| --- Log opened Tue Aug 08 00:00:55 2017 |
| 00:01 | | Jessikat`` [Jessikat@Nightstar-fq3g98.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 00:06 | | Jessikat [Jessikat@Nightstar-bni1j1.dab.02.net] has joined #code |
| 00:24 | <&[R]> | https://i.imgur.com/hPcXeUB.jpg <-- WTH |
| 00:29 | <&McMartin> | Does it really want the answer "work in base 13"? |
| 00:30 | <&[R]> | Maybe? |
| 00:32 | < Mahal> | Fuck common core |
| 00:33 | <&McMartin> | The other possibility is that it wants them to get "8+5 = 10+3" |
| 00:33 | < RchrdB> | McMartin, no, "make tens" is a term of art for the specific arithmetic mechanism they're using |
| 00:33 | < RchrdB> | yes |
| 00:33 | < RchrdB> | that |
| 00:34 | < RchrdB> | it wants them to learn that "8 is 2 less than 10 so I can make 10 from 8 by subtracting 2 from the thing on the other side of the + sign" |
| 00:35 | < RchrdB> | ...I think this is actually how addition with small numbers happens to work in my own head but I was definitely never taught it |
| 00:36 | <&McMartin> | At the time I was in high school, the fact that you could generalize "8+5 = 10+3" to " = 9+4" and so on wasn't explicitly covered on the default track until 9th grade. |
| 00:36 | <&McMartin> | I considered that a bit late, but I was hellaciously accelerated. |
| 00:36 | < RchrdB> | I think it's claimed that the Common Core way of doing arithmetic is supposed to have been studied formally with, randomised trials and control groups and so on and found to lead to better outcomes on some axis that someone cares about |
| 00:37 | <&McMartin> | Probably true. Reminder that Common Core is a number of US states deciding to share standards |
| 00:37 | < RchrdB> | I mean this specific "make tens" mechanism |
| 00:37 | <&McMartin> | Right. |
| 00:37 | <&McMartin> | You described it as a pre-algebra lesson, which is also the thing that was 9th grade default track back in '92. |
| 00:38 | < RchrdB> | I'm not suuuuuuper convinced, partly because doing science in pedagogy is hard, like, harder than medicine because you can't do double-blind, and *medicine gets results wrong and then corrects itself later all the time* |
| 00:39 | < RchrdB> | and partly because it sounds like the research may have been paid for by people who stand to make money from redoing all the math textbooks |
| 00:40 | < RchrdB> | (oh wait that's also the same thing as the difficulty doing medical studies right) |
| 00:40 | < Mahal> | I sort of understand the "make tens then add the rest later" idea, because that's how you function at Bigger Addition Later |
| 00:40 | < Mahal> | like 349+165 = 14 + 100 + 400 |
| 00:41 | < Mahal> | sorry, 4 + 10 + 100 + 400 |
| 00:41 | < RchrdB> | it does kind of sound dumb as all fuck when you talk through any of the steps aloud |
| 00:41 | < RchrdB> | OTOH gets you the right answer |
| 00:42 | < Mahal> | it's just badly worded. |
| 00:42 | < Mahal> | as McM said, 8+5 = 10+3 = basically how I learned math |
| 00:45 | | * [R] is now trying to do the mental excercise of "take two single digit numbers that have a sum greater than 10, how do you actually determine the correct result? What steps did you make?" |
| 00:45 | <&[R]> | AFAICT I'm at the point where I've just memorized the answers |
| 00:46 | | * Mahal erm |
| 00:46 | < Mahal> | honestly, I suck, hard at maths, always have. |
| 00:46 | < Mahal> | i have to do it manually every time |
| 00:47 | < Mahal> | and that there - making a ten & adding ther est - is pretty much how I do it |
| 00:49 | < RchrdB> | "sucking at arithmetic" and "sucking at mathematics" are very not the same thing though. |
| 00:49 | < Mahal> | OK, granted |
| 00:50 | < Mahal> | but on the basis that I suck so hard at anything leading up to and including algebra, I never got *to* calculus |
| 00:50 | < Mahal> | :P |
| 00:59 | <@himi> | My kid definitely remembers a lot of answers, but he's also learning to make use of those answers to work out things that can be related |
| 01:00 | | * Mahal nods |
| 01:00 | < Mahal> | I note, I'm not actually stupid. |
| 01:00 | < Mahal> | I just .. find it really hard to get things with numbers through my head. My logic doesn't ... work. |
| 01:00 | | * Mahal dunno. |
| 01:00 | <@himi> | So he remembered that 5*20 is 100, and he then used that to figure out that 130/5 was 26 |
| 01:00 | < Mahal> | because it's 20 + 30/5 which is 6 so 20? |
| 01:00 | < Mahal> | er, 26 |
| 01:00 | < Mahal> | ? |
| 01:00 | | * himi nods |
| 01:01 | <@himi> | Well, he counted up in fives to get that 30/5 was 6, but yeah |
| 01:01 | | * Mahal nod |
