| 00:23.00 | *** join/#utah On0bi (n=on0bi@75-169-252-2.slkc.qwest.net) |
| 00:54.09 | elg | -metar kelp |
| 00:57.43 | Supaplex | ~rlart |
| 00:57.44 | ibot | nukes fungus with a single large nuke |
| 00:58.43 | Supaplex | haha |
| 00:58.44 | fortyseventeen | giggles. |
| 01:00.12 | Sargun_Screen | is that 417? |
| 01:00.39 | fortyseventeen | 47teen, forty17, 40-17, whatever. |
| 01:00.46 | Sargun_Screen | Springfield, Missouri? |
| 01:01.00 | Sargun_Screen | what does it mean |
| 01:01.21 | fortyseventeen | es un numero desaparecido. |
| 01:01.28 | Sargun_Screen | english |
| 01:01.33 | fortyseventeen | Missing No. |
| 01:01.56 | Sargun_Screen | Interesting. |
| 01:02.25 | Sargun_Screen | I wonder if finhas invited his fiance in ere |
| 01:03.42 | *** join/#utah levi` (n=user@levi.dsl.xmission.com) |
| 01:04.05 | *** join/#utah emcnabb (n=emcnabb@cpe-071-070-229-150.nc.res.rr.com) |
| 01:04.06 | *** mode/#utah [+v emcnabb] by ChanServ |
| 01:19.55 | *** join/#utah redbeard2 (n=jphall@techhat.dsl.xmission.com) |
| 01:42.22 | *** join/#utah elg (n=fugalh@c-76-113-97-1.hsd1.nm.comcast.net) |
| 01:42.22 | *** mode/#utah [+v elg] by ChanServ |
| 01:47.21 | *** join/#utah elg (n=fugalh@c-76-113-97-1.hsd1.nm.comcast.net) |
| 01:47.21 | *** mode/#utah [+v elg] by ChanServ |
| 02:30.47 | *** join/#utah ccjoe (n=joeg@75-169-37-165.slkc.qwest.net) |
| 02:32.43 | *** join/#utah On0bi_ (n=on0bi@75-169-252-2.slkc.qwest.net) |
| 02:36.52 | On0bi_ | somebody google New World Order |
| 02:38.30 | *** join/#utah tiwula (n=lane@208.64.90.18.static.utahbroadband.com) |
| 02:39.21 | *** join/#utah ccjoe (n=joeg@75-169-37-165.slkc.qwest.net) |
| 02:40.50 | romanovic | is looking to rent a 3-bedroom residence, if anyone knows a place |
| 03:16.28 | eightyeight | maquis: actually, i do. trigger.pl. :) |
| 03:20.21 | eightyeight | interesting it starts off with '/' though... |
| 03:22.02 | eightyeight | let's try this again: |
| 03:22.13 | eightyeight | ~say http://icanhascheezburger |
| 03:22.14 | ibot | http://icanhascheezburger |
| 03:22.14 | eightyeight | http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/a359/zoom/ |
| 03:22.16 | brac | [ThinkGeek :: God Kills a LOLcat :: Zoom!] |
| 03:22.32 | eightyeight | heh. better |
| 03:31.15 | Supaplex | ouuu, I can has tEh lolcat t-shirt? |
| 03:41.45 | *** join/#utah fugue88 (n=dsowen@fugue88.dsl.xmission.com) |
| 04:01.24 | levi | Any math wizards awake? |
| 04:01.52 | fortyseventeen | what kind of math? |
| 04:03.05 | levi | Well, I need to figure out some sort of transformation of f(x) = e^x - e such that a computer can calculate it without loss of precision. |
| 04:03.33 | *** join/#utah _carmony (i=JustinCa@c-67-164-249-224.hsd1.ut.comcast.net) |
| 04:04.57 | fortyseventeen | hrm. exp(x)-exp(1) is too far off? |
| 04:06.44 | levi | Subtracting two floating point numbers that are close to one another results in loss of precision. |
| 04:08.11 | fortyseventeen | yeah, floats are silly. it would really take something to fix that, though, as far as I know |
| 04:08.40 | levi | Well, the trick is to use algebra to get rid of the subtraction. |
| 04:14.40 | fortyseventeen | all I can think of is exp(1)*(exp(x-1)-1)...perhaps better? |
| 04:16.27 | fugue88 | exp(f(x)) = exp(exp(x) - exp(1)) = exp(exp(x)) * exp(-exp(1)) |
| 04:17.06 | fugue88 | ln(...) of that last expression gives, I think, f(x) without doing a subtraction. |
| 04:17.22 | fugue88 | Icky. |
| 04:18.10 | fortyseventeen | and still lossy |
| 04:18.18 | fortyseventeen | for x close to 1, anyway |
| 04:18.29 | fugue88 | No doubt, but maybe in a more acceptable way? |
| 04:18.47 | fugue88 | Nah, guess not. |
| 04:19.52 | fortyseventeen | just doesn't like floats. |
| 04:20.28 | fugue88 | Okay, I'm going to backtrack on my backtrack and claim that it will be less lossy than the subtraction. |
| 04:20.57 | fugue88 | s/claim/hypothesize/ |
| 04:21.11 | fortyseventeen | it seems either way it could only be off by 1 or 2 bits... |
| 04:21.13 | fugue88 | There, amply guarded... |
| 04:21.28 | fugue88 | Hmm... |
