? Amazon's Split Personality: So What? |
| BW's IT 100 List ?
June 09, 2005
My "Gigahertz-per-second" Mea Culpa
Brain cramps happen, and sometimes a price must be paid. Let me explain.
I realized upon leaving work yesterday that I had inadvertently added the words "per second" to describe the speed of a PowerPC processor as 1.67 "gigahertz-per-second", in a story on Apple that we posted last night on BusinessWeek Online. Just a stupid, unexplainable mistake--one that I've never made before in my fifteen years as a technology reporter. Making matters worse, later in the same sentence we wrote the abbreviation as gHz, rather than the proper GHz. I'd hoped nobody would notice--but knew that I'd never be so lucky.
Alas, I was right. So far, a dozen or so BW Online readers have e-mailed to inform me of my error--some to politely educate me on why I was wrong (thanks to all of you!), others to flame me for being "ignorant" and "stupid", and one that suggested my error was yet more evidence of the shoddy state of journalism these days.
Now, I'm not going to cop to any connection to Jayson Blair or Newsweek's "Koran-down-the-toilet" problem, but I do hereby issue my apology. I realize such mistakes will call my credibility into question for some readers--particularly, it seems, with those who think I'm irresponsibily perpetuating the Megahertz Myth by even pointing out the gigahertz gap between PowerBooks and Wintel laptops.
So that's it. I've made my mea culpa. Now cut me some slack!
TrackBack URL for this entry:
I'd hate to have anybody accuse me of being soft on a journalist, but why not blame it on your spelling checker? Or get a better spelling checker. Or complain to or about software vendors who do a *lousy* job of integrating spell-checking features into their software.
MS Word 2002 alerts me that gHz should be GHz.
It does alert me to a problem with gigahertz-per-second, but merely that the "g" should be capitalized. Spell-checkers clearly need to do a better job with hyphen-connected terms.
And finally, it is in fact silly for us to use clock frequency as a measure of processor "speed". That's like confusing the tachometer and speedometer in a car. Ever see a speed limit sign that said "RPM" rather than "MPH"? Application-level measures of performance are much more useful.
I would also note that when talking about the performance of computers, it *would* be beneficial to quote the rotational speed of the disk drive, since they can vary and do affect performance. But, once again, application-level measures of performance are better. Remember the old Norton SI performance measurement? It had a number for the processor, and a separate number for I/O. Maybe that should be resurrected -- solely in the interests of simplying the lives of journalists, of course.
Blogger's motto: "No slack!"
-- Jack Krupansky
Posted by: Jack Krupansky at June 9, 2005 05:29 PM