| --- Log opened Thu Jun 12 00:00:13 2014 |
| 00:02 | | Xires is now known as ^Xires |
| 00:02 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 00:15 | | KiMo|autorejoin [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has joined #code |
| 00:15 | | ^Xires is now known as Xires |
| 00:16 | | Kindamoody[zZz] [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 00:20 | | KM|autorejoin [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has joined #code |
| 00:20 | | KiMo|autorejoin [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 00:21 | | KM|autorejoin is now known as Kindamoody |
| 00:29 | | KiMo|autorejoin [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has joined #code |
| 00:31 | | Serah [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 00:31 | | Kindamoody [Kindamoody@Nightstar-180u8i.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 00:31 | | KiMo|autorejoin is now known as Kindamoody |
| 00:33 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 00:44 | | Kindamoody is now known as Kindamoody[zZz] |
| 00:52 | | macdjord|slep [macdjord@Nightstar-7rac1r.mc.videotron.ca] has quit [[NS] Quit: âªRainbow Dash always dresses in style~âª] |
| 01:25 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 01:27 | | Nemu [NeophoxProd@Nightstar-kpuumf.gv.shawcable.net] has joined #code |
| 01:32 | | Xires is now known as ^Xires |
| 01:46 | | ^Xires is now known as Xires |
| 01:58 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 02:01 | | macdjord [macdjord@Nightstar-7rac1r.mc.videotron.ca] has joined #code |
| 02:01 | | mode/#code [+o macdjord] by ChanServ |
| 02:03 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 02:13 | | celticminstrel [celticminst@Nightstar-42flbc.dsl.bell.ca] has quit [Ping timeout: 121 seconds] |
| 02:21 | | celticminstrel [celticminst@Nightstar-tnvh3n.dsl.bell.ca] has joined #code |
| 02:21 | | mode/#code [+o celticminstrel] by ChanServ |
| 02:31 | | celticminstrel [celticminst@Nightstar-tnvh3n.dsl.bell.ca] has quit [Ping timeout: 121 seconds] |
| 02:34 | | Attilla [uid13723@Nightstar-ed0oqj.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 02:36 | | celticminstrel [celticminst@Nightstar-m2a422.dsl.bell.ca] has joined #code |
| 02:36 | | mode/#code [+o celticminstrel] by ChanServ |
| 02:59 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 03:04 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 04:00 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 04:05 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 04:09 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has quit [[NS] Quit: Program Shutting down] |
| 04:31 | | HotShot [fake@Nightstar-2cq4jk.sfldmi.sbcglobal.net] has joined #code |
| 05:01 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 05:05 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 05:06 | | Derakon is now known as Derakon[AFK] |
| 05:13 | | Turaiel[Offline] is now known as Turaiel |
| 05:21 | | RchrdB [RichardB@Nightstar-c6u.vd5.170.83.IP] has quit [[NS] Quit: Gone.] |
| 05:23 | | * McMartin prints out two pages of 25-year-old assembly code. SURRENDER UNTO ME YOUR SECRETS |
| 05:24 | | * Alek monocles McMartin. |
| 05:24 | | * Alek passes him the red pen. |
| 05:24 | | RchrdB [RichardB@Nightstar-c6u.vd5.170.83.IP] has joined #code |
| 05:29 | <&McMartin> | Oh man, I may use it |
| 05:29 | | * McMartin writes VERY POOR, SEE ME AFTER CLASS with it. |
| 05:29 | <&McMartin> | The whole reason for looking at this was to see how they handled the NMI problem, and as best I can tell, it was "by bricking the machine if the user hits RESTORE at the wrong time" |
| 05:31 | <~Vornicus> | ;_; |
| 05:31 | <~Vornicus> | "The NMI problem"? |
| 05:32 | <&McMartin> | The C64's equivalent of a BIOS can be swapped out for RAM. |
| 05:32 | <&McMartin> | So (a) w00t, that's a free 8K of storage space, but (b) wait a sec, that's where the interrupt vectors are, and (c) the RESTORE key is physically wired to the NMI signal |
| 05:32 | <&McMartin> | (b.2) that's also where the interrupt *handlers* are |
| 05:34 | <&McMartin> | So here we have a little BASIC-enhancement routine that's intended to swap out all the ROM chips to give you an extra 16KB of RAM... but that code has no defence against an NMI hitting while the BIOS-equivalent is swapped out. |
