dtucker: Thanks for the response. I'm not suprised I didn't find that, though. Since it is that easy, it really seems appropriate for that to be added to the warning message. I'll suggest that.

If anyone knows a good reason for the warning not to look like this, let me know:

The authenticity of host '192.168.0.123' can't be established.
RSA key fingerprint in md5 is: 59:94:5a:d7:2b:1f:ad:6e:ef:24:4c:71:1d:3c:3b:4a
If you have access to '192.168.0.123' you can run
ssh-keygen -l -f /etc/ssh/ssh_host_rsa_key.pub
on it to verify this key fingerprint.
Are you sure you want to continue connecting(yes/no)?yes
Warning: Permanently added '192.168.0.123' (RSA) to the list of known hosts.

I've wondered about SSH md5 fingerprints. When I SSH somewhere new I get this:

The authenticity of host '192.168.0.123' can't be established.
RSA key fingerprint in md5 is: 59:94:5a:d7:2b:1f:ad:6e:ef:24:4c:71:1d:3c:3b:4a
Are you sure you want to continue connecting(yes/no)?yes
Warning: Permanently added '192.168.0.123' (RSA) to the list of known hosts.

Now I understand that the idea with the md5 fingerprint is that if I know the correct md5 fingerprint on the other side I can know if I'm getting a man-in-the-middle attack. The thing is. There are cases in which I could easily verify that md5 fingerprint except I don't have the slightest idea how to do it. Does anyone know how to check it? If so, why doesn't SSH include breif insturctions in the above warning?

Even the security-minded folks I know just say “assume you arent' getting attacked on your first connection and then you get security from then on.”

I've been doing a lot of debuging at work using GDB. There are a bunch of useful commands among GDB's huge command list. Here's a quick list of the one's I've found useful:

• make – run make. It can take arguments just like on the command line.
  
• enter – do the same command again.
  
• n (next) – move to the next source line in the file.
  
• s (step) – step into the function.
  
• finish – step out of the current function.
  
• break filename:line number – set a breakpoint at a particular line.
  
• clear – clear the breakpoint you are stopped on.
  
• bt (backtrace) – show the current stack.
  
• frame n – go to the nth stack frame.
  
• info threads – list the current threads.
  
• thread n – look at the nth thread.
  
• print something – print a variable's value, or evaluate something (such as print sizeof(foo)
  
• info locals – print all local variables.
  
• info args – print all of the arguments to the current function as they are now (as opposed to as they were at the top of the function).
  
• call function – call a function.
  
• whatis something – print the type of a thing.
  
• x – just look at help x. This does memory dumps in many formats.
  

It's been quite a week. It started Saturday morning when I got up early and drove down to San Diego to have a scuba lesson diving from the Lois Ann onto the sunken Yukon as well as diving some kelp beds. It was a lot of fun. To anyone who's always thought “hey, I've always wanted to try that scuba thing,” go do it. It's fun.

After that I meat up with friends in San Diego and went to the horse races in Del Mar. I'd never been before. I won ten cents on Megahertz. (After parking, admission, and food I lost $13.90.) It's interesting to see the mass histeria that can arise when thousands of people bet a little money.

The following morning, siggraph began. The rest of the week was a blur (or was it just a convolution?) of fragment shaders, GPUs, 3D printers, clouds, and much, much more. (Including some G5s.)

Aside from the cool technology, this was the first academic conference I'd been to and I was struck by seeing the culmination of the scientiffic process in action; seeing researchers present peer-reviewed results to their peers.

Many people used Slithy, a Python-based OpenGL presentation system. The output quality appears to be much better than that of PowerPoint.
This morning started off badly due to fun with my cable modem. My connection died a while ago and today was the first day that worked to get it fixed. Their window was 8AM–noon. They showed up at 11:58, after I had called Adelphia to reschedule.

It turns out it wasn't Adelphia's fault. It looked like someone got into the junction box and messed things up. The line to my appartment went ito a 3-way joint. The “in” to that joint was from another joint... which was connected back to an out from the first joint; neither joint was connected to the outside world.

Still all I can say is that this commic is too true.

I wish Advogato supported the font tag so I could use small caps without having to use bold or itallic at the same time.

I'm having a lot of fun at SIGGRAPH. I've wanted to go to this conference ever since I first heard about it years ago. It's lots of fun, 'though I havn't had time to breathe. I got some pictures so far, 'though I don't yet have captions on them.

I'm looking forward to some down time this weekend.