I need help

It was starting to feel this way around here yesterday.

I’d finally had it with my AppleTV STILL trying to finish syncing less than 20gb from the night several months ago that my friend Noel helped me mess with channels, realized that 2 of my 3 Expresses were not 802.11n (or whatever that number is) capable, and reconfigured it as a WDS with two remotes and told me to put my oldest Express in my briefcase to use in hotels. It was better but obviously my AppleTV’s inability to sync a lousy two seasons of “Mad Men” in two months demonstrated that it still had serious problems.

So finally I did two days of seriously geeking out on wifi networking, even downgrading two base stations’ firmware to 7.3.1 and playing with KisMac for several hours to determine exactly what all my neighbors have going on with their wifi networks. (Most are running Cisco routers on the default channel 6 with WPA encryption, and only half have bothered to hide their easily-guessed ssid names. One has something like a dozen devices attached to his network, so despite inferences clearly available from this email, I am not the neighborhood’s worst geek. And FWIW, KisMac is one scary-ass powerful piece of open source software. With a few more days of KisMac monitoring I could be the evil neighbor who announces “all your base stations are belong to us” and then reconfigures everybody’s networks to be more secure and to use different channels for less interference with each other. Really, it would be a public service! And it would be more convenient for me when I need to Google something from my iPhone while walking the dog! Please explain these things to the public defender assigned to my case.)

When none of this improves anything, I finally break down and RTFM.

Apple’s “Designing AirPort networks” reveals that WDS was a bad idea, because it pretty much forces everything to use the slow 802.11b/g protocol. For 802.11n you’re supposed to use Airport Utility’s much-simpler checkboxes to configure your boxes to extend and join networks.

The doc explains dual radio mode on p48, which was 47 pages further into the details of how wifi works than I’ve ever wanted to get before, and that idea was the ticket. So the new plan is this: run my main base station (the Time Capsule in my office), the music room’s Airport Express station, our laptops, and the AppleTV on a primary network that is 5Ghz only. Run Ethernet from the Time Capsule to another Airport Express in the office, and use that as a bridge to broadcast a second, differently-named 2.4Ghz network with another Airport Express remote upstairs on the channel KisMac revealed to be least busy for benefit of my iPhone and guests with older laptops! Ten minutes and very few mistakes later, I’m done. Problems solved! AppleTV finished syncing!

Only took me two and a half months! Let’s hear it for RTFM being within the first ten things you try to troubleshoot computer problems instead of being what you do when ten weeks of Googling and futzing doesn’t produce results!

So I see this comic this morning while lying in bed and playing with my iPhone, because several friends have posted this web comic’s link to FaceBook. That’s when I realize that if I explain all these things to my friends, they will help me get help. They will know that it’s time to send me away for a while…

But make sure I figure out why Victoria’s laptop isn’t joining the new network before you do… I probably need to delete her Airport preferences, then repair permissions, restart, empty caches, zap her PRAM, upgrade to Leopard, and upgrade my base stations back to 7.3.2…

Hmmm, time to wrap this up–I have to get out of bed and get to work now…

The MacArthur Maze is back in business: Lessons for project managers

Here’s something that never happens: the California Dept of Transportation has been working faster than Google! 😉 Traffic has been flowing through the rebuilt sections of the MacArthur Maze for a couple weeks now, but Google Traffic still shows traffic from the Bay Bridge to the East Bay detouring through West Grand Ave:

Google Traffic

Speaking of amazing, the Maze was put back together in less than a month, when most people in the Bay Area including me thought it would take several months to a year. How was this possible? The story teaches several lessons for project managers.

1. Smart chartering. The Governator declared a state of emergency and chartered a rebuilding project the same day as the Maze Meltdown (as we locals have been referring to the disaster). By declaring a state of emergency, the State of California was able to waive literally years worth of red tape in procurement regulations, environmental impact reports, equal opportunity hiring requirements, design requirements, and on and on. They had bids submitted, a winner chosen, and a contract signed the very next day.

2. Smart contracting. California built significant incentives for early completion into the contract: $200K/day. The winning contract was a below-cost bid of $867K. The state estimated it would cost $5.2MM and take 50 days. By completing it over a month ahead of schedule, C.C. Myers’ firm earned the maximum bonus of $5MM—a handsome $2.5MM profit over the builder’s actual costs—not to mention immeasurable positive publicity.

2. Smart resourcing. Before submitting its bid, the winning firm had made a deal with a steel supplier that included sharing 25% of the profits if the steel could be delivered to spec, on time, for early completion.

3. Smart scheduling. C.C. Myers had workers on the job site 15 minutes before the contracts were signed, and he had crews working 12-hour shifts around the clock.

Another scheduling trick: while the massive steel girders were being manufactured in Arizona, construction work was already underway in Oakland. Normally girders and construction would be in a finish-start relationship, but they did it in parallel, literally in a finish-finish relationship: the last of the concrete was poured right after the last girder arrived.

And another: normally a full batch of girders would be completed and then all shipped together, but to save time on this project, they sent each girder up to Oakland as it was finished, with two drivers in each truck so that they could take turns and drive nonstop.

4. Smart compensation. The firm shares a percentage of profits with all its workers, so the workers all had an incentive to take on the hard work it would take to finish the project quickly.

5. What about risk management? Remember the old project manager’s triangle: time, cost, quality. Push one side, another suffers. California prioritized time by allowing higher costs, but did this affect quality? The typical problem with fast-tracking construction projects like this is that often corners get cut and safety suffers. On this point, time will tell. C.C. Myers has an unusually good track record with these kinds of emergency projects, but it’s not spotless.

So obviously there’s a big risk. According to the Project Manager’s Body of Knowledge (PMBOK), a project managers’ choices for responding to risk are to avoid, transfer, mitigate, or accept it. Let’s consider these options for the Maze:

Avoid. The risk is obviously unavoidable; even without time and cost constraints, there is no such thing as perfect, risk-free, overhead highway construction, especially in the land of earthquakes. I don’t think avoid was ever an option here.

Transfer. I haven’t seen the contract between California and C.C. Myers, but I would be surprised if it didn’t have a whole bunch of clauses spelling out C.C. Myers responsibility to repair and redo any defects shown to be its fault. This transfers the cost and responsibility for responding to quality problems from California to C.C. Myers, but is cost really the issue here? No. Lives are the issue: if this highway is unsafe, people could be injured or killed, and the real risk is borne by the citizens of California. So much for transfer.

Mitigate. Again, I imagine there were some contractual provisions to mitigate the State’s risk. I imagine they also mitigated some risk by opting to rebuild to the original design, which had stood the test of time and already survived several major earthquakes (this section of the Maze had been undamaged by Loma Prieta in 1989).

Accept. Under the circumstances, probably the only real choice was to accept the risk. Avoiding was impossible, transferring was a limited possibility, and mitigating would have broken the schedule requirements. That leaves accept.

How did they do?

I drove through the rebuilt section twice last week, and I couldn’t tell where the repairs had been made. Yesterday I drove through a neighboring stretch of the Maze and could finally recognize the section by its scaffolding and the progress of painting.

More reading for extra credit

Here are some links if you want to read more:

NYTimes article, 2 June 2007

SFChronicle’s article mid-construction about C.C. Myers

SFChronicle’s Maze Meltdown article roundup