Shallow Thoughts : : 04

Snowy Grapevine, Rainy Central Valley, Unappetizing Sagebrush

The Grapevine, the pass through which Interstate 5 crosses the mountains north of LA, was covered in snow today. Gorman, near the highest point of the pass, was blanketed in white, not even bushes or grass poking through.

We'd hesitated before coming this way -- the Caltrans web site had listed the pass as closed until a scant half hour before we left. Signs on the highway at Castaic still said the pass was closed, but we put our trust in the web, and forged on. Happily, the road was open, clean of snow, and barely even wet, giving a lovely view of the snowy Transverse Ranges as we passed through this unexpected white christmas. Also fun was seeing a double semi trailer full of oranges passing through this wintry landscape.

Descending into the central valley, we saw the first "Food grows where water flows" sign at Buttonwillow, pinned to a trailer in a field of sagebrush and tumbleweed. Perhaps a goat would have found some food there. At least sage (which I do like in cooking) is closer to culinary than the cotton that all the farms here were growing for the last two years (presumably due to subsidies) the remnants of which still litter most of the empty fields along the I-5 corridor.

"Farm water feeds the nation", fifty miles farther north, also stood in a field of tumbleweed, but the California Aqueduct was nearby, so it was at least somewhat topical. The next "Food grows where water flows" adjoined a vinyard. Does wine count as food? Maybe they were table grapes.

The Buttonwillow rest stop features lovely woven hanging birds' nests, visible now when the trees are bare of leaves and looking like something out of an African weaverbird documentary. I didn't get a good look at the birds occupying those trees now; usually those I-5 rest stops are populated mostly by blackbirds and ravens, but I'll have to keep a sharp eye out next time I pass through in spring.

Tags: travel, sign, politics, water
[ 19:33 Jan 04, 2005    More misc | permalink to this entry | ]

"Works With Linux!"

Why is it that devices which claims Linux support almost never work with Linux?

When my mom signed up for broadband, we needed an ethernet card and router/firewall for her machine. The router/firewall was no problem (a nice Linksys with a 4-port switch included) but ethernet cards are trickier. First, it turns out that lots of stores no longer sell them, because they're trying to push wireless on everybody. ("Hey, I have a great idea! Let's take Windows users who don't even know how to run Windows Update, and set them up with an 802.11b network that opens their connection to the whole neighborhood, plus anyone driving by, unless they take extra security precautions!") Second, ethernet cards are in that class of hardware that manufacturers tend to change every month or so, without changing the model number or adding any identifying information to the box so you know it's not the one that worked last time.

The sale card at Fry's was an AirLink 101, and it claimed Linux support right on the box. The obvious choice, right? We knew better, but we tried it anyway.

Turns out that the driver on the floppy included in the box is for a RealTek 8139 chip: a file called 8139too.c, which has already been incorporated into the Linux kernel. Sounds great, no? Except that it turns out that the card in the box is actually an 8039, not an 8139, according to lspci, and it doesn't work with 8139too.c. Nor does it work with the ne2k driver, which supports the RealTek 8029 chip. No driver we could find could make head nor tail out of the AirLink chip.

Amusingly, the Windows driver on the floppy didn't work either: it, too, was for a RealTek 8139 and hadn't been updated to match the chip that was actually being shipped on the card. So the AirLink is a complete bust, and will be returned.

Fortunately, the other likely option at Fry's, a Linksys LNE100TX, is still the same chip (DEC Tulip) that they've used in the past, and it works just fine with Linux.

It's sad how often a claim of Linux support on the box translates to "This is a crappy product which probably won't work right with any operating system, since we change it every couple of months. But three revs back someone tried it on a linux machine and it worked, so we printed up all our packaging to say so even though we didn't bother to retest it after we completely redesigned the board."

Tags: linux
[ 10:57 Jan 04, 2005    More linux | permalink to this entry | ]