Older blog entries for berend (starting at number 338)

Another day, another crash.

Just back after my linux's screen saver got stuck. compiz.real at 100%. Could ssh in, but could not kill, nor reboot the machine. I think my average uptime has been about 5 days for the last 6 months.

OS News has quite an interesting comparison between dominant processors in 1997 versus 2009.

Comunications of the ACM highlight the student enrolment crisis. But isn't it so hard to understand why the perception is negative?

What's the experience most people have with computers? They're promising machines, but often cease to work without apparent reason. No one seem to able to fix the problem, restart the program, kill it, or reboot are the only remedies. Mysterious behaviour coupled with having to press the reboot button doesn't inspire confidence that these computer programmers know what they're doing.

And it's not just my windows XP subnotebook that loses its wireless network all the time. Sometimes you can disconnect and reconnect, often you have to reboot. Or its get a blue screen. Reboot again.

It's not just Microsoft. The kids notice that FireFox keeps eating its memory and needs to be restarted now and then. A friend installs Ubuntu, and now his laptop locks up regularly, requiring a hard reboot. I lose my wireless on Ubuntu after a few sleeps and/or Microsoft VPN connects, and need to rmmod/insmod the relevant kernel modules.

Computing is just an unstable mess. And programmers must take the blame for that. But until the time programmers stop using pointer based languages and manual memory management, forget about progress in this area. It's time for the humble programmers that admit they can't be trusted with certain technologies.

I'm trying to close my Boingo account. That's a company that really makes it hard to close your account. They just love billing you every month.

I love the emacs nxml mode. It turns Emacs into one of the best XML editors I know. Especially its support for Relax NG is wonderful. But sometimes this mode just stops working with "Cannot complete in this context". I have no clue what it means, nor how to solve it. Anyone?

Bought a Navman F20 Europe in the Netherlands for just 99 Euros! I think these units went close to 300 EUROs two years ago. Probably it was so cheap because the product line is discontinued and the maps are from 2006. I'm probably going to buy a map update, which would be 50 EUR, so that means that for 150 EUR I have a fully up-to-date unit, still cheaper than the cheapest TomTom I've seen.

As there is a lot of bad information out there, which already has cost me 22 EUR, here some information for hackers.

You don't need to a buy a USB connectivity pack. The USB cable of any digital camera works fine. If you connect it through the USB cable, the unit will possibly even charge, but haven't tested that yet.

After you connect it, let Windows install the driver (yes, you need Windooze for this), else Service Pack 2 won't find the unit. Windows will install some Navman driver. You can now download, install and run the service pack update.

As I installed Service Pack 1 first, inadvertedly, it might be that this was the reason the navman driver was already in my system. Anyway, if you run Service Pack 2 it will update your system and possibly update the driver.

Perhaps if you didn't run the service pack before connecting the device, things might work a bit more smoothly, but in case you confused the order, just make sure to install the navman driver for the device, and then run the update.

There doesn't seem to be a way to unplug the device, so I just pulled the cable. After turning on he unit, the software was indeed updated.

UPDATE: yes, unit charges when connected to USB cable.

Best programming quote of 2008:

Programming is where enthusiasm meets reality.

What Makes the History of Software Hard by Michael S. Mahoney, IEEE Annals of the History of Computer, Jul/Sep 2008.

It's actually unbelievable. Europe is in melt-down mode, and the newspapers and news websites hardly mention it at all. Iceland nationalised its banks, Germany is talking about doing it shortly, England has almost done it. And the small countries like The Netherlands already nationalised one of their banks. Why o where are the blogs that can give us the real news from Europe?

As we will now discover if Friedman (and his follower Bernanke) is right, or Mises, it's interesting to look at the plan that could have saved the US, but no one was interested. McCain wasn't because he didn't know anything about the economy, and Obama wasn't interested, because this is the best chance ever for socialism. Just look at Europe where the countries are sweeping up their banks. It'll take a few weeks before the countries themselves fall over.

But back to the plan that could have been:

The solution to the trust issues in our financial system is elegant and it will work.

1. Force all off-balance sheet "assets" back onto the balance sheet, and force the valuation models and identification of individual assets out of Level 3 and into 10Qs and 10Ks. Enact this requirement beginning with the 3Q 2008 reporting period which begins next month. Total taxpayer cost: $0.00
2. Force all OTC derivatives onto a regulated exchange similar to that used by listed options in the equity markets. This permanently defuses the derivatives time bomb. Give market participants 90 days to get this done; any that are not listed in 90 days are declared void; let the participants sue each other if they can't prove capital adequacy. Total taxpayer cost: $0.00
3. Force leverage by all institutions to no more than 12:1. The SEC intentionally dropped broker/dealer leverage limits in 2004; prior to that date 12:1 was the limit. Every firm that has failed had double or more the leverage of that former 12:1 limit. Enact this with a six month time limit and require 1/6th of the excess taken down monthly. Total taxpayer cost: $0.00