ejwaxx

Some friends of mine and I came across this film while channel-surfing in the hotel room on a trip one night in late 2001-early 2002. The sheer ridiculousness of one particular scene stuck with all of us; unfortunately, the title of the film did not.

The scene which we remember involves a hulking, large guy strangling a small girl on a dock. As he is thrashing her body about, a teardrop flies from her eye, which sets in motion an astonishing series of coincidences that results in her liberation. Some of the events in this chain, in no particular order, because I can't remember the order for the most part:

(1) Said teardrop flies from the dock, where it lands upon a spiderweb.

(2) A dog, disturbed by the movement of the spider after its web is disturbed, starts barking in an abandoned warehouse.

(3) A hobo, disturbed by the barking dog, flings a whiskey bottle at it,

(4) Which frightens off a bird roosting in the warehouse, that proceeds to fly off.

(5) Said bird defecates on the windshield of a taxicab, which obstructs the view of the cabbie.

(6) Said cabbie, unable to see, veers of the road and wrecks into a fire hydrant.

(7) Said fire hydrant floods the nearby sewer drains, evacuating lots of rats into a nearby strip club.

(8) Strippers working at said club are frightened by influx of rats, and run out into the street, still nude.

(9) Said strippers, being still nude, distract an electric company worker who, not paying attention, brings down a wrench on a transformer which kills power.

(10) Power being killed, the light in the nearby lighthouse goes out.

(11) Without light to see by, a ship cannot successfully navigate, crashing into the very dock where the young girl is being strangled, and whose teardrop set this entire chain of events into motion.

Please note that this is not verbatim correct -- it is only our recollection after a decade and after plenty of beer tonight. Any assistance in identifying this film would be most welcome.

So, I grow tired of having people ask me, "What kind of music do you listen to?" I'd like to respond to this query with a single 700 MB (+/- a few MB, of course) music CD of the best songs in my music collection.

The problem, of course, is condensing the 40 GB or so of music I have (about a 2% turnaround) into one epitomic audio CD. I came up with a solution to this conundrum, but I'm having trouble implementing it.

It seems to work well in all matters of sport, so why not implement it with my music: a tournament of music. Single, double, ..., n-ary elimination...maybe even a regular season, so to speak, of songs, with an n-ary playoff to follow the initial selections.

It's a lot of work, but I'm willing to invest the months, even years, necessary to carry out the evaluations, as I'm constantly listening to music anyway. I've tried implementing this in the past using shuffle in my audio player of choice, but it's simply too much to track manually. Not to mention the fact that every time I restart the computer and audio player, it makes a new random playlist with a new random seed, which precludes the possibility of having an unbiased evaluation of all the music.

So here's what I'm trying to implement: a program or script that traverses my music directories, randomly assigning matches between tracks. One implementation that came to my mind immediately is one that crawls the music directory, making copies of all the track names, and randomly assigning "matches" between two songs, creating a directory with (quite a lot) of subdirectories consisting of each match. I have no problem manually opening the two tracks in each directory, comparing, and moving one track to a "win" and one to a "loss" folder.

This program should be universally applicable to any directory I feed it (and thus, extensible to any n-ary tournament format, given that I am willing to manually move tracks to win or loss folders). Here are my particulars:

(1) I run Ubuntu Linux, specifically Karmic Koala x86-64.

(2) All music is located in /home/ejwaxx/Music

(3) I have a (mostly unused) 2TB HDD, so caching or copying data within reason is OK. Specifically, a given implementation can copy all of the folder it is applied to, if needed.

(4) I have a rudimentary knowledge of Linux directory structure and programming in general.

(5) I use Rhythmbox 0.12.7 as my primary audio player.