| 05:35 | <&McMartin> | Which will hard-loop the CPU on a BRK statement |
| 05:35 | <&McMartin> | It also appears to not be using the BASIC ROM's floating point interop routines. |
| 05:35 | <&McMartin> | I bet I can drop the size of this routine by a factor of 16 and have it still be safer than the original. |
| 05:49 | <@Reiv> | Is this a physical chip swap? |
| 05:53 | <&McMartin> | No, it's bankswitching |
| 05:54 | <&McMartin> | My rewrite is 22 instructions long compared to the original 99. So not 16 instructions shorter, I guess~ |
| 05:54 | <@macdjord> | So swap out, then restore the critical bits? |
| 05:54 | <&McMartin> | Er, 16x |
| 05:54 | <&McMartin> | You can "write to" the ROMs and it writes to the underlying RAM |
| 05:55 | <&McMartin> | So before I swap out the BIOS I write a new NMI vector and point it to a one-instruction handler that is just "return from interrupt" |
| 05:55 | | Turaiel is now known as Turaiel[Offline] |
| 05:56 | <&McMartin> | Another plus: I have no branches or jump instructions to stuff that is not somewhere in the ROM, so my version is completely relocatable. |
| 05:56 | <@macdjord> | Which, of course, has the side-effect of disabling the reset button. |
| 05:59 | <&McMartin> | Sure. |
| 06:00 | <&McMartin> | But you're only doing this for about a fiftieth of a millisecond at a time. |
| 06:00 | | * McMartin runs his version through an assembler. |
| 06:00 | <@Reiv> | McMartin: ... how'd they write so 'badly' then? |
| 06:00 | <@Reiv> | I mean it sounds like you just refactored the code to a fraction of the size in a couple hours. |
| 06:00 | <&McMartin> | My version is 44 bytes long, compared to the original's 228. |
| 06:01 | <~Vornicus> | woo, 5x reduction |
| 06:01 | <~Vornicus> | Reiv: having seen the sheer number of ways people can screw up fullscreen in Windows... |
| 06:01 | <&McMartin> | Reiv: I'm exploiting two things. |
| 06:01 | <&McMartin> | Possibly 3 |
| 06:02 | <&McMartin> | There are three problems with their code, and two of them are inexcusable |
| 06:02 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 06:02 | <&McMartin> | But the big one is excusable, because I'm exploiting a Dangeroud Forbidden Technique~ |
| 06:02 | <&McMartin> | *Dangerous |
| 06:02 | <@Reiv> | ... oh go on then, courter of the apocalypse |
| 06:02 | <&McMartin> | To wit, I am exploiting the fact that there is a specific routine in a specific place in the BASIC ROM that does about 80% of what I want here |
| 06:02 | <@Reiv> | ... that's a forbidden technique? |
| 06:03 | <&McMartin> | If the guy who wrote this 25 years ago didn't have that disassembly available, he would not have been able to do that. |
| 06:03 | <&McMartin> | Hrm |
| 06:03 | <&McMartin> | Actually, the book I got that from is dated 1984. |
| 06:03 | | * McMartin checks some publication dates~ |
| 06:03 | <&McMartin> | Anyway, if he didn't have This One Book, he didn't have that |
| 06:03 | <@Reiv> | Well, fair enough then I guess |
| 06:04 | <&McMartin> | However, then doing that in reverse at the end is somethign that *is most certainly* documented and was in the standard reference |
| 06:04 | <&McMartin> | So he totally should have used that |
| 06:05 | <&McMartin> | So that is inexcusable |
| 06:05 | <&McMartin> | And as for the NMI problem, well, presumably he figured "what are the odds that the user hits that key exactly when he's in the middle of this routine?" |
| 06:05 | <&McMartin> | viz. every software rant ever~ |
| 06:06 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 06:07 | <@macdjord> | Huh. Yeah, I would not consider a 0.0002s window of opportunity for failure to be a major issue. |
| 06:12 | <&McMartin> | It concerns me mainly because the consequences of failure are so drastic. |
| 06:12 | <&McMartin> | (I use a similar technique in my C64 equivalent of crt0.s) |
| 06:12 | <&McMartin> | (That is, the stuff that runs before and after main()) |
| 06:13 | | * macdjord shrugs |
| 06:13 | <@macdjord> | Still sounds like a Don't Do That Then to e. |
| 06:13 | <&McMartin> | Also because What The Hell Dude It Is Like Seven Instructions~ |
| 06:28 | <&McMartin> | There is a minor issue in that my code seems to not be working~ |
| 06:29 | <@celticminstrel> | Such a minor issue. |
| 06:29 | <@celticminstrel> | The most minor of all. |
| 06:33 | <&McMartin> | OK, some of this is me being a dummy |
