The Mythical Bridge Engineer. Myth: All Engineers only ever work on bridges, and are selected from a small group of people inberd to have high ethics and no capacity for corner cutting. If only software developer could be like that.
Reality: Bridges are not immune to economic, political and eye-candy pressures. Selected quotes discussing just the east span of the Bay Bridge:
As with many expensive-to-solve potential problems, there was no political will to act.
This posed a serious conflict of interest, as members of the Engineering and Design Advisory Panel (EDAP) who were selecting the bridge design reviewed proposals by their own firm and rejected all proposals that did not have a representative on the EDAP.
Another qualified potential bidder did not bid due to a number of construction uncertainties owing to the innovative design … now expected to cost $6.2 billion (as of July 2005), up from a 1997 estimate of $1.1 billion (for a simple viaduct) and a March 2003 estimate of $2.6 billion that included a tower span.
While a number of proposals were submitted for a “signature” bridge, and the design chosen is considered by most critics to be acceptable from an aesthetic viewpoint, many questions have been raised by engineering experts as to its survivability under various scenarios - both natural and otherwise.
So yes to applying ethics to your job, but let’s not put bridge engineering on a pedestal. They’re not without their faults.
Apply in reverse. How to ruin a Rails project.
Better get working. Open-source project philosophy. Hopefully nothing you’re not already doing, but a good reminder nonetheless.
Valet service. As Nathan notes, the biggest problem with Scala seems to be the prominent lack of a valet service and books that read themselves. The original post concludes “I’m not at all convinced that the average software developer is able to grasp Scala.” That’s why we need to keep The Better COBOL around. Let’s not disrupt the natural order of the universe. People who know how to park their own cars already have several better languages to choose from.
Ruby 1.9. Faster than 1.8 on most things I tested against, but not without a catch: it’s not backwards compatible. On Buildr, a third of the Gems we depend on refuse to install out right and one caused havoc on Ruby Gems itself. And given that most that do install don’t even work, I can’t even run the test cases to figure out how long it will take to bring Buildr up to spec.
I’m in no particular rush, but it will be cool if we can get 1.9 usable by mid 2008. Meanwhile, James Edward Gray II has some good tips for upgrading from 1.8 to 1.9.
Picture of a Threadless T-shirt design (via ffffound!)


