I'd never been to a funeral before. My other grandfather might've even died in the States, but I was too young. Being that my father was his father's only son, and me my father's only son, it was important that I go this time, but I would've gone regardless.
They say that funerals are for the living, and this case was no different. It was good to see my father's family together (minus one sister as she couldn't make it). And it was good to see our families paying our last respects. I'd written my grandfather off a long time ago as his mind had already gone: Alzheimer's. Technically, I think he died of heart failure, but really, all his major organs were fighting for the title.
The Saturday following the Friday I arrived in Taiwan, the family went to pray for him in a ceremony that I think was to feed him in his next life. There was a lot of bowing, a lot of food offerings, and a lot of chanting by the temple priestesses (I don't know their actual title). I don't know the exact religion, but I think it's a mixture of Buddhism and Taoism.
The actual funeral was on Tuesday. We arrived in the morning and we met some family that arrived later. There was no rehearsal, but it was a fairly complex proceeding, with a man directing each of us as to what to do.
I'd never seen a dead body before, either. Of course I've seen them on TV. And real ones on the internet. Still, seeing my grandfather that day was startling. His body was emaciated from malnutrition and twisted from rigor mortis. It wasn't a pretty corpse like in Western funerals. I didn't recognize him. I'm not sure my Aunt did either as she confirmed the body with the tag. He was covered up to his neck, and the head that stuck out had its eyes closed, but its mouth opened like the mouth in "The Scream" by Munch, and the skin was waxen. After identifying the corpse, they brought it back and dressed it in some traditional clothes and laid it in an open casket. The casket was placed behind the ceremonial altar, behind a wall of cloth.
My father asked me to prepare something to say. I didn't know my grandfather too well. Particularly since the Alzheimer's took hold and he lived in Taiwan, there really wasn't much of a chance. I scraped together what I had and realized it didn't amount to much, so I wrote a poem instead. Afterwards, we each placed the papers we wrote our last words on into the casket.
After all the words and all the chanting, the body was taken to be cremated, casket, papers, and all. The furnace door went up as the family stood outside the threshold of the crematorium, the coffin slid in, and the door went back down. Apparently the process of cremation, as performed there, takes an hour and a half, so in that time, we sat around and chit-chatted.
When the cremation was done, they called us to a table and brought us the remains, among which were quite a few large pieces of bone, including skull fragments. A large pair of chopsticks was used by each of us to place a piece of bone in the urn. The bone was very light; much lighter than I thought it'd be; and very fragile as I'd soon see. After we each took our turns, the person who brought us the remains scraped up a mound of bone and dumped it on top, smashing it down with very audible crunching. The skull fragments were saved for last and placed specifically on top. Then the lid was glued down and we made our way to the temple along with the urn for my grandmother's ashes.
I think the temple was built specifically for the military deceased. A custodian showed us to the specific cabinet that we were to place the urns. A ladder was used to open a double-sized cabinet, and the custodian asked if we wanted him to put the urns in, but we said we would. With my grandfather's urn, I climbed up the ladder, slowly, but the ladder was very sturdy. My father followed with his mother's urn. The cabinet was closed and we were told that if we had anything left to say, we should do it now. Of course, the temple would always be available, but just very far away.
Friday, March 24, 2006
Wednesday, March 15, 2006
Some inappropriate photoshops
Filenames have credits, and they were most likely swiped from fark.com and somethingawful.com.
Wednesday, March 01, 2006
Something Random
My sister was complaining about her mp3 player the other day. She didn't like the randomness of her random function, as it seems to play a few songs too many times and a few songs not at all. I have the same mp3 player, and I think I notice it, too. I think the problem is a couple of things: the size of its playlist buffer and the Gaussian normalization of the distribution. It seems like if you have a list of 100 things and one is picked randomly, each thing has a 1 in 100 chance of being picked. So why does such a simple concept result in an mp3 player that can't pick songs correctly?
Of course it'd be a simple thing for a player to keep track of the last few songs played and not to play them, but it may also depend on the operating system's method of keeping track of files as to how big the list can be. Assuming it can't keep track indefinitely, at some point the buffer is going to overflow and there's a chance a song you've already heard's going to be played. It may be song you've heard before a lot, too. In fact, perhaps the size of the buffer is fixed, so you don't want it to be too big anyway because maybe some person doesn't put that many songs on their mp3 player and it'd just screw up the random function anyway? Or maybe with that in mind, the size of the buffer is set as a certain fraction of the number of songs in the overall list?
