Three Monkeys, Three Typewriters, Two Days

March 31, 2004

Linker fun

I just tried updating binutils to version 2.14.90 because I noticed that the changelog mentioned better (faster) handling of ELF symbols by ld. Link times for debug gklayout dropped by a factor of 3 (from 127 seconds to 40 seconds).

Posted by bzbarsky at 11:11 AM

March 17, 2004

Firefox, again (modulo name)

Today I tried Firefox again (for the first time since the 0.7 days). It took me a minute or so to locate the download link on the firefox page. I guess I automatically ignore all sidebars on websites nowadays, since 99% of the time they just contain garbage.

Much progress has occurred. DOM inspector actually works. The toolbar customization dialog no longer thinks it knows better than I where it should be located. I ran into my first polish bug in about 45 seconds (since checking the functioning of the JS console was one of the first things I did), but other than that minor snag the JS console and View Source window worked just fine.

The preferences dialog was, and still is, quite excellent. Much more friendly than SeaMonkey's, with not much less in the way of features. The inability to add types in the helper app section is a bit bothersome, but I understand the reasoning, and may get used to it (though so far it has led to a few more bugs getting filed, as I was testing adding a type). The dropdown for choosing where to save files by default is a nice touch, though I can't figure out where it thinks "My Downloads" is...

No more time to play right now, but when I get back I should give this thing a nice long spin. It looks like I could get through a day of browsing with it now (what with DOM inspector finally existing), so I should be able to really try it out.

First step at that point will be to find a less anemic theme. Something more like Modern.

Posted by bzbarsky at 4:41 PM

March 16, 2004

Standards support claims

I find it interesting that the people actually implementing support for a standard don't spend much time talking about how great their implementation is. I attribute this to two primary causes:

  1. They are very much aware of a wide array of nasty problems in their implementation or architecture.
  2. They're a little too busy fixing those problems to have time for grandstanding.

So it seems that the solution to the problem of obnoxious advocates is to make them learn something about the things they're advocating....

Posted by bzbarsky at 12:45 AM

March 2, 2004

Electronic voting

Joe writes about his electronic voting experience and echos what a lot of people have been saying lately—that a paper trail would make him much more confident in the voting procedure.

Perhaps a way to do that is to stick with the punch ballots but have the voting machine itself punch the ballot for you based on what options you selected (in addition to electronically recording the vote). The idea is that recounts could be done either by hand or by punch-card-counting machines (and there should be no hanging chad issues, if the card was automatically punched and the punching machine is designed at all well).

In fact, with this system you could do both counts—the electronic count and the punch card count; preferably using machines from different (and unaffiliated) manufacturers. If there's a discrepancy, hand counting comes in (though chances are this will just introduce even more of a discrepancy).

Posted by bzbarsky at 7:11 PM

March 1, 2004

Attinasi was right

This needs to stop.

Posted by bzbarsky at 9:35 PM


  1. No more reading the "bugs filed today" list. Hopefully others are doing it and making sure that thins like reporters posting patches don't get lost in the shuffle.
  2. No more reviews of random UI changes.
  3. Reviews for anything that takes longer than five minutes happen 10-1 on Saturdays.

Here's hoping that this is sufficient.

Posted by bzbarsky at 1:59 PM