| 06:36 | <&McMartin> | Now I've got 2/3s of it working by incrementing one constant by 1 |
| 06:40 | <~Vornicus> | hooray, off by 1 errors |
| 06:47 | | Kindamoody[zZz] is now known as Kindamoody |
| 06:59 | | AverageJoe [evil1@Nightstar-fb1kt4.ph.cox.net] has joined #code |
| 07:02 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 07:08 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 07:09 | <&McMartin> | This is still super-bizarre |
| 07:10 | <&McMartin> | It's successfully swapping out the BASIC and BIOS ROMs, but it's not successfully swapping *in* the character ROM. |
| 07:10 | <&McMartin> | There seems to be no good reason for that. |
| 07:11 | <&McMartin> | ... and setting a debug breakpoint makes the entire system hang. Good job, Linux porters of VICE. |
| 07:15 | <&McMartin> | Oh, I see. |
| 07:15 | <&McMartin> | I misunderstood what that register did |
| 07:16 | < simon_> | has anyone here implemented differential equations in haskell? I end up with a *lot* of zip4s and folds and the like. |
| 07:16 | <&McMartin> | And as a result instead of adding 16K of RAM and also exposing the character ROM, I instead exposed 24KB of RAM. |
| 07:16 | <&McMartin> | So that *will* complicate the logic some if I want it to do that. |
| 07:17 | < simon_> | (I'm basically pretending to do for-loops that index arr[i], arr[i-1], arr[i+1] or something similar.) |
| 07:18 | < simon_> | so I quickly end up with something like: zipWith3 f xs (x:xs) (tail xs) |
| 07:18 | <~Vornicus> | wait are you doing runge-kutta? |
| 07:18 | <~Vornicus> | If so, can you show me how? |
| 07:18 | < simon_> | I'm thinking maybe there are some nicer patterns. |
| 07:19 | < simon_> | Vornicus, I recently did Runge-Kutta in Python.. |
| 07:19 | <~Vornicus> | GIVE TO ZIIIIIIM |
| 07:20 | < simon_> | https://gist.github.com/sshine/7271208e54177daa3e0e |
| 07:23 | < simon_> | I'm not sure if using arrays in Haskell is preferrable because at least I can index wildly without folding like crazy. |
| 07:23 | < simon_> | using arrays in haskell is so close to coding imperatively I begin to wonder if there's any benefit ;) |
| 07:25 | <~Vornicus> | that code is a lot simpler than I've been led to believe. |
| 07:28 | < simon_> | yes, I even understand it! |
| 07:29 | < simon_> | it is very convenient to define the formula in terms of those F1, F2 |
| 07:46 | | celticminstrel [celticminst@Nightstar-m2a422.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 07:51 | <&McMartin> | Aha, got it working |
| 07:51 | <&McMartin> | If I don't put in the NMI defenses, it's now 32 bytes exactly |
| 07:52 | <&McMartin> | I do have one branch, but it's relative |
| 07:52 | <&McMartin> | ... and into the middle of an instruction. :D |
| 08:00 | <@macdjord> | Er, why? |
| 08:00 | <~Vornicus> | YOu are a terrible person |
| 08:01 | <@macdjord> | I don't find him particularly scary. |
| 08:04 | <&McMartin> | LDA #$32; BIT $30A9 |
| 08:04 | <&McMartin> | That $30A9 is secretly LDA #$30 |
| 08:04 | <&McMartin> | And BIT is normally a comparison instruction that's kinda a no-op |
| 08:05 | <&McMartin> | so I can branch past the first LDA and into the second to do a "set accumulator to one of these two values" things |
| 08:05 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 08:05 | | macdjord is now known as macdjord|slep |
| 08:07 | <~Vornicus> | so what's BIT actually do |
| 08:08 | <&McMartin> | It performs a logical AND of the accumulator and the specified value/memory location |
| 08:08 | <&McMartin> | If the result is 0, it sets the zero bit |
| 08:08 | <&McMartin> | it copies the sign bit of the value/memory location to the negative bit |
| 08:09 | <&McMartin> | And it copies the next-most-significant bit into the overflow bit. |
| 08:09 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 08:11 | <&McMartin> | CMP, meanwhile, does a subtract of the value/memory from the accumulator without modifying anything but the sign and carry bits. |
| 08:27 | <~Vornicus> | so it may mangle the math bits but if you're not using it |
| 08:34 | | Serah [Z@Nightstar-484uip.cust.comxnet.dk] has joined #code |
| 08:34 | | mode/#code [+o Serah] by ChanServ |
| 08:35 | | Nemu [NeophoxProd@Nightstar-kpuumf.gv.shawcable.net] has quit [Ping timeout: 121 seconds] |
| 08:49 | <&McMartin> | Right. It's basically poor man's conditional execution |