Now, if you put it on the "play randomly, but only once" mode, this may be entirely different. If I were to implement it, it'd probably work by flagging each file as it's played and picking one that's not flagged, randomly, for each new song. In short, it wouldn't really depend on a playlist buffer at all, and so, isn't limited by that as far as randomness is concerned. There's a term for this in probability class, and even an equation. But I forget.
It's been a long time since I've taken any statistics classes, but one thing I vaguely remember is convolution. Actually, 'vaguely' isn't the right word. 'Poorly' is a better one. One interesting/pertinent thing about convolution, though, is that given any type of distribution (such as our evenly distributed 1 in 100 one above), if you convolve (the verb form of convolution) it with itself enough, it becomes a Guassian distribution: normalized. What this means is that there are going to be some on one end that are played few or no times, a lot toward the middle that are played an average number of times, and some toward the other end that are played a whole lot. A playlist buffer might counteract it initially, but eventually, it becomes useless as far as desired randomness is concerned. I believe this concept is "randomness with reseeding", which forms the basis for the normalization of any distribution convolved with itself enough.
Sometime before my "hard core" statistics class, I took a physics class on optics. The professor showed us this one demonstration where photons were separated using grates, each grate having an equal chance of being the one the photon went through. The professor said that in the original experiment, one photon, passing through the grate, would strike some photon-sensitive backing opposite the grate, making a mark to show where it landed. So with that in mind, what did we, as the classs, think the resulting backing would look like over time as more and more photons hit it?
As it turns out, there was a big clump in the center, and corresponding but smaller clumps to either side of the big one, and so on, further out, in a line. The professor said it was interesting, but didn't really elaborate on why it looked that way. It wasn't until I took my statistics class that I put two and two together. Well, that is to say, I'm pretty sure that's why it looked like that. Just like I'm pretty sure that that's why mp3 player "random repeat" functions can sometimes suck.
Of course it'd be a simple thing for a player to keep track of the last few songs played and not to play them, but it may also depend on the operating system's method of keeping track of files as to how big the list can be. Assuming it can't keep track indefinitely, at some point the buffer is going to overflow and there's a chance a song you've already heard's going to be played. It may be song you've heard before a lot, too. In fact, perhaps the size of the buffer is fixed, so you don't want it to be too big anyway because maybe some person doesn't put that many songs on their mp3 player and it'd just screw up the random function anyway? Or maybe with that in mind, the size of the buffer is set as a certain fraction of the number of songs in the overall list?
Now, if you put it on the "play randomly, but only once" mode, this may be entirely different. If I were to implement it, it'd probably work by flagging each file as it's played and picking one that's not flagged, randomly, for each new song. In short, it wouldn't really depend on a playlist buffer at all, and so, isn't limited by that as far as randomness is concerned. There's a term for this in probability class, and even an equation. But I forget.
It's been a long time since I've taken any statistics classes, but one thing I vaguely remember is convolution. Actually, 'vaguely' isn't the right word. 'Poorly' is a better one. One interesting/pertinent thing about convolution, though, is that given any type of distribution (such as our evenly distributed 1 in 100 one above), if you convolve (the verb form of convolution) it with itself enough, it becomes a Guassian distribution: normalized. What this means is that there are going to be some on one end that are played few or no times, a lot toward the middle that are played an average number of times, and some toward the other end that are played a whole lot. A playlist buffer might counteract it initially, but eventually, it becomes useless as far as desired randomness is concerned. I believe this concept is "randomness with reseeding", which forms the basis for the normalization of any distribution convolved with itself enough.
Sometime before my "hard core" statistics class, I took a physics class on optics. The professor showed us this one demonstration where photons were separated using grates, each grate having an equal chance of being the one the photon went through. The professor said that in the original experiment, one photon, passing through the grate, would strike some photon-sensitive backing opposite the grate, making a mark to show where it landed. So with that in mind, what did we, as the classs, think the resulting backing would look like over time as more and more photons hit it?
As it turns out, there was a big clump in the center, and corresponding but smaller clumps to either side of the big one, and so on, further out, in a line. The professor said it was interesting, but didn't really elaborate on why it looked that way. It wasn't until I took my statistics class that I put two and two together. Well, that is to say, I'm pretty sure that's why it looked like that. Just like I'm pretty sure that that's why mp3 player "random repeat" functions can sometimes suck.
