Time for a core dump

I’ve been keeping a low profile lately. That’s for a lot of reasons. I’ve been doing mostly routine sysadmin work lately, which is mind-numbingly boring to write about, and possibly just a little bit less mind-numbingly boring to read about. While a numb mind might not necessarily be a bad thing, there are other reasons not to write about it.
During my college career, I felt like I had less of a private life than most of my classmates because of my weekly newspaper column. I wrote some pretty intensely personal stuff in there, and frankly, it seemed like a lot of the people I hung out with learned more about me from those columns than they did from hanging out with me. Plus, with my picture being attached, I’d get recognized when I went places. I remember many a Friday night, going to Rally’s for a hamburger and having people roll down their windows at stoplights and talk to me. That was pretty cool. But it also made me self-conscious. College towns have some seedy places, you know, and I worried sometimes about whether I’d be seen in the vicinity of some of those places and what people might think.

Looking back now, I should have wondered what they would be doing in the vicinity of those places and why it was OK for them to be nearby and not me. But that’s the difference between how I think now and how I thought when I was 20.

Plus, I know now a lot fewer people read that newspaper than its circulation and advertising departments wanted anyone to think. So I could have had a lot more fun in college and no one would have known.

I’m kidding, of course. And I’m going off on tangent after tangent here.

In the fall of 1999, I willingly gave up having a private life. The upside to that is that writing about things helps me to understand them a lot better. And sometimes I get stunningly brilliant advice. The downside? Well, not everyone knows how to handle being involved in a relationship with a writer. Things are going to come up in writing that you wish wouldn’t have. I know now that’s something you have to talk about, fairly early. Writing about past girlfriends didn’t in and of itself cost me those relationships but I can think of one case where it certainly didn’t help anything. The advice I got might have been able to save that relationship; now it’s going to improve some as-yet-to-be-determined relationship.

There’s another downside too. When you meet a girl and then she punches your name into a search engine, if you’re a guy like me who has four years’ worth of introspective revelations out on the Web, it kind of puts you at a disadvantage in the relationship. She knows a whole lot more about you than you do about her. It kind of throws off the getting-to-know-you process. I’d really rather not say how many times that’s happened in the past year. Maybe those relationships/prospective relationships were doomed anyway. I don’t have any way of knowing. One of them really hurt a lot and I really don’t want to go through it again.

So I’ve been trying to figure out for the past few weeks what to do about all this. Closing up shop isn’t an option. Writing strictly about the newest Linux trick I’ve discovered and nothing else isn’t an option. Writing blather about the same things everyone else is blathering about is a waste of time and worthless. Yes, I’ve been saying since March that much, if not all, of the SCO Unix code duplicated in Linux is probably BSD code that both of them ripped off at different points in time. And now it’s pretty much been proven that I was right. So what? How many hundreds of other people speculated the same thing? How could some of us be more right than others?

I’m going to write what I want, but I’m having a hard time deciding what I want to write. I know I have to learn how to hold something back. Dave Farquhar needs a private life again.

For a while, this may just turn into a log of Wikipedia entries I made that day. Yes, I’m back over there again, toiling in obscurity this time. For a while I was specializing in entries about 1980s home computing. For some reason when I get to thinking about that stuff I remember a lot, and I still have a pile of old books and magazines so I can check my facts. Plus a lot of those old texts are showing up online now. So now the Wikipedia has entries on things like the Coleco Adam and the Texas Instruments TI-99/4A. Hey, I find it interesting to go back and look at why these products were failures, OK? TI should have owned the market. It didn’t. Coleco should have owned the market, and they didn’t. Atari really should have owned the market and they crashed almost as hard as Worldcom. So how did a Canadian typewriter company end up owning the home computer market? And why is it that probably four people reading this know who on earth I’m talking about now, in 2003? Call me weird, but I think that’s interesting.

