Category: Computers

— USS Clueless – Interviewing college grads says he likes to give syntax tests to job applicants, such as one where the right answer is:

do {
*destination = *source;
} while (*source != 0);

Personally, I would code this as:

while ( (*dest++ = *source++) != 0);

since this form is more compact and it’s an idiom that “C” programmers had better understand if they are maintaining other peoples’ code.

I interviewed for several jobs last year before settling on the pre-IPO startup I selected, and several of them gave me coding tests. I’ve been programming for over 20 years, and this kind of thing annoys me immensely. I think it’s reasonable to ask people system design questions, such as “how would you go about solving this problem?”, but programming tests that focus on syntactic details are just plain wrong unless you’re looking for an obedient code-monkey with no ability to think creatively or to solve problems. You have only so much time in an interview, and wasting it with this kind of thing takes away from the time you should be spending looking for more important things.

I’m working for a company that builds products conforming to the 802.11 standard. As it turns out, I was one of the people who designed the original MAC protocol spec for 802.11. Should anyone seriously have asked me about “C” syntax in an interview for this job?

One interesting experience I had at [high-profile router startup] was with a guy who was supposed to be the OS expert who wanted to talk about inhibiting scheduling in a pre-emptive OS. I explained all of the various methods for doing that, including system calls and the CLI machine instruction on Intel processors. The yahoo wanted to know what happens if two processes issue a CLI but only one issues a STI, believing that this instruction nests.

When I pointed out to him that there’s a single bit interrupt mask in the CPU, so the example was far-fetched at best, and then explained that the first STI would restore scheduling, he got mad at me and said I didn’t know what I was talking about.

I don’t think that company is going to do very well.

— I hereby add Verio to the Axis of Evil. These yahoos upgraded their perl installation to a version that uses a different database format than the old one, and they didn’t provide users access to the conversion program or the path to the old perl. So all Movable Type users on Verio have been unable to login to their own Blogs since 9:00 AM Monday.

Verio must die.

— A couple of readers take issue with my mini-rant on the design bug in Movable Type and Blogger where font-size is specified in pixels rather than in the handy font-size keywords (medium, small, and x-small) that work so well, pointing me to the Urban Legend that moved designers in this wrong direction (A List Apart: Fear of Style Sheets 4)

Who goofed? The W3C or the IE/Windows team? It doesn’t really matter. What matters is that the keywords don’t map to expected sizes, and an incompatibility exists not only between different manufacturers’ browsers, but between the Mac and Windows versions of the same browser.

…
So what can you do? Sadly, until browser makers agree on the right way to render absolute font size keywords, all you can do is ignore the W3C recommendations and use pixels in your Style Sheet. Or do not use sizes at all.

With all due respect to the folks who run A List Apart, this is bullshit. IE 6 and Netscape 6 render fonts specified by keyword exactly the same way, and just as you would expect them to be rendered – medium is a normal size font, small is smaller than medium, and x-small is smaller than small. They also render “medium” at the same size as Netscape’s old font size “3”, for whatever that’s worth. And – and this is the big deal – if your visitor wants to see the font-size defined by keyword larger or smaller, he can do so simply by changing the Browser’s View->Text Size setting; if you defined it in pixels, this adjustment does nothing.

Now it may very well be the case that IE 3 or Netscape 4 didn’t render these keywords correctly; those browsers have so many problems with CSS that it really doesn’t matter. So the fact remains that a font defined in pixels won’t display as you envision on high-res displays, and there’s not a damn thing the user can do about it, while a font defined by keyword will display correctly in a modern browser, and the user has control over it in any case. If this is some sort of unique issue for Mac people, c’est la vie.

— Macintosh user Virginia Postrel really puts her foot in her mouth in this essay on productivity in the New York Times (Lessons in Keeping Business Humming, Courtesy of Wal-Mart U.)

“Surprisingly, the primary source of the productivity gains of 1995 to 1999 was not … information technology … though companies accelerated the pace of their I.T. investments during those years,” reports a summary of the findings published in The McKinsey Quarterly. “Rather, managerial and technological innovations in only six highly competitive industries — wholesale trade, retail trade, securities, semiconductors, computer manufacturing and telecommunications — were the most important causes.”

When academic economists come out with crap like this you wonder if they’ve ever held a job in the Real World. The industries listed in the McKinsey report Postrel derives this derivative essay from are the most highly info-systems-dependent ones in the entire economy; they are able to make their “managerial and technological innovations” because they have networks of computers that feed back vital information about buying trends (“sell-through”) to the warehousing, shipping, and purchasing operations that keep the enterprise humming. This is exactly what info systems are about, when they’re not used for blogging. What a completely idiotic opposition she sets up.

— If you have a hard time reading certain blogs because the font is too small and they’re designed in such a way that the normal IE view->text size doesn’t do anything, drag this link into your favorites and click on it when you’re looking at a squinty page:

Zoom In

To turn off the zoom effect, just re-load.

linked by the 802.11b blog comparing WiFi to the cell data services it’s killing:

If wireless carriers are slow-moving bureaucratic governments of the wireless world, then Wi-Fi technology is its guerrilla movement. The technology is relatively cheap, anyone can install it and, at 11 megabits a second, it’s fast.

Whole article at: The Seattle Times: Wireless where you want: Wi-Fi is the guerrilla revolution of wireless computing

and it’s shame that so many sites with good content are going that way. I could go into a fine rant on the nature of the eyeball and the design goals of video monitors, but it mainly comes down to common sense. Sullivan has given his readers the option of choosing a color scheme, but most who use quirky design don’t. Please, folks, black on white. The eye needs contrast and definition or the brain can’t digest your thoughts.

— In the course of apologizing to Andy for crimes both real and imaginary, InstaPundit made a very passionate defense of blogspot.com. I wanted to re-read that part, but it seems that blogspot has crashed again.

Update: OK, it’s up again and here’s what he said: “Yeah, Blogger and Blogspot have their problems — but the hosting service I was planning to move to was shut down twice by hackers last month, so I’m just reminded that it’s the nature of the Internet to be less-than-fully reliable.”

Hmm. Most ISPs don’t ever crash, and certainly not on a daily basis. Really. But nevermind.

Rob Flickenger’s skepticism over CringelyNet leads him to create a realistic version of long-haul WiFi in Sonoma County:

In Sebastopol, being neighborly has gone beyond sharing a cup of herbal tea. People today share high-speed Internet access.

Using a garden-variety wireless technology popularly called WiFi, tech-savvy, community-minded Sebastopol residents are creating a free wireless Internet network dubbed NoCat.

For those not in the know, Flickenger is the sys admin for O’Reilly, the tech book publisher, and he’s been around this stuff forever.
link: Tapping in freely

Researchers crack new wireless security spec

A UNIVERSITY OF Maryland professor and his graduate student have apparently uncovered serious weaknesses in the next-generation Wi-Fi (Wireless Fidelity) security protocol known as 802.1x.

Back to the drawing board. Security wasn’t part of the 802.11 MAC protocol, because those of us who designed it didn’t know anything about the subject, and subsequent attempts to graft it on (WEP and 802.1x) have problems. As things stand, you’re not truly secure on 802.11 unless you use an end-to-end VPN, but neither are you on a wired network, easily tapped at many points. The man-in-the-middle flaw that these folks think they’ve found isn’t a genuine flaw beause they assume a directionality of signal that’s not generally possible, but I don’t want to spoil their fun. (source: 802.11b Networking News)