| 04:25.13 | fugue88 | Okay, I'll agree with fortyseventeen here. If e^x is so close to e so as to differ only in the least signficant bit, then the normal subtraction will reveal exactly that least significant bit. |
| 04:25.46 | fugue88 | You can't git more precise than that because that's the limit of the data that you're starting your calculation with. |
| 04:28.14 | levi | Floating point subtraction is kind of a weird thing thanks to the way it's represented. |
| 04:29.09 | levi | Let's say you're doing 37.593621 - 37.584216 |
| 04:30.25 | levi | The distance between the two (1 - y/x) is 0.0002501754. To figure out the number of bits of precision you lose, you have to figure the powers of 2 that the distance lies between. |
| 04:30.58 | levi | In this case, it's between 2^-12 and 2^-11, which means between 11 and 12 bits of precision are lost. |
| 04:31.22 | *** join/#utah fugue88 (n=dsowen@fugue88.dsl.xmission.com) |
| 04:33.07 | fortyseventeen | doesn't that mean that 11 bits are correct? |
| 04:35.12 | levi | Doesn't what mean that? |
| 04:36.43 | fortyseventeen | eh, I dunno. either way it seems like way more error than it should have |
| 04:37.18 | fugue88 | If only digital computers had real reals... |
| 04:47.39 | *** join/#utah Sargun (n=Sargun@atarack/Staff/Sargun) |
| 04:53.31 | fortyseventeen | OK, this is useful http://en.wikipedia.org/wiki/Unit_in_the_Last_Place |
| 04:55.31 | *** join/#utah fozzmoo_ (n=fozz@209.41.95.5) |
| 04:55.37 | *** join/#utah On0bi (n=on0bi@75-169-252-2.slkc.qwest.net) |
| 04:57.26 | fortyseventeen | so converting floats to decimal to binary will have an error of less than half of the ULP, meaning that subtracting or adding them will have an error no larger than the ULP. the difference can't be off by more than 1 bit. |
| 04:59.10 | fortyseventeen | s/the ULP/1 ULP/ |
| 04:59.26 | fortyseventeen | forgot /g, oops |
| 05:00.22 | levi | The gap between two floating point numbers is not a constant value. |
| 05:01.03 | fortyseventeen | no, but you can bound it |
| 05:01.42 | levi | I just told you the way to calculate the bounds on loss of precision for subtraction of floating point numbers. |
| 05:04.19 | fortyseventeen | I don't see how (1 - y/x) means anything |
| 05:05.34 | levi | That's the distance between two numbers. |
| 05:05.39 | Supaplex | m00 |
| 05:07.04 | fortyseventeen | right, so you should be comparing the expected and actual results that way, not the operands |
| 05:12.06 | levi | Okay, let's look at this another way. |
| 05:12.24 | levi | Let's say we've got f(x) = x - sin(x) |
| 05:13.46 | levi | And we're calculating it at x = 1/15, which gives us something like 0.6666666667 |
| 05:14.13 | levi | So we take sin(x), which gives us 0.6661729492 or something close to that. |
| 05:14.50 | levi | So, then we subtract the two, and we get 0.0004937175 |
| 05:15.17 | levi | You can see right there that we lost some precision. When it's normalized again, it goes to 0.4937175000 |
| 05:15.37 | levi | But the full-precision answer is 0.4937174327 |
| 05:16.42 | fugue88 | Lost significant digits, but not precision. |
| 05:17.20 | levi | When you lose significant digits, you lose precision. That's what significant digits are about. |
| 05:17.34 | fugue88 | (0.6666666667 +/- 0.00000000005) - (0.6661729492 +/- 0.00000000005) = 0.0004937175 +/- x |
| 05:17.45 | fugue88 | where x >= 0.00000000005 |
| 05:18.11 | fugue88 | So, okay, lost some precision. |
| 05:18.18 | fortyseventeen | so the error is 0.000000136312, which is less than 2^-22. for normal floats, that means 1 bit off. |
| 05:19.33 | fugue88 | Here's another: (2.000 +/- 0.0005) - (1.000 +/- 0.0005) = 1.000 +/- x |
| 05:19.57 | fugue88 | Exact same loss of precision (just pretend), but different effect on the sig figs. |
| 05:20.29 | fugue88 | It really is the precision that matters; the sig figs are just an easy proxy for the precision. |
| 05:21.38 | levi | Yes, but the point is that subtracting two numbers that are very close to one another can lead to a significant and avoidable loss of precision. |
| 05:22.01 | fortyseventeen | yeah. it's just usually no more than a bit or two |
| 05:22.18 | levi | That depends. |
