OK, so /. is only putting nofollow on links when the comment has been modded down. That isn't QUITE so bad. It protects /. from comment spam. However, it does raise one other problem: bad moderators. If a comment gets modded down, and then gets modded back up because some idiot moderator chose poorly, the comment links STILL have nofollow on them.

Ah, so /. has enabled rel="nofollow" on links posted by users. How brain-dead have they become? Nofollow causes search engines not to follow the link. Have you really thought this through? This isn't so much about PageRank as it is about the ability for search engines to find things in the first place! Sure, people can follow my link back to my home page, but Google can't? Therefore Google has no idea my page is even there, unless someone else links to it without nofollow.

Once the conversion of the entire web to nofollow is complete, Google will be unable to calculate the PageRank of anything, and its results suddenly become entirely irrelevant, and look like AltaVista, which can't bring up microsoft.com if you search for "Microsoft."

Maybe the nice folks in charge of /. are trying to do bloggers a service and protect them from comment spam. STOP IT! It isn't the right way to go about it. It doesn't work! And it's going to have the opposite effect. You're also "protecting" those bloggers from being found in search engines.

Thanks a fucking lot, /.

Here's a suggestion: When moderating, set to -1, Newest First (Ignore Threads).

Why? In reading /. articles it usually turns out that the highest rated comments are near the beginning, or replies to posts near the beginning. So if someone comes along later and posts something especially insightful or funny, he has almost no chance of getting modded up. That is, unless you read in reverse chronological order. Then these people will have a better chance of being modded up (or down, as the case may be).

Comments?

If you bought one of those shiny new 802.11{abg} access points so you could be lazy and use your laptop in bed without a bunch of cords dangling all over the place, you have a decision to make. Do you want your neighbors and random strangers using your Internet connection?

If you decide you don't want other people using your connection, then don't do these things:

• Hide your SSID. Your access point will broadcast it anyway whenever your computer associates, and if you're using Windows XP then it associates every few seconds.
• Use MAC filtering. Your access point will broadcast valid MAC addresses whenever those stations are in use, and anybody can pick those up and change their MAC address to match yours.
• Use WEP. It's easy enough to crack that anybody listening can recover your WEP key in a fairly short time if you actually use your wireless connection for anything.
• Use a Microsoft access point. Microsoft access points will gladly send their WEP key to anybody who asks, making WEP completely useless.
• Use LEAP. It is based on Microsoft CHAP and a poor implementation at that. It's easy to crack.

Hm, what's the point of enabling all that security if it's so easy to get around? Here are some other things you might try:

• Turn off the access point's DHCP server. Won't do you much good, since somebody can just "borrow" your IP address when you aren't using it or use an unused IP address in your subnet.
• Reorient the access point's antenna. Then you'll just have the people on the other side of your apartment using it.

Hm, you may as well just take the damn thing back and get a refund, and suffer the Ethernet cord.

Help! I have an IBM ThinkPad R40 2681CU1 with one of those built in winmodems. Or I THINK it's a winmodem. Here's the lspci output:

0000:00:1f.6 Modem: Intel Corp. 82801DB AC'97 Modem Controller (rev 03) (prog-if 00 [Generic])
Subsystem: IBM: Unknown device 0524
Control: I/O+ Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Interrupt: pin B routed to IRQ 5
Region 0: I/O ports at 2400
Region 1: I/O ports at 2000 [size=128]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=375mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-

And again with numbers:

0000:00:1f.6 Class 0703: 8086:24c6 (rev 03)
Subsystem: 1014:0524
Control: I/O+ Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Interrupt: pin B routed to IRQ 5
Region 0: I/O ports at 2400
Region 1: I/O ports at 2000 [size=128]
Capabilities: [50] Power Management version 2
Flags: PMEClk- DSI- D1- D2- AuxCurrent=375mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
Status: D0 PME-Enable- DSel=0 DScale=0 PME-

I already tried the ltmodem-8.26 driver, and no dice. It doesn't even think there's a modem installed. Is that the right driver? Does it need a patch? What's the right driver for this modem?

How to install Oracle 10g on Gentoo Linux:

1. Create a file /etc/redhat-release as root which contains:
   Red Hat Enterprise Linux 3 ()
2. Run the Oracle Universal Installer you downloaded and extracted from ship.db.cpio.gz. It will be fooled into thinking you're on a Red Hat system and will install without complaint. Be sure to follow the directions carefully!
3. At the end of the install, it will ask you to run a root.sh script. You need to edit this script before you run it, find the two lines where it runs $LNS to create two symlinks in the /etc/rc.d/init.d directories (which you don't have on gentoo) and delete those lines. Then run the script.
4. You'll have to write your own init.d scripts to start up the database, but they aren't too hard. If you want a copy of the scripts I wrote, let me know.
5. Install Oracle Client, Cluster Ready Services, etc., the same way according to Oracle's directions.

Finally, you need a metric ton of RAM and disk to run this puppy. Count on needing at least 1GB of RAM and 5GB of disk, and that's before you make any databases.

Recently the keyboard on my three-year-old Dell Latitude CPx (P3-650, 256MB, ATI Rage Mobility 8MB, 20GB, Red Hat 9) decided to stop working. Specifically the F6, 8, I, K, comma, Page Up, and Page Down keys stopped working. I've been using it with an external keyboard for a few days.

So after exchanging a few nasty emails with Dell Support, they are actually sending someone out to me to fix it. The tech called me earlier today and he is waiting for the replacement keyboard to show up.

Amazing, all the evil things I've heard about Dell support, and they're actually coming through for me. The only mysterious part is this: I have 28 days left in my extended service contract. Don't parts usually fail right AFTER the warranty expires?

Update 6/3 7:17 pm: The technician came by, swapped out my keyboard with a brand spanking new one, did it in about 10 minutes much to my surprise, and after he let me test it to make sure I was satisfied, off he went. Now my only problem is my keyboard looks new, but the rest of my laptop's outer shell doesn't.

Will someone please tell me what the appeal is of these blogs? Why do people feel the need to share their innermost thoughts and feelings with the whole world? Who reads blogs besides maybe their close friends or family? I don't get it. Update: I'm not anti-blog per se, I just really want to know what people see in them.