Stardate 20020814.1242

(Captain's log): I receive an enormous amount of reader mail these days, upwards of 300 letters a week. Many of them are long and carefully composed, and I just don't have the time to answer all the letters which deserve responses, at the length that they deserve. I have to prioritize.

Sometimes people send me letters informing me of mistakes I've made, and if they make a convincing case and if the mistake substantively weakens an argument I've made, then I usually try to respond to it online (such as here).

I also receive a fair proportion of hate mail. I'm always surprised about what I write which brings that kind of response. I'll post something I think is totally innocuous and get a pile of ranting; and then I'll post something totally inflammatory and get nada. Of course, hate mail on some subjects is pretty much a foregone conclusion. Given that I don't have the time to even answer all the letters I think are worthwhile, I rarely bother responding to hate mail at all. I don't tend to react to demands, or to righteous indignation. I have better things to do.

In reaction to a post about Apple and the Mac, I received mail on July 23 from Gene Steinberg. He uses an email address at "mac.com", runs a web site with URL http://www.macnightowl.com, and claims in his signature that he is the author of "The Mac OS X Version 10.1 Little Black Book". You can imagine how he felt about my post. He reacted in particular to this which I wrote:

Non-Mac-users would, I think, be far more inclined to believe Mac users when they point to something that actually is good about the Mac if it weren't for the fact that Mac users lie about all the things which are blatantly obviously bad about it. (For instance, how about a moratorium on "Megahertz Myth" blathering, OK? How about admitting that the Mac is slower because Motorola is being outpaced by Intel and AMD? How about ceasing to insult my intelligence with rigged demos based on carefully selected Photoshop filters run against a deliberately crippled PC running a buggy version of the program?)

Here is the letter he sent me:

You make some good points in your article. Apple needs to use faster processors and build beefier systems. It also needs to enhance its "Switch" campaign and make it more compelling to the folks who aren't persuaded by soft sell.

However when you state something as a fact without any supporting evidence, as in the statement that I quoted from you, you insult the intelligence of your readers.

Having tried Apple's tests on a production version of Photoshop 7 with a production PC equipped with a 2.2Ghz Pentium 4, I can tell you that Apple isn't lying to you in any way about these benchmarks.

In fact, the test photo and the filters they use are remarkably unsophisticated, and are not far removed from what a graphics artist might work with in normal production.

I don't know where you got the idea that a special version of Photoshop is being used or that the PC has somehow been crippled, but it's not so. Apple makes its tests and protocols available to members of the press for their review. I assure you if anything was truly fishy about the tests, the truth would be quickly exposed.

You have fallen for unfounded rumors here, and you owe your readers an apology.

He doesn't happen to be correct. The specific benchmark to which I referred was the famous on-stage one between a 733-MHz G4 and a 1.5 MHz P4, and it used Photoshop 6, not Photoshop 7.

I suppose that as such things go this wasn't really hate mail, but as soon as I found a demand for an apology, it went straight into the bozo bin. I have no use for presumption, from European heads of state or from anyone else.

Today, I received a second copy of this same letter, with an addendum:

How about the courtesy of a response?

You'll get a response, all right. First, on the issues and second on the entire philosophy of personal web sites.

On the substantive issues, it is a demonstrable fact that Photoshop 6 for the Pentium, which was the version that Jobs used for his demo, had a substantial flaw which severely slowed it. Scott Wasson at Tech Report has been all over this, and he did benchmarks of Photoshop 6 against Photoshop 5.5 here. The performance degradation is quite profound. It came out later that Adobe had set a flag incorrectly on the linker when they created the final build of the PC version of Photoshop 6.0 before sending it to manufacturing, and left a whole lot of debugging code in.

When Adobe released Photoshop 6.0.1, they didn't fuck up the link and the performance was vastly improved. Then they released a version specifically optimized to use SSE2 on the P4, though it didn't seem to have much effect.

Apple's benchmark used the crippled Photoshop 6.0, and Apple used the result of that single test in its advertising for a year, and i