The resolution for me was to:

• Use maven jar:jar to build the jar
• Use dependency:copy-dependencies to get the jars in one place
• Clean out the multiple versions of the jars to avoid conflicts (yes…)
• Invoke on the command line using ‘java -cp lib/*: my.class.name’

This worked, and it’s easy to update just the code jar. I can’t believe it’s so painful to try and create a jar with dependencies that worked using maven assembly:assembly. Maybe I missed a better solution, but this one worked.

There are a few things I really don’t like about it (the iPad) including:

• Closed, closed, closed. You mostly have to talk to it through iTunes, you can’t just mount it and copy to it, you can’t put an app you wrote onto it, etc. It’s a peripheral, folks. It’s a fat-a** iPod. Pricey, too.
• There’s no USB port, which is a big clue.
• It’s all about one app at a time. Halting between them bugs me. My Android phone does better.

Now for the good, since I’m finally starting to find things I genuinely like about it:

• It’s a really nice reader for PDFs and e-books. I just read some Winnie-the-Pooh to my daughter at bedside and it was a pleasure. Bright, self-lit, crisp images in full color, wonderful.
• The battery life is nice and long, depending on usage of course.
• Now that I am using Dropbox+GoodReader, I have a reasonable way to get PDFs from my laptop onto it for reading, and that’s really nice. Stuffing tons of reading material in there makes sense and I love that unlike my bag it doesn’t get thicker or heavier.
• Showing pictures to friends is pretty fantastic, too. Even once they get over the ooh-ahh reaction.
• Having quick/easy access to email and web seems to fill a little niche, too, around the house.
• And I do love the touch interface, just for the simplicity of pointing at what you want. Beats keyboard+mouse for several classes of activities.

So now I guess I’m just waiting to get to the point where I’ll feel comfortable whipping it out on the subway. At which point, I will be really happy to have my mail to browse, the morning Times to look through, and some tech PDFs to read. It’s enough to make me start buying my O’Reilly’s in e-form, and finally stop piling up once-read 3-month half-life tech tracts.

Alright, I am finding some things to like about it. But it’s not a general-purpose computer and I wish that it was more of one. It does feel pretty futuristic, though.

I will just add that over the past year I have  found Groovy to be a lot of fun and surprisingly satisfying to work in. Over the years, I’ve worked in so many languages that at this point I just want something that allows me to clearly express the ideas I’m working with, and support the evolution of those ideas over time. For me, a lot of the pleasure in programming is still the process of coming to understand what it is that you are modeling. One of the things that I like best about Groovy, is that you can evolve your code alongside your understanding. So you can begin with a three line script that just does something very basic, and useful. You can then iterate those same lines into a more structured class, statically typed variables, interfaces and packages for managing name spaces and complexity, access modifiers for encapsulation, and build and deploy your code as a jar file or a war file just like any other Java component. And that component runs on the mature, high-performance Java virtual machine that we really know how to tune, monitor and scale.

Now if I could only get voice-recognition support for programming, I could use some of this recovery time to get some coding done…

1. Uninstall the old copy
2. Install the 64-bit version from the MySQL site

Install the two packages (a separate one for the StartupItem) and copy the prefs pane to /Library/PreferencePanes/.

Update

The MySQL prefs panel had an annoying dialog about having to reopen, etc. So I found this rebuilt version from the good folks at Swoon that works just fine.

But if you’ve got a Mac, and you need to get through to (and past) a Checkpoint VPN gateway, then here’s what finally worked for me, in brief:

1. Download the most recent client (which is ancient anyway) from Checkpoint.
2. If you already tried to install it, completely uninstall it.
3. If you’re running Snow Leopard, then patch the installer and follow the other instructions on this page.
4. If you use a hardware token (e.g. RSA SecurID) you’ve been having weird problems connecting to the gateway, consider trying a new hardware token (happened to me).
5. Finally, make sure that your account is enabled by the admin, since a few failures seems to trigger a disabling of the account automatically.

