mult.ifario.us

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.

People sometimes talk about a company having a "startup environment". This usually means less organizational structure and with that the perception of greater agility for the business, but at least to me, a relaxed organizational attitude is not the kernel of truth at the heart of a (successful) startup. A "startup environment" should indicate the presence of a single shared vision backed by clearly articulated, simple plans.

It's all too easy to start a company (or other endeavor) with "This idea is going to make me rich" as both motivation and success criterion, and it's possible to have that measuring stick rob you of all joy during the journey or lead you to make inappropriate decisions. I'd like to propose the promotional scratch-off -- the kind of gamepiece you get when you buy a fast food meal during a movie tie-in or something similar -- as a life lesson. (Life lessons for the price of a burger and coke are definitely among the most inexpensive available...)

You pay a fair price for a burger, and the gamepiece rides along as a bonus. Maybe you win fries, maybe a burger, maybe nothing... Even if you do win through persistence, eating burgers to win a prize is only going to make you fat.

I've been spending time on airplanes lately, so I'm slowly getting caught up on my weblog reading queue. I recently dequeued Paul Graham's note on hiring, and it reminds me of my advisor's metric -- whether a person has "bright eyes" (or not).

Assessing the brightness of someone's eyes has numerous advantages over degrees (which must be bought -- with tuition if nothing else -- and can be lied about), letters of recommendation (which are always guarded and almost always positive), references (which are selected to create a positive image), standardized test scores (which measure standardized capabilities), or other standard metrics; not the least of these is that it's an inclusive measurement as opposed to an exclusive measurement.

The downside of using ocular luminescence in hiring decisions is that it requires both talking and listening to candidates, so it's impossible to apply to a large applicant pool. It is a reasonable measure to apply to a small, pre-culled pool, but the probability is large that pre-processing already culled a number of bright-eyed candidates.

At least in my experience, bushy tails are irrelevant to hiring decisions...

I've been waiting for this for a while -- on 43 Folders today, a post about Act-On, a bundle for Mail.app that binds key sequences to rule actions. It's like a little slice of Quicksilver in Mail.app, so I'm one step closer to a mouse-free user experience. Even better, unlike AppleScript actions, Act-On has the property that a "move to folder" command correctly repositions the current message cursor.

While I was neither at the table nor a fly on the wall, I claim that the investment thesis for JBoss was straightforward -- Surely there's some way to make money on 30% marketshare. And while some of the analysis of IBM's purchase of GlueCode focuses on open source, I expect that the investment thesis was as simple and pragmatic as the one I imagine for the JBoss investment -- Surely there's some way to make money by using the Apache brand to drive business through the IBM services channel. (IBM already gives their software (a.k.a. "serviceware") away to drive the services channel, so it's not about free software.)