| 08:59 | | HotShot [fake@Nightstar-2cq4jk.sfldmi.sbcglobal.net] has quit [[NS] Quit: Trespassers will be shot, Survivers will be shot again! [Time wasted on Mirc 4hrs 27mins 37secs]] |
| 09:06 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 09:10 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 09:18 | | AverageJoe [evil1@Nightstar-fb1kt4.ph.cox.net] has quit [[NS] Quit: Leaving] |
| 09:18 | | Kindamoody is now known as Kindamoody|afk |
| 10:06 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 10:11 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 10:21 | | gnolam_ [lenin@Nightstar-lca497.cust.bredbandsbolaget.se] has joined #code |
| 10:21 | | gnolam [lenin@Nightstar-lca497.cust.bredbandsbolaget.se] has quit [Ping timeout: 121 seconds] |
| 10:49 | | mode/#code [+o RchrdB] by ChanServ |
| 10:58 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 11:02 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 11:08 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 11:30 | | VirusJTG [VirusJTG@Nightstar-6i5vf7.sta.comporium.net] has joined #code |
| 11:45 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 12:26 | <@Azash> | http://i.imgur.com/K8BQxfH.png |
| 12:27 | < simon_> | haha |
| 12:27 | | * TheWatcher facepalm |
| 12:48 | | gnolam_ [lenin@Nightstar-lca497.cust.bredbandsbolaget.se] has quit [Connection closed] |
| 12:49 | | gnolam [lenin@Nightstar-lca497.cust.bredbandsbolaget.se] has joined #code |
| 12:49 | | mode/#code [+o gnolam] by ChanServ |
| 13:29 | | Kindamoody|afk is now known as Kindamoody |
| 13:53 | | macdjord|slep [macdjord@Nightstar-7rac1r.mc.videotron.ca] has quit [[NS] Quit: Off] |
| 14:05 | | Serah [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| 15:14 | | Serah [Z@Nightstar-ro94ms.balk.dk] has joined #code |
| 15:14 | | mode/#code [+o Serah] by ChanServ |
| 16:07 | | Kindamoody is now known as Kindamoody|afk |
| 16:08 | <@Azash> | http://lifeofasoftwareengineer.tumblr.com/post/55789573990/git-reset-hard |
| 16:12 | | celticminstrel [celticminst@Nightstar-m2a422.dsl.bell.ca] has joined #code |
| 16:12 | | mode/#code [+o celticminstrel] by ChanServ |
| 16:14 | | You're now known as TheWatcher[afk] |
| 16:23 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has quit [Connection closed] |
| 17:11 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 17:11 | | McMartin [mcmartin@Nightstar-rpcdbf.sntcca.sbcglobal.net] has quit [Ping timeout: 121 seconds] |
| 17:11 | | McMartin [mcmartin@Nightstar-rpcdbf.sntcca.sbcglobal.net] has joined #code |
| 17:11 | | mode/#code [+ao McMartin McMartin] by ChanServ |
| 17:38 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 17:38 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 18:07 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 18:11 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 18:12 | | Attilla [uid13723@Nightstar-ed0oqj.irccloud.com] has joined #code |
| 19:07 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 19:12 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 19:21 | | HotShot [HotShot@Nightstar-v7se27.try.wideopenwest.com] has joined #code |
| 20:08 | | You're now known as TheWatcher |
| 20:08 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 20:10 | <&McMartin> | OK, after doing a little bit more research from last night: |
| 20:10 | <&McMartin> | (a) the really long program I drastically improved was published six months before the reference book I used to drastically improve it |
| 20:11 | <&McMartin> | (b) that program is credited to someone with the same first and last name as the crazy-awesome A/V tech at my high school back in the early 90s o_O |
| 20:11 | | Kindamoody|afk is now known as Kindamoody |
| 20:12 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 20:27 | | HotShot [HotShot@Nightstar-v7se27.try.wideopenwest.com] has quit [Connection closed] |
| 20:43 | <~Vornicus> | Coincidence? I THINK NOT. |
| 20:57 | | Kindamoody is now known as Kindamoody[zZz] |
| 21:04 | <&McMartin> | That would also indicate that (c) dude was probably like 15 when he wrote it |
| 21:04 | <&McMartin> | So these are all good excuses for my AUGH WHY questions. |
| 21:09 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 21:13 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 21:23 | <~Vornicus> | Indeed. |