And baseball, well, Darrell Porter and Dick Howser didn’t have entries. They were good men who died way too young, long before they’d given everything they had to offer to this world. Roger Maris didn’t have an entry. There was more to Roger Maris than his 61 home runs.

The entries are chronicled here, if you’re interested in what I’ve been writing lately while I’ve been ignoring this place.

Much ado about nothing and other stuff

Much ado about nothing. The most recent report I read indicates that AOL/Time Warner and Red Hat are talking, but not about an acquisition. Sanity has entered the building…
Good thing User Friendly got a chance to get its two cents’ worth in. I got a couple bucks’ worth of laughter from it.
Much ado about something. On Sunday, Gentoo Linux developer Daniel Robbins announced that an obscure AMD Athlon bug slipped past Linux kernel developers, resulting in serious problems with Athlon- and Duron-based systems with AGP cards. This confirms some suspicions I’ve heard–one of the Linux mailing lists I subscribe to occasionally has rumblings about obscure and difficult-to-track-down Athlon problems.

The result was that Gentoo’s site was slashdotted into oblivion for a while, but hopefully it also resulted in some extra exposure for the distribution. Gentoo is another source-based distro. Lately I’ve been resigned to just using Debian to build my Linux boxes, but I’m still awfully fond of the idea of compiling your own stuff. As CPUs get faster and faster, I expect that to become more commonplace.

But I digress. The bug involves the CPU’s paging function. Older x86 CPUs used 4K pages. Starting with the Pentium, CPUs started allowing 4MB pages. But a bug in the Athlon’s implementation of this extended paging causes memory corruption when used in conjunction with an AGP cards.
Alan Cox is working on a workaround. I’m a bit surprised a patch isn’t already out there.

CPU bugs are discovered all the time, but it’s fairly rare for them to be serious. If you ever run across a Pentium-60 or Pentium-66 system, boot up Linux on it sometime and run the command dmesg. You’ll find workarounds for at least two serious bugs. A TI engineer named Robert Collins gained a fair bit of notoriety in the last decade by researching, collecting, and investigating CPU bugs. Part of it was probably due to his irreverant attitude towards Intel. (As you can see from this Wayback machine entry.) Sadly, I can’t find the story on the site anymore, since he was bought out by Dr. Dobb’s.
Catching up. I haven’t been making my rounds lately. The reason why is fairly obvious. I used my day off yesterday to have lunch with someone from my small group, then when I got home I read the e-mail I absolutely had to read, responded to those that absolutely had to get responses, answered a couple of voice messages, wrote and sent out a couple of other messages, looked up, and it was 5 p.m.

“Alright God,” I muttered. “I just gave the day to Your people. Time to go spend some time with You.” So I whipped out my handy-dandy Today’s Light Bible and read about Moses. Seemed appropriate. The inadequacy and jumping the gun and making excuses, that is. The Biblical “superheroes” were human just like us, and the book doesn’t gloss over that. Today’s Light is designed to divide the Bible into pieces so you can read the whole thing in two years. I can’t decide if I want to get through it in a year or in six months. A few years ago I read it in its entirety in four months, but that pace is a bit much. If you’re willing to spend as much time reading the Bible every day as the average person does watching TV, you can make it through in a few months. But it’s not exactly light reading, and I’m not sure I recommend that pace. If you’re willing to dedicate that kind of time to Bible study you’re probably better served by learning Greek so you can read the New Testament in the original. Then if you’ve still got your sanity you can think about tackling Hebrew.

I finally got around to reading Charlie Sebold’s entries for the last few days. One especially poignant observation: “I continue to be surprised at how much I remember about computers, and how much I forget about everything else (including far more important things).”

I sure can relate. I wish I could trade everything I remember about IBM PS/2s and Microchannel for something more useful. But I remember goofy baseball statistics too–I can recite the starting lineup and pitching rotation of the 1980 Kansas City Royals (I’ll spare you). But I can’t tell you the names of all seven people I met Sunday night.