| 01:01 | < Mahal> | I rote-learned my times tables. |
| 01:01 | <@himi> | (which appears to be how he's doing division - counting up) |
| 01:01 | < Mahal> | sometimes I still have to recite them in my head to get to the answer I want, but I get there |
| 01:01 | <@himi> | Yeah, likewise |
| 01:02 | <@himi> | Apparently rote-learned times tables is discouraged now |
| 01:02 | < Mahal> | I kiiind of understand why |
| 01:02 | < Mahal> | but also find it a bit sad |
| 01:02 | <@himi> | I think it might be worth learning /after/ you already learn other stuff, but putting a lot of effort into teaching it in school is probably a waste of time |
| 01:03 | <&McMartin> | Rote-learning past 9 is definitely a waste. |
| 01:03 | <@himi> | I suspect rote-learning 2, 3, 5, and 7 would be useful |
| 01:03 | <&McMartin> | Heh |
| 01:03 | <&McMartin> | That's the kind of optimization that just makes things harder overall~ |
| 01:03 | <@himi> | Recognising how the others are related to those is /powerful/ |
| 01:04 | <&McMartin> | Yeah. I would expect it's harder to really get a feel for that unless you had 4, 6, 8, and 9 though. |
| 01:04 | <@himi> | Yeah, like I said I don't think it'd be worth learning as part of school, but already Xander's showing that being able to count in twos and fives is a very useful tool |
| 01:04 | <&McMartin> | Yep |
| 01:05 | <&McMartin> | A table from 1-5 and then "to get 6-9, do 5, then add the rest, woo distributive property, but we don't say that part out loud to first graders" makes some sense |
| 01:05 | <@himi> | . . . . though to be honest, I think Xander may be heading towards maths geniusy kind of thing, so I may not have a very good handle on the more general problems of maths education |
| 01:07 | <@himi> | Memorising times tables as a tool for use when applying other rules that you've already learned is going to be more valuable than learning the times tables by rote just so that you can do a table lookup on the basic problems |
| 01:07 | <@himi> | If you don't have the rules to make use of it then it's just a table lookup, not mathematical skills |
| 01:08 | <&McMartin> | Yeah. |
| 01:09 | <&McMartin> | I've been making the assumption that the algorithm for multidigit multiplication is hopelessly difficult to apply unless you have the 0-9 table handy for lookup. |
| 01:10 | <@himi> | The impression I've gotten from the common core problems I've seen is that they're teaching a whole range of approaches to solving problems, many of which don't need you to know those tables, or only part of them |
| 01:12 | < Mahal> | How old is Xander |
| 01:18 | <@himi> | Turning 7 in a few weeks |
| 01:23 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has quit [Ping timeout: 121 seconds] |
| 01:57 | <@macdjord> | himi: ... I remember cheering for you guys when you announced his birth. I feel OLD now. |
| 01:57 | | * Mahal too |
| 01:58 | <@Alek> | I remember them hooking up. -_- |
| 02:11 | | macdjord is now known as macdjord|slep |
| 03:18 | <&McMartin> | Heh. |
| 03:18 | <&McMartin> | // filthy, right? but it took me 8 seconds and covers all our cases. |
| 04:48 | <@himi> | ugh |
| 04:48 | | Degi [Degi@Nightstar-va3g87.dyn.telefonica.de] has quit [[NS] Quit: Leaving] |
| 04:48 | <@himi> | The more I look at the MSI docs the more I wonder what on earth is /wrong/ with MS developers |
| 05:02 | <&McMartin> | MSI is a very bad technology |
| 05:02 | <&McMartin> | What are you using to interact with it? |
| 05:03 | <@himi> | Nothing! |
| 05:03 | <&McMartin> | Why is it impinging on your world? |
| 05:03 | <&McMartin> | (If you have a choice of installer technologies, please use NSIS or Inno) |
| 05:04 | <@himi> | I'm trying to figure out how to make the Python distutils/setuptools msi installer option more flexible - specifically, how to run a script with different arguments on install and remove |
| 05:04 | <&McMartin> | Mmm. |
| 05:04 | <&McMartin> | OK |
| 05:04 | <&McMartin> | *They* should be using Wix. |
| 05:04 | <&McMartin> | Since that's the open-source tool for dicking around with MSIs |
| 05:04 | <@himi> | Or, I guess, how to tell it to create start menu shortcuts, but that's a whole new feature set for it - the post install script support is there, it just doesn't distinguish between install and remove |
| 05:05 | <@himi> | Hm |
| 05:05 | <&McMartin> | (Wix is also good enough that MS just flat-out started using it internally.) |
| 05:05 | <&McMartin> | (But "good enough" is a very, very qualified term here) |
| 05:05 | <@himi> | I noticed some wix references while googling |