| 05:22.36 | levi | Again, I told you the method for calculating the lost bits of precision. |
| 05:22.57 | levi | I'm sorry if you don't believe me, but I'm not going to copy the proof out of my textbook here. :) |
| 05:24.06 | fugue88 | It's not "Numerical Recipes" is it? |
| 05:25.12 | levi | No. |
| 05:25.35 | levi | "Numerical Mathematics and Computing" by Cheney and Kincaid. |
| 05:25.40 | fugue88 | Ah, good. |
| 05:25.47 | fugue88 | Good that's it's not Recipes. |
| 05:26.02 | fugue88 | Okay, I'm about to give up on typing. |
| 05:27.04 | fugue88 | I think I've heard good things about Cheney and Kincaid's book. |
| 05:27.23 | fugue88 | Not the one I used though. |
| 05:27.43 | levi | I'm sure the information is good, but it's not a very helpful book, and my teacher is not very helpful either. |
| 05:27.53 | fortyseventeen | maybe what you're saying is: adding/subtracting can only be off by 1 bit, but the value of that bit can vary with the exponent? |
| 05:28.14 | fortyseventeen | that's what ULP is, the value of that bibit |
| 05:28.17 | fortyseventeen | *bit |
| 05:28.55 | fortyseventeen | is that what you meant by 'not constant'? |
| 05:28.56 | levi | fortyseventeen: What I'm saying has nothing to do with ULPs, it's to do with spurious zeros at the end of the mantissa after shifting to normalize the exponent. |
| 05:29.46 | fortyseventeen | Ah, gotcha. yeah, you're not losing precision. |
| 05:29.55 | levi | So, no one in class knew what to do with this problem, so we asked the teacher (who assigned it) and he floundered on the board for 15 minutes or so until he gave up and said he just wanted us to flounder to, basically. |
| 05:30.27 | levi | fortyseventeen: You must not have paid attention to my last example, because there was precision lost there. |
| 05:30.52 | levi | s/flounder to/flounder too/ |
| 05:30.53 | fortyseventeen | like fugue said, the number of meaningful bits (sigfigs) is smaller, but the actual degree of precision didn't change |
| 05:31.57 | fortyseventeen | those extra zeros are still smaller than the last bit of either operand |
| 05:32.17 | levi | If you don't believe 0.4937175000 is less precise than 0.4937174327, I'm not sure what to tell you. |
| 05:32.32 | fortyseventeen | I'm not arguing that |
| 05:35.02 | fugue88 | 0.4937175000 +/- 0.00000000005 is more precise than 0.4937174327 +/- 0.00005 though. |
| 05:35.42 | fugue88 | Unless you know where a number came from, you *can't* declare its precision based on its representation. |
| 05:36.05 | levi | Geez. |
| 05:36.38 | fugue88 | Okay, I shouldn't give such a hard time. |
| 05:36.48 | fortyseventeen | what you're saying is that normalizing a number by increasing its exponent and shifting the bits up affects its precision, right? |
| 05:36.48 | levi | I'm aware of all this. |
| 05:38.44 | levi | fortyseventeen: I'm saying that subtraction of machine floating point numbers that are very close together causes significant loss of precision that can be avoided by manipulating the equation such that the subtraction is no longer required. |
| 05:40.06 | levi | fortyseventeen: The bits of precision were lost *before* the shift, the normalization just moves the zeros to the end. |
| 05:42.48 | fortyseventeen | then I can't say anything else other than that you're confusing precision with bits |
| 05:42.53 | levi | It's not so much the subtraction operation itself that causes the loss of precision, but the fact that you're setting up your equation using the subtraction of close-together numbers. |
| 05:43.27 | levi | I am NOT confusing precision with bits! Yeesh! |
| 05:43.37 | levi | You seem to be thinking I'm saying something other than I'm saying. |
| 05:43.55 | levi | I clearly demonstrated the phenomenon I'm trying to avoid, and you said you weren't arguing against that. |
| 05:46.10 | fortyseventeen | oh, wait, OK. so in your first example (37.593621 - 37.584216) you would instead do (0.093621 - 0.084216), right? |
| 05:46.37 | Supaplex | converts a vmware player image to a xen guest . . . and waits on qemu-img convert ... |
| 05:47.03 | levi | No, no, no. |
| 05:47.23 | Sargun | Haha |