I just thought I would throw this question out there. I am not asking for a complete solution, only any pointers (and no, I don't mean memory addresses) that might aid me in accomplishing this task. If you do find this task intriguing enough to submit a substantial amount of code, I can be reached at "my Y!A username" at hotmail dot com. The language of use is entirely up to you: this is primarily a learning exercise for me, but I learn better with examples to observe. It seems like the task would be most amenable to a bash script, but I simply don't know enough about Linux directory structure, or regexes, or anything really, to get a good handle on how to start.

I am familiar with the rudiments of Python, C, C++, Java, and x86 Intel assembly (though why anyone in their right mind would attempt to do this task that way is beyond me), so these are my preferred methods; however, other methods provide a learning experience, and I am open to those as well. Implementations in other OS's are also good, especially if I can see a way to extend or port them to Linux.

Best answer *WILL* be provided to the respondent whom I deem most helpful, regardless of how close their response brings me to completion of this task.

Consider a square conducting loop of total resistance R and side length s, with a long wire carrying current I parallel to one of the sides of the loop. The wire is initially located a distance d0 from the closest side. At time t = 0, the loop is pulled away at a constant speed v, orthogonal to the wire and in the initial plane of the loop.

Determine whether or not there is a theoretical upper limit to the energy that can be generated in the loop. If such a limit exists, find it, in terms of the given R, s, d0, v and fundamental constants.

I wrote this question some time ago, and am curious as to what others would think its difficulty level is. To me, it seems like it would either be a good extra-credit problem, or good for a homework set on induction, provided sufficient time, for an intro calc-based physics class.

Consider a circular loop of radius r in the x-y plane. This loop is constructed of wire of negligible resistance, except for a resistor of resistance R. There is a uniform magnetic field of magnitude B in the positive z-direction. At time t = 0, this loop begins rotating about a diameter parallel to the x-axis at a constant angular velocity ω. Determine, in terms of one or more of r, R, N, B, ω, as well as fundamental constants, the energy dissipated in the resistor after N complete turns.

Perhaps also someone can confirm if I am even doing it right: I come up with a final answer of (πr²B)²ωπN/R.

My apartment is wired for two phone lines using six-pair cat 5 cabling (at least that's what I presume it is -- there are six pairs, and the first four pairs follow the cat 5 coloring convention). I was originally just taking a look to see which jacks are wired for which lines in preparation for running my new ATA through the IW in the apartment, but now that I see I have six pairs available, here's my question:

My roommate currently gets poor wireless reception (his room is on the other side of the rather large apartment with tons of stuff in the way) and I'd like to run an Ethernet line over there. Running it straight across the apartment will end up killing someone when we trip over it, and since it's an apartment, I'm limited in the damage I can do to the walls. So what I'm wondering is: since the inside wiring is six-pair, could I get combo Ethernet/RJ11 wallplates for my and his rooms, switch the first four pairs to Ethernet, and run a patch cable from the jacks to his comp/my router?

I know that each computer has to have a dedicated connection to the router/switch, and this is the only PC that will be hooked up in such a manner.

I currently have a computer running an Intel Core2Duo E6850 operating at 3.2GHz with at 1333MHz bus speed. It works in tandem with DDR2 RAM operating at 800MHz. I have heard tell that, since 667 is essentially 1/2 of 1333, that my RAM would ironically perform better if I purchased some DDR2 operating at 667MHz, since the bus speed is an integral multiple of the RAM speed. Is this, in fact, the case, and if so, what degree of performance enhancement could I expect from running my RAM at 667MHz?

A question I posit to the community:

It is well known that

∞

∑ 1/n

n=1

is a divergent series. In his calculus text, Stewart posits the problem (which is rather easily resolved) to show that

∞

∑ 1/n

n=1

over precisely those n that do not contain zeroes in their decimal expansions is convergent to a sum less than 90. It is relatively easy to write a program which computes an upper bound on the sum over the said domain to be no more than 11.65 (if I recall properly). The question I ask is: is it possible, using elementary mathematical methods, to refine the upper bound on the sum to less than 90, without resorting to numerical methods?