Bitscape's Lounge

Powered by:

Y2K (soon to be Mega?)Fest: Night 1

Started: Monday, January 3, 2000 04:07

Finished: Monday, January 3, 2000 05:12

Bouncing and I decided to get a jump start on festing and begin last night. After the conclusion of the second X Files rerun, Bouncing brought Scully into the living room and began almost instantaniously, and I proceeded to bring Dagobah's 34 day uptime to a close (the crashing problems appear to have subsided after I disabled the external hardware cache, at Jaeger's suggestion). Dagobah was transported it into its usual festing location with minimal fuss, and I booted for the first time to the 2.2.13 kernel I compiled weeks ago. Which is good, because I was afraid a new version would come out before I would have a chance to use this one.

After a few minutes of tinkering, which included me discovering that Nist's (of all places!) clock is not y2k compliant, we made a beverage run to Albertson's in the Bouncing Mobile. (Safeway and King Sooper's are both closer, but both require those nasty "bye bye privacy" cards to get stuff (particularly soda) at a reasonable price.) To my surprise, snow was coming down hard; I hadn't looked out a window all day, so it was quite a shock to me.

Besides my usual 2 liter bottles of Pepsi (Wild Cherry variety this time) and Mountain Dew, I decided to try some Sobe Energy drink, since one of the active ingredients listed was Guarana. "It had better be good, since it's twice as expensive as normal soda." Bouncing bought some various coffee accessories to augment his carbonated caffeine mediums (he likes to get all these wacky specialty flavors, syrups, and creams; I generally couldn't care less about such things as long as it's nice and strong with some sugar and milk). We then proceeded home in the snow.

I led the opening ceremony, in celebration of the end of the world, the crashing of all PC's, the death of the Internet, the destruction of mankind, and the nuclear implosion of our solar system's star (or something like that).

Interestingly, our projects were on a similar track, and we both found ourselves perusing the Postgres documentation, and exchanged tidbits we found about how to do various tasks.

Since lag became especially bad on ssh sessions with core while Bouncing downloaded the M12 rpm, I decided to research what the kernel could do in the way of packet prioritizing, since bound to be far worse with around a half dozen people, surfing the web, downloading, and doing whatever else online when Tuesday night comes. In the process of reading one of the FAQ's, I found out how to change that nasty little behavior that causes telnets to get cut after 15 minutes of idle time; Bouncing upped Festery's timeout to 2 hours, which is what the FAQ recommended.

While I didn't quite reach the goal of figuring out how to prioritize packets, I think I got pretty close. Found the kernel option that I think needs to be checked, but the kernel config help was lacking in real info about how to make use of it. The next thing will be to figure out how to actually configure such things. Probably some user space interface tool akin to ipchains, or files buried deep in the /proc partition that will need to be edited. More docs to read...

I wiped and recreated all the tables for the next Bitscape's Lounge. I had done some things wrong before, such as getting the "default 'now'" confused with "default now()", and also wanted to make use of some knowledge gleaned from reading more docs and discussing ideas with Bouncing. This time, I also did it the smart way and put all my create queries in a big text file instead of entering them interactively at the sql prompt. Screw something up? Just edit the file and re-parse it. Retyping creates from scratch everytime, or wandering through the history to find them is just stupid. Now, I'm really satisfied with the new database.

Also, rewinding time a bit, yesterday afternoon I finally worked out precicely how the hotpoints comment system is going to be calculated. I jokingly told Bouncing I was going to take out a software patent on it, and described the general idea of how it would work. I haven't seen a system that quite accomplishes the same thing anywhere else before, although any reasonably competent programmer could divise such a scheme. He replied that sadly, the patent office probably would shew it right through. I had to agree. Were I less morally inclined, and in possession of a few thousand dollars to waste on such nonsense, I could get in on the scam game too. I'd rather do something fun and constructive, like make cool web pages. :)

Well, the next step will be to code some nice juicy perl modules to tackle this beast of a database. I'll leave that for later. Right now, my brain is just too mushy, and I need to think about exactly how I want this whole thing to be structured. No spaghetti.pl this time around. :)

So, I think I'll wander off to bed, and dream up (literally) some wild algorithms to use in the next stage of the project. Woohoo!