| 05:47.37 | Sargun | It'd be quite funny if levi was a college professor |
| 05:48.22 | Sargun | levi, how would you modify the IVT under linux (0x00-0xff*4)? |
| 05:48.40 | levi | That was just an example of how to calculate the number of bits of precision you lose through that subtraction. To fix it, you would eliminate the subtraction altogether. |
| 05:49.23 | Sargun | levi, isn't subtraction just addition (for ints) |
| 05:49.39 | levi | Sargun: I'm not talking about ints. |
| 05:49.45 | Sargun | oh, doubles? |
| 05:49.49 | levi | floats |
| 05:50.25 | fortyseventeen | same thing. it just changes the sign and adds |
| 05:50.48 | Supaplex | where's the rootbeer? |
| 05:50.57 | fortyseventeen | but I think I see what you're saying |
| 05:50.58 | Sargun | at work, we took a bunch of code that uses floats, and converted it to use ints, and inexact rounding (use a scale of 0-360 (float), and scale to (-2**31-2**31) |
| 05:51.06 | Sargun | and it got significantly faster |
| 05:51.23 | Sargun | levi, anyways, my question about interrupts |
| 05:52.01 | fortyseventeen | no idea how it could apply to e^x - e though |
| 05:52.27 | Sargun | fortyseventeen, IIRC there is an x86 instruction for that |
| 05:52.29 | levi | fortyseventeen: In order to get the same precision in an algorithm with subtraction as without, you would have to have several more significant bits in your operands. |
| 05:53.19 | levi | If you use an algorithm that doesn't subtract number close to each other, you can retain all the precision you started with. |
| 05:53.26 | levi | Or more of it, anyway. |
| 05:53.42 | eightyeight | anyone familiar with openfire here? |
| 05:53.52 | levi | Sargun: I don't know how you modify the IVT under linux. |
| 05:55.10 | levi | fortyseventeen: Well, my first approach was to use a series version of e^x, and then subtract the series for e^1 from it. But I still ended up with an x - 1 in every iteration. |
| 05:55.27 | fortyseventeen | yeah, I thought of that too. :P |
| 05:55.32 | levi | And I tried factoring an e out. |
| 05:55.47 | levi | Like I said, the teacher floundered around trying a bunch of things and then gave up. |
| 05:57.54 | Sargun | fortyseventeen, why do you want to calculate e^x? |
| 05:58.25 | levi | Sargun: Dude, you are so behind in this conversation, don't even bother trying to catch up. |
| 05:58.31 | Sargun | ok, I give up |
| 05:58.32 | fortyseventeen | it's levi, and he wants (e^x - e) for values of x close to 1 |
| 05:59.00 | Sargun | 0 |
| 05:59.08 | Sargun | just assume x is 1, and call it a day. |
| 05:59.37 | Sargun | if there is a little loss of precision, call it a pentium |
| 06:20.05 | fortyseventeen | f(x) = (1-e) + x + x^2/2! + x^3/3! ... |
| 06:21.33 | fortyseventeen | how does exp() |
| 06:21.45 | fortyseventeen | how does exp() work anyway? |
| 06:22.19 | *** join/#utah On0bi (n=on0bi@75-169-252-2.slkc.qwest.net) |
| 06:23.22 | fortyseventeen | oh jeez. check out expm1() |
| 06:24.29 | fortyseventeen | thats...not quite what you want, but close |
| 06:25.32 | levi | I thought of that, but then I disregarded it, but now I think it may be the right solution. |
| 06:26.05 | fortyseventeen | expm1(x)-expm1(1) |
| 06:26.24 | levi | What's expm1? |
| 06:26.31 | fortyseventeen | e^x-1 |
| 06:27.11 | fortyseventeen | and of course (e^x - 1) - (e^1 - 1) = e^x - e |
| 06:29.57 | fortyseventeen | uh, that does very little good, actually |
| 06:30.08 | levi | No, I think the series the way to go. |
| 06:30.50 | fortyseventeen | have it it, then |
| 06:30.56 | fortyseventeen | s/it/at/ |
| 06:31.01 | fortyseventeen | needs sleep. |
| 06:31.33 | levi | G'night. |
| 06:52.59 | tensai | man, what's up with all the scary math? glad I missed that. |
| 06:56.02 | levi | tensai: My lame numerical methods class. |
| 06:56.10 | levi | I wish I could miss it. |
| 06:56.30 | tensai | college level math is enough to make anyone cry |
| 06:56.52 | high_roller | didn't cause me any problems |
| 06:56.58 | high_roller | <-- math major :) |
| 06:57.07 | tensai | *shudder* |
| 06:57.57 | tensai | I got along just fine with math in high school. college beat that out of me. |
