Eventually we'll get to whatever philosophical differences separate computing on the fancy new MacBook Air ultra-light machine from computing on a classy high-end ThinkPad T60 running Vista with its latest updates. (Hint: differences less profound and sweeping than many people assume.)

Also, some of the practicalities involved in shifting the center-of-gravity of your work from one platform to another. (Hint #1: If you have a lot, lot, lot of info stored in Microsoft Outlook .PST files, as I do, a full shift is not as easy as you've been told. Hint #2: At least for me, neither VMWare Fusion nor Parallels, the two programs that let Intel-powered Macs run Windows programs, has been all that simple to configure and get running -- though I remain hopeful that I'll get one of them to work!)

Let's talk today just about numbers: ones that have gotten my attention. They involve the lost-time overhead a computer imposes on you while you wait for it to work. What I've found:

Putting the computer to 'sleep' (so you can save battery power when you step away for a while). Time from issuing command till end of disk activity and screen display, three trials:
Thinkpad T60 / Vista: 12 seconds, 13 seconds, 15 seconds, average about 13 seconds. I am excluding as an anomaly the first time I ran this test, which took 80 seconds.
MacBook Air / Leopard: 3 seconds, 2 seconds, 3 seconds, average about 3 seconds.

Waking up from sleep. Time from issuing command until system is fully usable again, which means responsive to keystrokes and with wifi connection restored:
TP: 40 seconds, 75 seconds, 25 seconds, average about 47 seconds
MBA: 9 seconds, 7 seconds, 9 seconds, average about 8 seconds

Full shutdown. Time from issuing command, with no programs running but a wi-fi connection active, until end of all activity. Did this only twice because it takes longer.
TP: 75 seconds both times.
MBA: 27 seconds, 25 seconds, average 26. I am excluding as an anomaly the first time I ran this test, which took 115 seconds

Full startup. Time from pressing power button to having computer fully usable, with programs responsive to commands and a wifi connection active.
TP: 165 seconds, 170 seconds, average about 167.
MBA: 70 seconds both times.

The outlier/anomalous readings suggest that this is not a perfect science. But the pattern of these numbers matches my impression about the time-tax penalty imposed by each system -- which is indeed the biggest difference between them so far.

Previously in the MacBook Air series: #1 and #2. More to come.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.