19 Feb 06 Things I’d like to see in OpenBSD: more smp work

OpenBSD does a good job on pretty much everything they take up, but they have to proritize what they work on, because they don’t have as much manpower as other projects (i.e., FreeBSD, and especially Linux).

Anyway, SMP is one of the things that gets the short shrift sometimes because other projects tend to take priority (like bgpd, fixing bugs, and device support).

I have a dual AthlonMP box here at home, and it doesn’t quite work right. OpenBSD drops into ddb during an interrupt handler on cpu1 when the dhcp client starts up (doesn’t happen if dhcp doesn’t run). I’ve spent a tiny amount of time trying to debug it a little bit, but I just haven’t really had the time to really dive into it.

17 Feb 06 The Second Mexican War

FrontPage magazine.com :: The Second Mexican War by Lawrence Auster

This article is the best single piece that so aptly describes what the real problem with immigration is. He really does a good job of sticking to the big picture, giving background info, and more. I haven’t even finished reading it myself (it’s fairly lengthy), but I already feel like standing up and shouting “Amen!”.

This piece ought to be syndicated everywhere. Please, please, someone find a way to get it into the WSJ (Note to those who don’t know: the WSJ would sell America’s cultural soul to make a buck).

17 Feb 06 Software Development’s Evolution towards Product Design

Lost Garden: Software Development’s Evolution towards Product Design

Great article. I couldn’t say that I disagree with him on anything except for the bit about offices vs. open spaces.

I will *always* opt for an office over a cubicle or otherwise open working space. Always. If you have an office space, you can still get communal working environments by having conference rooms, a general communal area where people can congregate to work in groups, and at a minimum, you could simply leave your office door open. Other things make teamwork very doable even with an office-centric space. For example, corporate IM (think Jabber).
The opposite isn’t true; if everyone is in an open carrel or in a cubicle, that’s how they’re stuck. Their personal space is constantly invaded, their attention divided and interrupted. Sure, they can work together, but it makes it very difficult to ever hunker down and concentrate on something that doesn’t involve other people.

I think some companies merely like to push the concept of open spaces and cubicles as being better because it’s cheaper, not because they really believe it’s better.

17 Feb 06 boost; alas, I knew thee well

Hop on over to boost.org and see for yourself; it’s a collection of neat libraries that are all peer-reviewed, and adhere to the traditional (and standardized) C++ paradigm.  I like it a lot.

But jeez louise.  Did they really have to make their own build tool?  Boost.Jam is unbelievably complicated, and more importantly, all the boost libraries combined really aren’t that large.  It wouldn’t have taken — relatively speaking, of course — all that much more effort to just use autoconf and make, like *every other project*.  When bjam works, nobody cares that it sucks.  When it doesn’t work…well, there’s not much to be done about it if you don’t want to expend a ton of time on it.  Really.

Another thing – I think one thing that still prevents C++ from being more widely deployed is the lack of really great libraries in the standard.  Boost is working toward that, but many, many people do not even try to use boost just becuase of the stupid bjam tool.  I would bet money that boost could instantly pick up a bunch more users if a simple Makefile were available.

16 Feb 06 Letter from Senate Majority Leader Bill Frist

I received a letter from the NRSC the other day. It was a typical GOP fundraising letter, filled with lowbrow, ra-ra nonsense. I really hate those letters; the Republican platform (which is modeled after the conservative agenda) is reduced to a few talking points that are designed to prod you into a panic about the current state of affairs, thus feeling obligated to send money to the requesting branch of the GOP.

Did I mention I don’t like those letters? I’m a Republican. I’m active in my local party, and I follow politics pretty closely. These letters still get on my nerves.

They often have “surveys” attached, with questions like this:

Do you think President Bush should protect the nation instead of doing nothing as the Democrats want him to do?

I am particularly frustrated by this latest letter “from Bill Frist, M.D.” (a.k.a. the National Republican Senatorial Committee). I didn’t even bother reading the letter. The first question of the enclosed survey was something like this:

What issues would you like to see the Republicans in the senate tackle? Rate the following in order of importance (1-7):

• Homeland Security / National Defense
• Social Security reform
• Education
• Insurance and Healthcare Reform
• Creating Jobs, Growth, Balanced Budget, Economics, etc.
• War on Terror
• Other

There was one other thing on the list that I’ve forgotten. Anyway, immigration wasn’t on the list. Are these people insane? How much more out of touch could you be? I wrote immigration in the space provided next to “Other” and ranked it #1.

Anyway, the survey was actually not that bad, except for a few asinine questions that have obvious answers.

At the end of the survey, I lectured that if I realized the importance of keeping “our” senate majority, that it was important for me to donate. I boldly marked “NO” and wrote in the comments box below that until the NRSC ends support for Lincoln Chafee’s re-election campaign, they wouldn’t be seeing another red cent from me.

What a bunch of morons. Chalk up another item on the list of reasons to abolish the 17th amendment.

16 Feb 06 [Review] Dark Lord: Rise of Darth Vader

10 Feb 06 Animated GIFs — From MathWorld

Animated GIFs — From MathWorld

These make for some pretty nifty visualizations of highly abstract mathematical notions.

09 Feb 06 A Game Theoretic Approach to the Toilet Seat Problem

The Science Creative Quarterly » A GAME THEORETIC APPROACH TO THE TOILET SEAT PROBLEM

This is hilharious. I may come back to this to prove the conjecture referenced in Remark (5).

04 Feb 06 The calculus of celebrity

I recently sent an email to John Derbyshire, a contributor to The Corner over at National Review Online. I was pleased to see him post my message.

04 Feb 06 C++ exceptions with stack traces

Crud. I’ve spent some time trying to figure out a portable way to do this, and it’s not simple. At all. If you don’t care about portability and you use win32 or linux, then there are simple solutions because the OS maintains a list of symbols and can give them to user applications at any time.

My first thought was that I could write a macro, called “THROW” or something like that, that would help keep track of the frames in the stack trace. Not only is that not as simple as it sounds, it also only works for exceptions that are thrown and re-thrown all the way back up to main(), which doesn’t really solve the issue for me.

My second attempt involved templates, because templates are supposed to replace a lot of functionality that macros once provided. It looked something like this: template class st_exception : public std::runtime_error; int f() { throw st_exception<__file__ , __LINE__, __func__>("message"); }

Not only would this be very cumbersome to use, it was tricky to implement. Even after I got the template voodoo worked out (hint: you can’t pass a string literal as a template parameter), it still didn’t even work.

My next attempt will try to exploit information available in the C++ ABI. This way (assuming it works), I can implement exceptions with stack information and have it be portable too.