Here’s a little story about a cam named Jed…

Or something like that. Anyway, yesterday afternoon I decided to take a stab at bringing a webcam online. Any webcam at all. First I went to that wonderfully-supported but broken-thanks-to-DLink DSB-C300. Still broken, in that the kernel module sees it just fine but it hoses the entire USB system within mere seconds of use.

Fine, we’ll try the Logitech cheap-o-cam. Yes, the latest CVS drivers can indeed talk V4L to it, and it still provides the yellowest, grainiest image known to camkind.

After a frustrating hour spent on each of those devices, I went online in the faint hope that support might be emerging for the Intel PC Camera Pro attached to Ryoko. What do you know, there is such a project. Too bad it’s in such miserable beta that the best I could get from the cam was a really washed out greyscale image.

And then the jackpot: A userspace app that snags raw image data from my prized 3Com Homeconnect! But wait, that’s at home…

I dragged the little black beauty into the office this morning, and after recompiling the kernel so I could get the USB /proc stuff to work I was able to snap images. Yay!

As an extra bit of fun, I hacked up a little script and an fcron job that takes snapshots every minute and saves them to a specific file on the test server I’m using.

Another couple hours of puttering around with this thing, and I may even be able to put up a webcam link again. I’m probably going to spend time trying to make the image a little bit prettier and making a proper web interface to it.

And up from the Web comes a bubblin’ crude…

Due to my inability to match the configuration of the webserver files and the machine that generates the random email address, that particular feature was broken within 24 hours of my putting it into service. Whoops.

It’s all better now, I assure you. From now on, all features get fully tested before being announced. Honest.

Some things I just can’t let go of. One of these is my sense of nostalgia, especially for drivel I’ve set to paper or electron storage at some time in the past. You wouldn’t believe some of the crap I still have cramming old three-ring binders in my closet.

In other words, the Thoughts channel is back, and repopulated with the four journal entries from the old Zero database. Read and enjoy. I mean hell, you’re reading this, so you might as well go read those too.

Instead of leaving Zero’s webserver shut down, I’ve decided instead to have her index page redirect to this journal page, just like the index page here on greyduck.net itself. Slick, huh?

There’s a method to my madness, of course. Now I can use Zero’s webserver to do various “extra” things that Invite’s servers aren’t really set up to do, like provide the randomized email address trick. (It should be working now. Feel free to test it! The link is right below the Calendar.)

The relationship between greyduck.net and zero.kgon.com is now firmly incestuous. Fun, no?

Here’s something all Linux newbies and journeymen should tape to every monitor in their office:

When switching Linux systems from one root partition to another, always check the bootloader config carefully when making any changes to the bootloader of any kind.

For instance, today I tried to completely hose my LFS rig. How? By removing all references to my stub Mandrake install from lilo.conf and running ‘lilo -v’ to update. Great, but there’s a problem. My old lilo.conf also had this problem, but I fixed it. Which is to say, I had the wrong partition marked as ‘root = ‘ in the lilo.conf. This error was in both copies of said file… and while I fixed it on one, I didn’t fix it in the other.

So there I am, staring at the kernel panic message from the fact that it was trying to use my old /usr partition as the / partition. Idiot. It’s a damned good thing you can pass the ‘root = ‘ parameter at boot time!

Have I mentioned that I love Linux? No? Well, I do. It’s even good enough to let clumsy oafs such as myself get back out of the trouble they put themselves into.

I got the index-redirect to work, though I will continue to hope that Invite lets me set an index.php instead.

The dropcap CSS trick works, but I need to tinker with the spacing and alignment parameters. It looks a bit off in Mozilla/NS6+, and it looks way off in IE6. The beauty of using CSS, though, is that once I fix it, all of the journal entries that use it will look good automagically. Go stylesheets!

Yes, I’m going to do a better “logo banner” include section. I had to do something, right? Give me a few days in which to play with Photoshop/Fireworks/Gimp and I’ll do this place up nicely.

I haven’t decided on which other Channels I’m going to create. I’m probably going to add in Work, and I think I’m also going to add a couple of Media channels so I can do reviews. (I’m going to salvage my book review from the old journal database, dammit!)

Oh, and I’ll figure out someplace to put my name. If you’re reading this, it’s probably because I told you to… but eventually I’m going to have visitors who don’t know who the hell I am. (And once they’ve left, they still won’t! Muahaha! Er. Ahem.) I’m trying to think of a non-cheesy way of doing so, which probably means I’m going to spend some time websurfing so I can rip off somebody else’s cool technique. Viva la Web, baby.

All this and more, just as soon as I get through today’s festivities: Gaming with Ben, Legend Of The Rangers in the evening. It’s a good day.