| 06:58.06 | levi | Now I'm trying to figure out some way to get (1/x)(sinh x - tanh x) to a form that preserves accuracy at small values of x using identities instead of series. |
| 06:58.12 | tensai | same with a lot of things I used to like. chemistry, physics |
| 06:58.51 | high_roller | man I don't remember all my trig identities |
| 06:59.04 | levi | This might have been easier if it hadn't been like 7 or 8 years since I last did math like this. :) |
| 08:21.01 | *** join/#utah mrpull (n=Jake@67.208.136.68) |
| 09:27.19 | *** join/#utah levi (n=user@levi.dsl.xmission.com) |
| 10:05.14 | *** join/#utah mrpull1 (n=Jake@67.208.136.68) |
| 11:59.38 | *** join/#utah eightyeight (n=aaron@ubuntu/member/atoponce) |
| 12:38.34 | *** join/#utah brac (n=rbot@c-76-113-97-1.hsd1.nm.comcast.net) |
| 12:42.21 | *** join/#utah romanovic (n=romanovi@c-67-186-209-120.hsd1.ut.comcast.net) |
| 13:04.38 | *** join/#utah emcnabb (n=emcnabb@cpe-071-070-229-150.nc.res.rr.com) |
| 13:04.38 | *** mode/#utah [+v emcnabb] by ChanServ |
| 13:27.27 | *** join/#utah jnbek (n=jnbek@unaffiliated/jnbek) |
| 13:41.35 | *** join/#utah elg (n=fugalh@c-76-113-97-1.hsd1.nm.comcast.net) |
| 13:41.35 | *** mode/#utah [+v elg] by ChanServ |
| 13:43.56 | *** join/#utah eggyknap (n=jtolley@jtolley.dsl.xmission.com) |
| 14:36.20 | *** join/#utah eggyknap_ (n=jtolley@jtolley.dsl.xmission.com) |
| 14:47.47 | *** join/#utah ^Migs^ (n=Migs@wsddevdude.weber.k12.ut.us) |
| 14:48.46 | *** join/#utah emcnabb (n=emcnabb@nat/redhat/x-664f2e72b8c6183b) |
| 14:48.46 | *** mode/#utah [+v emcnabb] by ChanServ |
| 14:59.07 | *** join/#utah thaddeusq (n=thaddeus@216.49.181.128) |
| 15:21.16 | *** join/#utah elg (n=fugalh@dhcp07.cs.nmsu.edu) |
| 15:21.16 | *** mode/#utah [+v elg] by ChanServ |
| 16:00.51 | *** join/#utah Jayce^ (n=jayce@office2.bluehost.com) |
| 16:00.51 | *** mode/#utah [+o Jayce^] by ChanServ |
| 16:08.08 | *** join/#utah pashdown (i=pashdown@sine.xmission.com) |
| 16:16.49 | *** join/#utah fungus (n=olsonl@bromine.sosstaffing.com) |
| 16:19.09 | *** join/#utah JoshH (i=josh@josh.xmission.com) |
| 16:26.15 | *** join/#utah fungus (n=olsonl@bromine.sosstaffing.com) |
| 16:26.55 | *** join/#utah Tene_ (n=tene@poipu/supporter/slacker/tene) |
| 16:27.27 | *** join/#utah Sargun (n=Sargun@atarack/Staff/Sargun) |
| 16:38.58 | *** join/#utah neybar (n=jalance@204.228.142.225) |
| 16:39.21 | *** join/#utah redbeard2 (n=jphall@208.110.152.42) |
| 16:42.42 | *** join/#utah toofer (n=toofer@kjordan.securesites.net) |
| 16:59.21 | *** join/#utah tiwula (n=lane@75.148.99.197) |
| 17:00.52 | *** join/#utah markl__ (n=mark@c-67-186-245-229.hsd1.ut.comcast.net) |
| 17:46.24 | eightyeight | maquis: http://techdirt.com/articles/20090128/1928243562.shtml |
| 17:46.25 | brac | [Reading Comprehension: A Terrible Thing To Waste | Techdirt] |
| 17:46.31 | eightyeight | if you're still at amazon, that is. :) |
| 17:49.45 | maquis | hmm? |
| 17:55.50 | maquis | ugh... |
| 17:57.06 | *** join/#utah eggyknap (n=jtolley@c-67-171-126-25.hsd1.ut.comcast.net) |
| 18:01.37 | eightyeight | i take it that's a sore spot? |
| 18:04.38 | *** join/#utah mabes (n=mabes@bmabey.fttp.xmission.com) |
| 18:05.39 | *** join/#utah TimRiker (n=timr@70-1-81-157.pools.spcsdns.net) |
| 18:05.39 | *** mode/#utah [+o TimRiker] by ChanServ |
| 18:23.55 | *** join/#utah sjansen (n=sjansen@97-117-62-121.slkc.qwest.net) |
| 18:27.02 | *** join/#utah undertakingyou (n=will@undertakingyou.dsl.xmission.com) |
| 18:28.07 | *** topic/#utah by sjansen -> http://plug.org/irc | Channel log @ http://ibot.rikers.org/%23utah/ | I can has #utah |
| 18:29.51 | *** join/#utah On0bi (n=on0bi@97-117-4-58.slkc.qwest.net) |
| 18:46.16 | ^Migs^ | this is hilarious: http://bbs.311.com/forums/showthread.php?t=14326 |
| 18:46.20 | fozzmoo | More git questions... |
| 18:46.58 | fozzmoo | So, I'm working with a team of "developers" in India. They're familiar with SVN and CVS, but git is foreign to them. |
| 18:47.22 | fozzmoo | They're concerned about two developers editing the same file and are proposing a process by which a developer may "lock" a file for editing. |
