Initramfs

So I'm back to working on my robot. This time, I want to add a camera (Logitech Quickcam Orbit AF - works with uvcvideo, even pan/tilt! Yeah! ) to it so I can implement a Visual SLAM solution and have it drive around the dorm and various campus buildings. To do so, I had to get a Verdex gumstix computer because the connex gumstix models didn't have USB host lines and the camera we have is a USB based device (duh).

As a result, I've been playing around with some of the nicer features of the verdex model I've just recently got. One of the best improvements in my eyes (aside from the lack of USB host lines! Grrr!) is the microSD slot included with the motherboard. Now I don't have to wait 6 minutes for a 115Kbps transfer of 30MB to flash the OS over the serial line. Plus, I can make changes to the filesystem and grab log files really quickly and easily by transferring the microSD card. (I would normally just scp it over, but it's nice to be able to do it faster sometimes, and when the network here at Rice is spotty, good luck!)

Now for the topic of this post, initramfs! I just wanted to link to a pretty good article that describes the motivations behind initramfs and why it's awesome.

http://www.linuxdevices.com/articles/AT4017834659.html

If there's one thing I appreciate online the most, it has to be the sheer number of records about computer history. Without these records, there would be no way for me to even understand why things are the way they are. Computers change so fast, 12 year olds growing up now will probably only see floppy drives in museums. =P That being said, Kernel development seems to go just as quickly. This means that for anyone new to the scene, (aka moi), understanding why the kernel is built the way it is gets harder and harder.


Anyway, that article was an interesting read on initrd vs. initramfs and the whole root= fiasco. Now I'm off to set up my Gumstix!