mult.ifario.us

Howard Foster, Sebastian Uchitel, Robert Chatley, Jeff Magee, and Jeff Kramer at Imperial College London have released an Eclipse plugin that performs verification operations for BPEL4WS 1.1 processes and WS-CDL descriptions by constructing corresponding labeled transition systems (LTS). (Kramer and Magee wrote Concurrency: State Models and Java Programs about general applications of LTS techniques in Java.) It would be interesting to see how the PXE compiled representation and the FSP version of the WS-Engineer representation compare, but I have my doubts that I'll get a chance to play with it.

0xDECAFBAD — or so they say, and it certainly tastes that way... (Via /..)

Via Charles Dietzel and others, the upcoming Java Studio Enterprise contains BPMN tooling and integration with PXE! The getting started guide for BPEL functionality has some nice screen shots and a walkthrough of a classic travel reservation example process.

I've moved the venerable pbblog over to typo on lighttpd (and renamed it mult.ifario.us), and the experience has been pleasant so far. (Thanks to Joseph Tam for helpful how-to articles here and here.)

The subject matter should stay mostly the same, i.e., whatever I feel like posting. As for migrating away from WordPress, the hosting company (TextDrive) supports both PHP and Ruby, so it's not an absolute imperative. Nonetheless, I'd rather experiment with RoR than muck around with PHP. The one and only experience that I've had with PHP was modifying WordPress to correctly report the content type and deal with the combination of encoding as declared in the HTTP response and the RSS feed. It could have been a property of the version of WordPress (1.2 at the time, I think), but I felt kind of dirty afterwards, the same way that I feel after writing an AppleScript...

The typo distribution includes a convenience script for migrating a WordPress database to typo, and it works well enough. The only hitch in migrating articles from the old WordPress ID-based permalink scheme is that typo article numbers are allocated sequentially based on the number of published articles, while WordPress doles out IDs based on when the draft for the article was created or when the article was published, whichever came first. Since some drafts remain unpublished, WordPress thinks that I have 128 articles with some gaps in the IDs, while typo thinks that I imported 121 with no gaps in the IDs. It should be a quick exercise to cook up a controller to map the old IDs to new IDs. At some point, I'll get the rest of the content from the Radio and SnipSnap incarnations of pbblog imported as well.

I'm a fan of thick clients for web applications, even those with front ends that would be nice by web or even Web 2.0 standards. I'm not going to trade my OS in for a browser anytime soon.

Like TimTam for Confluence, I'm enjoying 1001 as a non-browser front-end for uploading to Flickr — drag an image or two over from iPhoto, select a couple of tags, type blurbs, hit "Upload", and it's done.

After multiple bad experiences with Mac OS X updates from Apple, I bought a 0.5TB LaCie "Big Disk" and started religiously making backups with SuperDuper!. (The exclamation point is part of the name...) Most folks don't seem to have trouble with the updates, but they've consistently been problematic for me. The Big Disk didn't make the cut for items that we initially brought to Seattle with us, so I fell off the backup wagon for a couple of months while we house hunted.

As Murphy's Law would dictate, the original drive in my PowerBook reported a failure via S.M.A.R.T. ("Self-Monitoring Analysis and Reporting Technology", not "Specific, Measurable, Attainable, Realistic, and Tangible") and refused to mount due to corruption in the structure of the filesystem. To make matters worse, fsck from single-user mode failed for the same reason. It would have been nice if the S.M.A.R.T. functionality had warned me that the failure was imminent, but I suppose that's too much to ask.

Getting my data back was problematic, and it took several boots from an external drive to get the machine to recognize the internal hard drive as being present. I tried Disk Warrior first, but it was completely worthless because it can't recover data from a drive that the OS can't mount! (Nonetheless, lots of other people seem to have great affection for the tool.) Next, I tried Data Rescue II, and that exactly did the trick of getting the files that I needed off of the old drive. The old drive was in such sorry shape that access times were quite long, so copying 20 gigabytes of data took around six days.

Getting a replacement drive installed was painless - probably 15 minutes start to finish. I bought a new Hitachi 7K60 and followed the excellent write-up at MacFixIt. I also bought a couple of tools from them, since all of my tools were sitting back in Chicago.

After several attempts at getting a good configuration installed, I have the PowerBook back running again with 10.4.3 just the way I like it. I have also promised myself to always be disciplined about backups.

This is old news at this point, but Matz endorsed Io in a roundtable (transcribed by Obie Fernandez and the good people at TopFunky):

interviewer: For those that adhere to learn one new language per year which other languages should we learn?

Matz: (paraphrased) I recommend the io language. It's an object oriented, the prototype is, and everything is a method call, even the assignments, and the declarations, and the method definition. Everything is a message. So it's kind of a simple and interesting language.

I'm already a fan of Io because of some of its langauge features (prototypes, futures) and the emphasis on messaging, but hopefully this endorsement gets the language some additional visibility and use as it approaches a 1.0 release.