Digested

There won't be a CVS-Digest this week, since I'm not going to be home. A few days off. It's been far too long since I had some time off, and I am rapidly losing interest in almost everything except getting away.

All should be back to normal next week.

This is odd. I have sent letters to the editor 4 times in my life and have been published 4 times. Hmm. Either they don't get many letters, or, well I better not think about it. I have on good authority that my head is already too large...

There was an opinion piece in the Financial Post, financial pages of the National Post, where Paul Kedrosky pondered the software monoculture question with regards to security. He made some comments that prompted me to write: (published in the July 12 2004 Financial Post)

So Microsoft 'can't go rushing out and patch the problem'? It must take 8 months for a serious problem to be fixed in their browser?

Someone said that if you can't fix your problems in two weeks, you shouldn't be on the internet. Either Microsoft's browser is so bad, so buggy so full of holes it took 8 months, or Microsoft really doesn't care. The fact that the patches came out two days after two security organisations urged people to use another browser suggests that they don't care. It is commonplace for security patches to be available within hours or days for other products.

I suggest that Microsoft only fixed things because they were losing sales. Or maybe it dawned on someone that selling known flawed software that harms customers could set you up for a nasty class action suit.

It is not the monoculture of software that is the problem. It is the monoculture of business plans, business strategy. Security doesn't increase sales. It takes the best and brightest of your developers off doing things that increase sales. Security makes your product difficult to use and setup. And a secure piece of software will hit the market a year later than everyone else. Microsoft's success is based on being first to market with neat features. Releasing a first version with promise (and bugs) that freezes the market for anyone else. To change into a purveyor of secure software, Microsoft must change everything that has made it successful. Not only successful, but an overwhelming monopoly.

Imagine where we would be if there weren't free software alternatives available.

My box is quite busy. Compiling various things, long running scripts, web server, etc. The 2.4 series kernel was stable, but the GUI experience was poor when things got busy. I upgraded to 2.6.5 (devfs RIP) and found it better, but under certain high load conditions the mouse would get jumpy with random mouse events thrown in to make things interesting. 2.6.7 is much better. Right now am recompiling KDE, apache working away in the background, and everything is smooth. Kmail is even better, not blocking as much with spamasassin. Nice.

As usual, summer has been nuts. We had very hot weather in June. If you enjoy the heat, enjoy having lazy summers, don't become a refrigeration mechanic. On top of that, my coworkers were either on holiday or home caring for newborns and wives, leaving me holding the fort. I'm not bitter. So far it's a very good year. Our volume has almost passed last year's total. And the work has been interesting, predominantly product refrigeration and freezers. That is my preference. Lettuce doesn't talk back when it is too cold, warm, draughty, or generally miserable.

An interesting observation from having my daughter travel the world. We expected to communicate through email, which presumably would be cheaper than phones. The problem is that internet access isn't universal, as phones are. We received emails sent from kiosks in airports, internet cafes, and a hostel in Belgium, but phone calls from the homes where she stayed. Is it too expensive, limited usefulness, or people are avoiding the Microsoft Experience(tm) of worms, trojans and virus'? It isn't limited to Europe, Montreal being the same.

Not much progress on the Digest coding front. I have Quanta open most of the time to some php file, but can't seem to get past the staring blankly phase. I fixed a problem where a developer used --------------- to format his commit log. Dashes are used in the cvs log as a divider. It's fixed now.

Maybe it's time to try something completely different. A project that has been rolling through my mind for a while is a residential heat load calculation program. I've got a spreadsheet that works, but it takes forever. The calculations are quite straightforward, just grabbing appropriate entries from tables, multiplying and summing. Data entry is the challenge. It takes me a couple of hours to do one, and it is very tedious exacting work. There are commercial applications available, but from what I've seen, they enforce a way of seeing the project. I can do an educated guess heat load within 10-15% in 15 minutes. The data needs to be viewed and entered in 3 or 4 different contexts. An interesting usability challenge.