| 05:06 | <&McMartin> | I had to maintain some Wix documents in a past life |
| 05:09 | <&McMartin> | My memories are extremely vague, but isn't getting shortcuts into place just a matter of a few extra directory tags in your wxs fiile? |
| 05:11 | <@himi> | Quite possibly |
| 05:11 | <&McMartin> | ( e.g. http://wixtoolset.org/documentation/manual/v3/howtos/files_and_registry/create_s tart_menu_shortcut.html ) |
| 05:11 | <@himi> | It's just a matter of creating the right shortcuts in the right places |
| 05:11 | <&McMartin> | I won't say wix is sane, but it's the *most* sane tool for this stuff I've ever heard of. |
| 05:11 | <&McMartin> | (But it's a big part of why I've always preferred NSIS despite NSIS bragging about how it's basically a cross between PHP and assembly language) |
| 05:14 | <&McMartin> | (Also, if I had to guess what combination of tragicomic events led to MSI, it was probably "an initial bodged-together system from whatever was handy, that turned out to do a bunch of things wrong because the original conception of the problem was fundamentally misconstrued, and by then it was too late to take back fundamental design decisions and a collection of 'best practices' and kluges accumulated |
| 05:14 | <&McMartin> | around it for 20 years") |
| 05:14 | <&McMartin> | (tabs in Makefiles, on steroids.) |
| 05:18 | <@himi> | I recall it being introduced with some fanfare, to replace all the bodged up custom installers that used to be out there |
| 05:18 | < Mahal> | compared to some of MSI's prequels, MSI was a godsend |
| 05:18 | < Mahal> | it's still a heap. |
| 05:20 | <&McMartin> | I consider NSIS best of breed in the space, and that's frankly a goddamn embarassment, too. |
| 05:20 | <&McMartin> | The problem MSI was clearly trying to solve was that installs should be inherently automatically reversible. |
| 05:20 | <&McMartin> | The problem with this is that, well |
| 05:20 | <&McMartin> | It's not true; they aren't. |
| 05:21 | <&[R]> | They should be though |
| 05:21 | <&McMartin> | There's a subset that are! |
| 05:21 | <@himi> | Coming from a *nix perspective the idea that they wouldn't be reasonably easily reversible is rather confusing |
| 05:21 | <&McMartin> | They are unstalled by unpacking an archive in a fixed location. |
| 05:21 | <&McMartin> | himi: ... um |
| 05:22 | <&McMartin> | There's a reason one of the steps in every major package manager, install and uninstall, is "run arbitrary script" |
| 05:22 | <@himi> | They're just files on the filesystem, after all |
| 05:22 | <&McMartin> | One certainly never edits files instead of creating or deleting them. |
| 05:22 | <&McMartin> | Certainly not in etc |
| 05:22 | <@himi> | Of course not! |
| 05:22 | <&McMartin> | There's definitely no talking to the outside world as part of the process |
| 05:23 | <@himi> | What kind of animals are we? |
| 05:23 | < Mahal> | this is why I rather like the OSX approach~ |
| 05:23 | <&[R]> | Most of them have that an an optional step (the [un]install scripts that is) |
| 05:23 | <&McMartin> | [R]: Right. But as soon as you do that you've conceded the entire argument. |
| 05:23 | < Mahal> | my /favourite/ is, and always will be, "I can't find my install media, fuck you I'm not uninstalling" |
| 05:23 | <&McMartin> | Now, user expectations for Windows installers are quite a bit more aggressive. |
| 05:24 | | * Mahal bangs head against desk in frustration. |
| 05:24 | <&McMartin> | Mahal: Yes. |
| 05:24 | <&McMartin> | Important design note, kids: keep install manifests somewhere besides the package itself |
| 05:24 | <&McMartin> | (That's what causes that with MSI) |
| 05:24 | <&McMartin> | Anyway, yeah |
| 05:25 | <&McMartin> | By the time 0.4 had rolled around, for UQM, we finally got a workable install flow for UQM on Windows |
| 05:25 | <&McMartin> | This involved started with a ~1MB file that included all the code, and knew both how to download the 10-700MB that was the rest of the installation based on options, and could also verify that it didn't need to download things because they were already there and had valid MD5 sums. |
| 05:26 | <&McMartin> | It would fetch things as needed and only as needed, and only once, because it would notice it was there afterwards. |
| 05:26 | <&McMartin> | It turns out there's a *nix tradition for *that* too |
| 05:26 | <&McMartin> | In UNIX terms, NSIS is the Windows technology for sharchives. |
| 05:27 | <&McMartin> | (And for cases where all you need are some files sitting somewhere, the Windows technology for that is "a zip file") |
