src in these directories). These projects will contain the classes/files that are common to their sub-projects. The fact that these projects even exist will likely be transparent to the end user.

Dream Machine

Well done Apple, you've created my dream machine. Too bad it's so damn expensive. My ideal configuration is as follows:

• Two 3GHz Dual-Core Intel Xeon
• 4GB (4 x 1GB)
• Hard Drive - Bay 1: 250GB 7200-rpm Serial ATA 3Gb/s
• Hard Drive - Bay 2: 500GB 7200-rpm Serial ATA 3Gb/s
• 2 x NVIDIA GeForce 7300 GT 256MB
• Apple Cinema HD Display (23" flat panel)
• 2 x SuperDrives
• Both Bluetooth 2.0+EDR and AirPort Extreme
• Apple Wireless Keyboard and Apple wireless Mighty Mouse
• AppleCare
• Subtotal: $6,436.00

The nice thing about owning one of these bad boys is you could run several copies of Linux and Windows (using VMWare and/or Parallels). I could literally buy one machine and have 5 in a matter of hours. Hmmm, I should probably bump up that ram to 8GB.

It's tempting, but that's a lot of cash to shell out for a machine. Then again, the more write-offs the better.

Back from the Hills

Last week at the Cabin was a great experience. It was one of the first times I've been there in a while where I didn't want to leave. The kids loved it, partly because of the 4-wheeler, but also because they were outside the whole time. Allowing kids to get as dirty as they want is kindof fun. Abbie liked it so much she said she wanted to move there. Who knows, maybe she will someday.

When we arrived at the cabin on Sunday, there was lots of tall grass everywhere.

My mom, a.k.a. "The Burn Boss" (that's her new title at work), made quick work of the fire danger and spent the next 3 days mowing down the grass and killing any knapweed she found.

When we weren't hanging out, riding "Rockit" or building the new cabin, we enjoyed Glacier Park and putting a foundation under my parent's bedroom.

On the way home, we stopped and saw my aunt in Helena, then spent the night at the Old Faithful Snow Lodge. Saturday was a long 11-hour drive through Wyoming.

All in all, it was a great road trip. I think I've had enough driving for a while though. In the first week alone, I hit 7 states in 7 days!

Going off the grid

Today marks the beginning of Raible Road Trip #10, Leg #2. My parents and I, as well as Abbie and Jack, are leaving soon to drive to Spokane, Washington. We'll stop in Spokane tonight, where we'll meet up with my sister and Julie. Kalin arrives later tonight (hopefully with a case of cider from her cidery), Julie flies in tomorrow morning. From Spokane, it's about a 4-hour drive to the cabin.

Think we're the ones having all the fun? Nope, Julie is in Vegas right now with her sister. They flew in last night for 12 hours of good ol' Vegas fun.

The cabin, which you can read more about on my about page, is "off the grid" in Montana. There's no electricity, no running water, and 120 acres of lush Montana acreage. There's a sauna for bathing (good ol' Finnish roots) and a generator if we really need power. My mom prefers no generator, my dad likes power and bright lights. I suppose I could dial-up to the 'net if I wanted, but MacBook Pros don't have modems, which is how I prefer it. Vacation shouldn't involve a computer - so I'm pretty pumped that mine will be pretty much useless all week. Besides, we're pouring the foundation for the Cabin #2, and that sounds like a lot more fun than any of this computer stuff.

I'll be back in Denver a week from today, but I don't expect to have any desire to work/blog/etc. until Monday the 7th. And with that, I'm officially on vacation.

Moving from CVS to SVN at java.net

I like java.net. Ever since I moved AppFuse to java.net from SourceForge (in February 2004), I've been very happy with their service. I can't recall having an issue with CVS, and if I did, it was during a scheduled outage. The only thing that's ever made me consider moving AppFuse to another site was java.net's lack of Subversion.

A lot has changed in the last year, and now java.net does support Subversion. However, if you want a project converted from CVS to SVN, you have three options:

• Drop and re-create your project, losing all of your mailing-list/forum history
• Create an entirely new project, which gives you similar results to the first option
• Pay to have your repository converted

Since I like being at java.net (uptime being the major factor here), I've decided to go with option #3. It's costing $750 to convert AppFuse from CVS to SVN, and Virtuas has agreed to sponsor the move. I like that we'll be using SVN soon, but it does seem odd that we actually have to pay for it. AFAIK, moving a SourceForge project from CVS to SVN doesn't cost anything.

In recent days, I've thought about moving AppFuse to Google Code, but it looks like they might be having some early growing pains.