| 18:47.36 | fozzmoo | I'm thinking that's completely unnecessary if they're using git. |
| 18:48.15 | fozzmoo | Can someone with more experience than me validate my thoughts? |
| 18:49.32 | Jayce^ | unecessary |
| 18:49.40 | sjansen | Locking is a substitute for good communication and proper review. They may be right. If they're not able to talk to each other, and not careful enough to review merged code, then git may not be for them. |
| 18:51.34 | fozzmoo | thx |
| 19:02.03 | ^Migs^ | hahahahahaha http://www.foxnews.com/story/0,2933,484326,00.html |
| 19:02.10 | ^Migs^ | ZOMBIES AHEAD |
| 19:02.35 | elg | fozzmoo: agreed. locking in distributed revision control is a "you really don't want that" issue |
| 19:02.53 | elg | because you really don't want the mindset that requires locking. it's not compatible |
| 19:02.59 | fozzmoo | Right. |
| 19:03.03 | fozzmoo | That's what i was thinking. |
| 19:03.31 | tedp | they should [learn to] be able to resolve a merge conflict |
| 19:03.59 | tedp | as mentioned, good communication will reduce merge conflicts |
| 19:06.14 | elg | really i haven't seen a need for locking in cvs/svn either, when reasonable communication is in place |
| 19:06.27 | elg | but merges can be more hairy there, not because of the model but just because those tools suck at merging |
| 19:06.54 | elg | merging is first class in distributed. has to be |
| 19:08.32 | whiteinge | sjansen: you can sign up for a public google group without a goog account |
| 19:09.05 | elg | btw fozzmoo i just found this blog the other day and it's pretty nifty |
| 19:09.15 | sjansen | whiteinge: Did you find a URL with instructions? |
| 19:09.17 | elg | gitready.com |
| 19:09.18 | fozzmoo | elg: Do share. :) |
| 19:09.21 | fozzmoo | elg: Thanks! |
| 19:09.32 | elg | kind of like a daily git tip in your feed reader |
| 19:10.07 | whiteinge | sjansen: I put instructions up on our group page: http://groups.google.com/group/oalug |
| 19:10.38 | whiteinge | sjansen: basically you email GROUPNAME-subscribe@googlegroups.com |
| 19:11.27 | sjansen | cool |
| 19:49.27 | *** join/#utah tristanbob (n=tristanr@ubuntu/member/tristanbob) |
| 20:05.26 | ^Migs^ | http://notalwaysright.com/ah-parents/337 |
| 20:07.49 | elg | what grecian symbol is this? Î |
| 20:08.45 | jsmith | elg: Xi? |
| 20:09.18 | jsmith | http://en.wikipedia.org/wiki/Xi_(letter) |
| 20:10.53 | elg | my hero |
| 20:12.34 | sjansen | http://icanhascheezburger.files.wordpress.com/2009/01/funny-pictures-this-baby-panda-bids-you-hello.jpg |
| 20:12.36 | eightyeight | http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/a359/zoom/ |
| 20:12.37 | brac | [ThinkGeek :: God Kills a LOLcat :: Zoom!] |
| 20:13.52 | sjansen | http://icanhascheezburger.files.wordpress.com/2009/01/funny-pictures-your-cat-is-ready-to-admit-he-gained-weight.jpg |
| 20:13.53 | eightyeight | http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/a359/zoom/ |
| 20:13.54 | brac | [ThinkGeek :: God Kills a LOLcat :: Zoom!] |
| 20:14.25 | sjansen | http://icanhascheezburger.files.wordpress.com/2009/01/funny-pictures-horse-hates-drinking-from-lakes.jpg |
| 20:14.26 | eightyeight | http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/a359/zoom/ |
| 20:14.27 | brac | [ThinkGeek :: God Kills a LOLcat :: Zoom!] |
| 20:14.55 | sjansen | icanhascheezburger is the coolest |
| 20:15.17 | sjansen | Everyone should visit http://icanshascheezburger.com/ |
| 20:16.50 | sjansen | They've got a brilliant business model. They don't create content, readers do. And they don't host content. http://icanhascheezburger.files.wordpress.com/ does the hosting. They just profit. |
| 20:17.41 | sjansen | http://icanhascheezburger.files.wordpress.com/ so... I guess that means wordpress.com does the hosting. |
| 20:17.41 | eightyeight | http://www.thinkgeek.com/tshirts-apparel/unisex/popculture/a359/zoom/ |
| 20:17.42 | brac | [ThinkGeek :: God Kills a LOLcat :: Zoom!] |
| 20:18.31 | sjansen | http://tinyurl.com/dk5mws |
| 20:18.47 | sjansen | Didn't think of that one, did you eightyeight! |
| 20:20.58 | eightyeight | heh |