At that point, my life began getting much better. Back at home, I have a Linksys WRT160N router, which seemed to have all the good VPN protcols already enabled for passthrough. But no luck connecting with SecureClient. So I found this very helpful tip, and followed the advice to force UDP encapsulation. And now I finally have working VPN from the office and from home.

Whew.

I want to draw shapes again. Interact with them and the data they model.

Searching for a way in…

I looked into Java for awhile, couldn’t believe the GUI hasn’t really progressed since Swing…really? It’s insane that it’s still so ugly and awkward and heavyweight. The Groovy folks have made the best of it with Griffon, but it’s still too heavy a framework for what I want. Processing is a fantastic garden for growing beautiful visuals and interaction, but I don’t want to create sketches in the Processing language separately. I want to integrate those capabilities into a little application code. I looked at calling Processing from Java as an API, but that’s a tough path, the too-thin part of the curve. I came back around to thinking I should ditch rich clients and build on the HTML Canvas element, or maybe SVG. I set aside SVG because it just didn’t convince me to dive in. I kept expecting someone to have already built the kind of GUI framework I wanted on the Canvas API, but couldn’t find it. I guess no one really does that anymore, or they’re doing it privately. So I looked more closely at the admirable John Resig‘s amazing port of the Processing language to Javascript (processing.js), and I began to see that it was pretty concrete stuff layered on top of the Canvas 2D API.

At a certain point, you just have to dive in and write.

After more crawling and struggling to dust off memory and limber old muscles I am finally at a small base camp from which I can incrementally add functionality. I’ve got structured code (JS functions and objects), interactive logging (writing to a DIV), debugger support (Firebug), decent editor support (IDEA 9), and DOM manipulation and event handling (jQuery). My rudimentary model of objects is drawn and events are noticed.

I do feel older, though.

I am guessing it’s been about 20 years, actually, since I dealt with this stuff. I know I was writing Think C code back in ’90 and ’91 that dealt with component hierarchies and their visual representations. I even remember some earlier GUI frameworks for DOS back in the late 80′s…yikes. Talk about feeling old. Things seem a lot richer and simpler now — no dealing with offscreen buffers, or invalidating portions of the view…I guess today’s machines are fast enough that you just draw what you’ve got as often as you need.

…and the creative energy still beckons.

What’s satisfying is that the code required to make what I want seems like it can be pretty small and clear, and Javascript offers enough dynamic language sweetness to keep it appealing. All this takes place in a browser that’s already running, and whatever happens there can be integrated with the rest of a very big wide world around it now.

I was standing in a Barnes & Noble earlier today, scanning row after row of books on platforms, languages, disciplines…all software and making it. Even after all these different roads I am still excited to feel the rush of all that thinking, all those people learning and following and inventing. And I’m thrilled to be able to dive in and muck around too.

Who knows what might come out?

But this story has a happy ending, honestly. And through a miraculous combination of technologies I now have that dear old laptop running perfectly right here inside my new Mac, whenever I need it. It’s hardware can never fail again, and it will never get old and die.

Here’s what I did:

• I’ve been using Acronis TrueImage to maintain a full backup of that laptop’s drive image for a few years
• It turns out that VMWare Converter can now take Acronis images directly, and the Converter is free
• I couldn’t seem to install the Converter inside a VMWare Fusion vm on my Mac, so I recommissioned an old Linux box as a temporary XP box (installed XP, didn’t activate it)
• The images took up a few hundred gigs of space on my NAS and it’s a slow connection, so…
• I took the USB drive that backs up the NAS and plugged it right into the new XP box, installed the ext2 IFS for Windows, and there they were, ready to go
• About an hour later, I had my old laptop in a new vm, like magic.

What’s really amazing about this is that even though the laptop’s hard drive presumably was fine, I never even used it to resurrect the machine. This is the way it’s supposed to work but never seems to when you need it. The end of loss. Eternal life for your old devices.

Of course the irony is that you never really go back to use them…