| 21:39 | | HotShot [HotShot@Nightstar-v7se27.try.wideopenwest.com] has joined #code |
| 22:10 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 22:12 | | Serah [Z@Nightstar-ro94ms.balk.dk] has quit [Ping timeout: 121 seconds] |
| 22:14 | | Reiv [NSwebIRC@Nightstar-q8avec.kinect.net.nz] has joined #code |
| 22:14 | | mode/#code [+o Reiv] by ChanServ |
| 22:14 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 22:30 | | Derakon[AFK] [Derakon@Nightstar-5fqf0m.ca.comcast.net] has quit [Connection reset by peer] |
| 22:31 | | Derakon [Derakon@Nightstar-5fqf0m.ca.comcast.net] has joined #code |
| 22:31 | | mode/#code [+ao Derakon Derakon] by ChanServ |
| 22:38 | | Serah [Z@Nightstar-484uip.cust.comxnet.dk] has joined #code |
| 22:38 | | mode/#code [+o Serah] by ChanServ |
| 22:55 | | Orthia [orthianz@Nightstar-3tp.juj.184.203.IP] has quit [Ping timeout: 121 seconds] |
| 22:59 | | Orthia [orthianz@Nightstar-3tp.juj.184.203.IP] has joined #code |
| 22:59 | | mode/#code [+o Orthia] by ChanServ |
| 23:09 | < simon_> | heh |
| 23:09 | < simon_> | I'm in this small, newly started import cooperative |
| 23:09 | < simon_> | we import vegan products from germany and sweden because none are produced in denmark |
| 23:10 | < simon_> | almost all of the people involved do it for the cause and have no business experience |
| 23:10 | < simon_> | one guy stands out, though. he's this ex-professor who now works as the leader of a huge investment bank |
| 23:10 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has joined #code |
| 23:11 | < simon_> | also, his social skills are a little limited, so as far as I understand, what he has contributed is writing the bylaws, making a webshop and shouting at people on facebook for not living up to his expectations |
| 23:11 | < simon_> | so I made this mistake when I was in the shop and accidentally sold something that had been reserved through the webshop |
| 23:11 | < simon_> | the guy who'd reserved it didn't mind, but this one guy was upset about it because it undermines his effort. |
| 23:13 | < simon_> | but his way of saying that was "there's no possible way that you can derive your sales results without having served the guy who reserved his order, according to set theory [and then he gives a pretty decent proof of that]. also, if you look at the cosine similarity of your sales vector and his order vector, it seems like you did serve him. I'm not going to bother making the actual calculations." |
| 23:14 | < simon_> | so I thought, why not make them. :) (it *does* suggest that I did sell to him even though I didn't.) |
| 23:16 | | Syka [the@Nightstar-b1fuls.vividwireless.net.au] has quit [Ping timeout: 121 seconds] |
| 23:16 | < simon_> | > (order %*% sales) / (length(order) * length(sales)) |
| 23:16 | < simon_> | 0.3884298 |
| 23:16 | < simon_> | but shouldn't those vectors be normalized before this makes any sense? |
| 23:17 | < simon_> | (that's my question. the rest was ranting.) |
| 23:17 | < simon_> | he might've ordered 2 burgers and 2 hot dogs, and I might've sold 5 burgers and 7 hot dogs, and that's pretty similar. |
| 23:17 | | Serah is now known as JetpackWearingFireBreathingTRex |
| 23:20 | < simon_> | but shouldn't I normalize each vector according to something? like... if I served 5 people (who buy in the same way), maybe I should divide the sales vector by 5. or maybe I should consider some underlying distribution for each element in the vectors. I'm not even sure exactly... the only things I ever normalized, I knew were normally distributed. but we don't have the same amount of each item, and they're not equally popular. |
| 23:26 | <~Vornicus> | I'm so confused |
| 23:28 | < simon_> | hehe |
| 23:29 | < simon_> | ok. summary: I've got two vectors and I want to find the cosine similarity for them, but I suspect I need to normalize the vectors. |
| 23:29 | < simon_> | unfortunately, I have no idea how the vectors' underlying distribution is, but I think it's a mixture model of ten gaussian distributions. |
| 23:31 | < simon_> | this is our very first order on our very first day, though, so estimating the means (for normalizing them) is a trivial operation. |
| 23:56 | | JetpackWearingFireBreathingTRex [Z@Nightstar-484uip.cust.comxnet.dk] has quit [Ping timeout: 121 seconds] |
| --- Log closed Fri Jun 13 00:00:29 2014 |