| 20:24.43 | Jayce^ | sjansen, same with motivated photos, but MP one-ups by having the crack-like 4 "other" photos as thumbnail links on the side |
| 20:35.13 | tensai | Jayce^: I had to quite MP cold turkey |
| 20:35.22 | Jayce^ | very very busy |
| 20:36.19 | Sargun_Screen | http://blog.wired.com/27bstroke6/2009/01/fannie.html < I could have made a better bomb of logic :-P |
| 20:38.43 | ^Migs^ | http://snurl.com/aymzr |
| 20:39.01 | sjansen | MP is too obviously crack. After losing several hours of my life to it, I realized I had to walk away and never come back. |
| 20:39.41 | mrpull | ^Migs^: is that first hand? |
| 20:39.42 | sjansen | icanhascheezburger.com on the other hand... It's such a mild addiction "I can quit whenever I want, I just don't want to". |
| 20:39.46 | *** part/#utah mabes (n=mabes@bmabey.fttp.xmission.com) |
| 20:39.48 | ^Migs^ | 'fraid not |
| 20:40.21 | mrpull | still good stuff |
| 20:51.16 | eightyeight | is playing "Antonio Vivaldi - Presto" on last.fm |
| 20:52.16 | elg | good stuff |
| 20:53.30 | eightyeight | listening to liebesleid by rachmaninov now. changed stations |
| 20:54.00 | elg | <3 the emiliana torrini station |
| 20:54.05 | eightyeight | i tend to prefer classical piano. hopefully this station will stay on piano |
| 20:54.08 | *** join/#utah gshipley (n=gshipley@nat/redhat/x-7a0d22d441608bef) |
| 20:54.25 | eightyeight | but vivaldi as always a treat |
| 20:59.05 | maquis | thesuperbigfrog: ping |
| 20:59.08 | eightyeight | i was listening to piano sonata no. 23 in f minor op. 57 by beethoven last night with my wife and realized how much i missed classical piano |
| 20:59.25 | *** join/#utah gshipley_ (n=gshipley@nat/redhat/x-2445ee8fc027b9f5) |
| 21:00.01 | eightyeight | caused me to pull out lots of beethoven, chopin, rachmaninoff, liszt and others |
| 21:01.52 | *** join/#utah cgibreak (n=cgibreak@cgibreak.us.to) |
| 21:06.17 | *** join/#utah Sargun (n=Sargun@atarack/Staff/Sargun) |
| 21:06.26 | sjansen | questions eightyeight's manhood |
| 21:06.53 | fozzmoo | hopes eightyeight doesn't feel the need to prove his manhood |
| 21:07.06 | eightyeight | just question me. i'll answer you straight. no need to ask my manhood |
| 21:21.13 | *** join/#utah cgibreak (n=cgibreak@cgibreak.us.to) |
| 21:22.33 | *** join/#utah Heartsbane (n=hearts@unaffiliated/heartsbane) |
| 21:24.19 | ^Migs^ | man, I want this: http://www.engadget.com/2009/01/28/cigarette-lighter-camera-shows-theres-nothing-more-patriotic-th/ |
| 21:24.21 | brac | [Cigarette lighter camera shows there's nothing more patriotic than espionage - Engadget] |
| 21:40.35 | *** join/#utah romanovic (i=romanovi@wireless34.wireless.utah.edu) |
| 21:40.36 | *** join/#utah romanovic_ (n=romanovi@c-67-186-209-120.hsd1.ut.comcast.net) |
| 21:45.42 | fozzmoo | elg: Thanks for recommending I check out audacious. |
| 21:45.50 | elg | np |
| 21:47.24 | eightyeight | gah |
| 21:47.30 | eightyeight | is playing "Carl Orff - Olim lacus colueram" on last.fm |
| 21:47.45 | eightyeight | and the lead female vocalist is nothing compared to sylvia mcnair |
| 21:47.47 | elg | oooh, I love carmina burana |
| 21:47.52 | eightyeight | grabs his own copy |
| 21:49.59 | romanovic | yeah carmina burana is cool |
| 21:50.47 | eightyeight | the best version is directed by leonard slatkin with the saint louis symphony orchestra as chorus |
| 21:51.00 | eightyeight | sylvia mcnair, john aler and hakan hagegard |
| 21:51.13 | eightyeight | s/as/and/ |
| 21:52.41 | romanovic | i never knew what it was until the Lord of the Rings movies came out, then someone played it for me claiming in influenced the movie soundtracks |
| 21:54.13 | ^Migs^ | meh, everyone's trying to stick gothic chants into movies now |
| 21:54.15 | ^Migs^ | it's nothing new |
| 21:54.29 | ^Migs^ | same with games |
| 21:55.03 | romanovic | yeah. same with sean hannity, he stole carmina burana so now i can't listen to it without thinking of his opening theme music |
| 21:55.35 | eightyeight | michael jackson introduced it to me. :) |
| 21:57.35 | levi | All I know of Carmina Burana is O Fortuna. |
| 21:58.08 | levi | But I really like O Fortuna. :) |
| 21:58.46 | eightyeight | yeah. that's the recording everyone is familiar with |
