I’m away from my email today, so you get a short post. But this article was too humorous (and too right) to pass up:

There is a fundamental problem with GNU/Linux—it requires clueful people to exist in the IT food chain. Anywhere in the food chain. It doesn’t take an experienced kernel hacker to install GNU/Linux, run a web server, or teach people how to log on to the network. It just requires a user with an interest in the subject, the ability to solve problems, and the desire to achieve results.

At no point is GNU/Linux experience a prerequisite for learning GNU/Linux. You can learn it the way you learnt Windows, through experimentation, Self-study, user groups, and so on. If you can apply knowledge you can solve problems. Knowledge is easy to come by, clueful people that can apply it, are not.

So, if someone is unable to roll-out a few GNU/Linux boxes in this day of Ubuntu, Debian, Novell, Red Hat, manual pages, web sites, and Internet forums it’s not due to lack of experience, or lack of knowledge. It’s due to clueless people without any ability to solve problems. Or read.

And the clueless outnumber the clueful, at all levels of the IT food chain.

Continue Reading »

Popularity: 3% [?]

Doesn’t look promising for the Democrats retaking the Senate — at least they’ve got the House. That’ll mostly slow down the mess, and will give the Democrats (who would do nothing either way) an excuse as to why they’re doing nothing. That’s promising for a Democratic win in ‘08, for better or for worse.

Lieberman got re-elected, and now he’s not pretending to be a Democrat. Not like he did a good job of pretending before, but at least he hung out with you guys.

Looks hopeful for a Democratic Congress lead by Pawlenty. I hate Hatch. Hate him hate him hate him. And splitting the executive and legislative branches is consistantly best for the country, particularly since there’s such a strong party-then-country atmosphere in both parties. (Of course, both of them say that it’s Party-then-Country because only their party can save their country. Whatever.)

I’m kinda shocked by how close the Bachman/Wetterling and Gutknecht/Walz races are. The Dems would have carried that race easy if they hadn’t run Wetterling — she was a horrible presence in debates, and left herself wide open with that “I’m not going to run ’cause the numbers say I won’t win” comment. Bachman is such a horrible evangelical-pandering whackjob that only by choosing an equally whackjobish (yet in such a uniquely liberal way) candidate could they have lost it.

And I’m proud of Minnesota — first Muslim in Congress. Take that, our Brave New After 9/11 World.

Popularity: 2% [?]

There’s something bizarrely refreshing about the self-loathing that’s always triggered in me by Nine Inch Nail’s “The Downward Spiral”. It’s catharsis for my depression. Even though I’ve moved onto new music and only rarely hear any NIN, that album has got a permanent niche in my musical identity and a key to a very unique part of my emotions.

That’s all — just felt like sharing.

Popularity: 2% [?]

Had something of a victory this week, and I’ve really got to give kudos to PostgreSQL for it. There’s something magical about taking a peice of technology, pushing it well beyond what you would rationally expect it to be able to handle, and having not just work, but work well. These sorts of experiences tend to make one a rapid fan of the technology in question. Well, postgres just did that for me.

Here’s the set up. I got the following problem dumped on me tuesday. See, there’s this table we wanted to have in the database. Now, each row isn’t that large- we needed a file date the record was created, a symbol (4-8 characters), an expiration date, and two volatility quotients (floats). So we’re only looking at 40-50 bytes per record. But here’s the problem: we’re already generating something like a million records a day, and this is expected to increase, and they’re wanting to put six months worth of data, probably more, into the database. So I’m looking at a single table that’ll start with over 100 million rows, and could easily grow to half a billion rows. Worse yet, I need to access the table by at least two different ways- by file date and by some combination of symbol and expiration date, with the latter being time critical (it’s an interactive application- a human is waiting on this data).

Now here’s the punchline- the machine this monstrosity is going on. It’s an HP DL-145 2.2GHz Dual Core Opteron with 4G of memory, a nice machine- but only a single 70G 7200RPM SATA drive as storage. Doing this on $300,000 worth of hardware I’d expect- but doing it $3,000 worth of hardware? On a machine we just had kicking around? What do they think I am, a miracle worker?

Um, yeah. Fortunately for me, I was using Postgresql. One miracle, to order, comming up. Follow me for the details of how I did the (seemingly) impossible.

Continue Reading »

Popularity: 84% [?]