| 21:59.01 | eightyeight | it's fun to get the lyrics, and translate the latin to english |
| 21:59.23 | eightyeight | nothing but a bunch of drank, greedy, love-starved monks. :) |
| 21:59.34 | levi | Yeah, the text is pretty interesting. :) |
| 21:59.48 | levi | It's not all Latin, either. |
| 22:00.42 | levi | "Chramer, Gip Die Varwe Mir" is a Middle High German song by a female prostitute. |
| 22:00.43 | eightyeight | well, mostly. i guess there's german and french in there, but not much |
| 22:01.14 | eightyeight | yeah |
| 22:01.19 | levi | "Merchant, give me the color to redden my cheeks so that I may make young men love me whether they wish it or not. Look at me young men! Let me please you!" |
| 22:01.31 | eightyeight | :) |
| 22:02.32 | levi | Then there's one about a swan who's been killed and roasted. |
| 22:03.57 | eightyeight | is that in taberna? |
| 22:04.39 | eightyeight | yeah. |
| 22:04.53 | eightyeight | "Oh, misery! how black I am and thoroughly rosted!" |
| 22:05.23 | eightyeight | "The boy turns and turns the spit, the fire cooks me through, the butler bastes me." |
| 22:05.44 | eightyeight | "Now I lie in the plate, and will never fly again; I see munching teeth!" |
| 22:06.58 | sjansen | Now I see what you have against cute pictures of cats. |
| 22:07.47 | eightyeight | this is the best rendition, imo: http://www.amazon.com/Orff-Carmina-Burana-Hagegard-Slatkin/dp/B000003FH3 |
| 22:09.25 | eightyeight | it's the feeling in lead vocalists that just nail it |
| 22:15.37 | levi | So, do you recall the trailer music for The Two Towers? I think it was that one, at least. |
| 22:16.13 | levi | Yeah, it was The Two Towers. |
| 22:16.33 | eightyeight | did they play a carmina burana song? |
| 22:17.18 | levi | No, but it was an orchestral/choral rearrangement of Lux Aeterna, a Clint Mansell piece he composed for Requiem for a Dream. |
| 22:17.42 | eightyeight | ahh |
| 22:17.45 | levi | The original was performed by the Kronos Quartet. |
| 22:18.10 | levi | It's pretty good stuff. It's since been used in a zillion film trailers. |
| 22:18.20 | eightyeight | that reminds me. i need to get john rutter's requiem one of these days |
| 22:18.27 | *** join/#utah gshipley (n=gshipley@nat/redhat/x-dcee548a7508a713) |
| 22:19.05 | elg | eightyeight: i think i'll have to disagree on that, but I don't have the time at the moment to go look up who did the recording I have. |
| 22:19.56 | eightyeight | elg: i'm interested in what rendition you perfer. i've only heard a few, but love the one i posted |
| 22:20.18 | elg | it sounds like an ok rendition, but perhaps a bit sappy |
| 22:20.18 | elg | gtg |
| 22:20.18 | levi | http://www.youtube.com/watch?v=7ZCW5x8cmFE&feature=related |
| 22:22.57 | eightyeight | levi: ahh. yeah. i forgot i have the soundtrack to all 3. i should listen to them when i finish off carl orff |
| 22:23.17 | levi | eightyeight: It's not on the soundtrack, it was just for the trailer. |
| 22:23.36 | sjansen | Amazon sure does make it easy to blow a couple hundred dollars on book in just 3 minutes. |
| 22:23.56 | eightyeight | yeah |
| 22:25.42 | levi | It's available here, though: http://www.amazon.com/Requiem-For-A-Tower/dp/B0019QKN3S/ |
| 22:26.19 | eightyeight | you know. we've now come full cicle |
| 22:26.36 | eightyeight | 13:54 <+elg> <3 the emiliana torrini station |
| 22:26.48 | eightyeight | she sang gollums song |
| 22:34.20 | *** join/#utah romanovic__ (i=romanovi@wireless34.wireless.utah.edu) |
| 23:03.00 | eightyeight | is playing "Carl Orff - Veni, veni, venias" on last.fm |
| 23:03.41 | eightyeight | nope. listening to 'In Trutina'. my favorite track, where sylvia mcnair's voice sells the operetta |
| 23:03.53 | *** join/#utah mabes_ (n=mabes@bmabey.fttp.xmission.com) |
| 23:10.57 | levi | http://www.etsy.com/view_listing.php?listing_id=19677785 |
| 23:30.02 | eightyeight | dulcissime shows the dynamic range as well. another selling point for this rendition |
| 23:31.07 | *** join/#utah Rolcol (n=Rolcol@unaffiliated/rolcol) |
| 23:40.00 | *** join/#utah WoodsDog (n=woodsdog@nat/novell/x-238210e8ccc75249) |