| 05:28 | <&McMartin> | Mahal: It turns out the OSX approach is a gigantic hilarious nightmare of recursive failcannons, as is the Apple way, but as is also the Apple way, they only show up when you attempt to create software on your own instead of using it. |
| 05:28 | <&McMartin> | It turns out that OSX and "self-contained Windows solution" are the same technique, but Windows is better at it~ |
| 05:29 | <&McMartin> | From a "what goes into them" and "what they hardcode" perspective, .dylibs are worse by miles than ELF .so files or PE32 .dll files. |
| 05:30 | < Mahal> | McMartin: I'll grant you that I'm speaking from the power-user perspective. |
| 05:30 | <&McMartin> | Yeah. |
| 05:30 | <&McMartin> | That technique can be done on all three of the major systems. |
| 05:31 | <&McMartin> | And it seems like in the hobbyist Windows space, it's done more than in the commercial or enterprise spaces. |
| 05:31 | <&McMartin> | Linux is only starting to do it more in the wake of binary-only software distribution becoming big business in that space >_> |
| 05:32 | <&McMartin> | But after noticing it, a bunch of next-gen package managers seem to have dived into it |
| 05:32 | <&McMartin> | ... but as I noted the last time it came up, things like Nix and snappy seem to be reinventing not app bundles, but side-by-side installation (e.g., the 50,000 slightly different versions of libc and directx) |
| 05:33 | <&McMartin> | People who aren't MS fled screaming from WinSxS a long time ago, but perhaps the cases it's useful for are more common in serverland~ |
| 05:33 | <@himi> | On Linux that model is more targeted at containers |
| 05:33 | <&McMartin> | Yep. Containers + Deduping truly identical components is pretty much the core of WinSxS |
| 05:33 | <@himi> | At least, all the use cases I've seen (including snappy and the like) |
| 05:34 | <@himi> | I haven't seen anyone really concerned about dedup in those cases, though |
| 05:34 | <&McMartin> | Nix's version of co-existing hermetic builds with strictly versioned dependencies is the one that's the closest to it. |
| 05:34 | <@himi> | In theory, sure, it'd be nice, but disk is cheap |
| 05:34 | <&McMartin> | Er. NixOS, sorry. |
| 05:34 | <&McMartin> | Nix is the package manager. |
| 05:44 | <@himi> | Well, looking at the wix doco and having attempted to mangle my brain with MSI internals I can see how it's a reasonably sane way to model the same data |
| 05:45 | <@himi> | Also probably not hard to extend distutils/setuptools to support . . . |
| 05:45 | <@himi> | Not sure of the compatibility of the license with common open source licenses, though |
| 05:46 | <&McMartin> | Yeah |
| 05:46 | <&McMartin> | For Ophis I used MS's libc installer and made an NSIS script that unpacked my stuff into a place and then ran the runtime environment installer silently. |
| 06:07 | <@himi> | The Wix devs definitely aren't kidding when they talk about a steep learning curve |
| 06:27 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:28 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has joined #code |
| 06:28 | | mode/#code [+o celticminstrel] by ChanServ |
| 06:29 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 06:30 | | Jessikat` [Jessikat@Nightstar-94t9ha.dab.02.net] has joined #code |
| 06:30 | <&[R]> | "Task 'Published Calendars' reported error (0x88304005) : 'Server returned an error.'" Thanks MS. That's amazingly helpful. |
| 06:33 | | Jessikat [Jessikat@Nightstar-bni1j1.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 06:56 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 07:31 | | Kindamoody[zZz] is now known as Kindamoody |
| 08:03 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 08:07 | | Jessikat` [Jessikat@Nightstar-94t9ha.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 08:17 | <&McMartin> | New Bumbershoot post: https://bumbershootsoft.wordpress.com/2017/08/08/project-mehitabel-unixlib-and-s dl/ |
| 09:04 | | Kindamoody is now known as Kindamoody|afk |
| 10:34 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 10:34 | | mode/#code [+o macdjord] by ChanServ |
| 10:35 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 12:17 | <@gnolam> | <McMartin> I consider NSIS best of breed in the space, and that's frankly a goddamn embarassment, too. |
| 12:18 | <@gnolam> | Have you used Inno? I switched from NSIS to Inno, and I'm never going back. |
| 12:52 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 13:18 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has joined #code |
| 13:18 | | mode/#code [+o celticminstrel] by ChanServ |
| 13:48 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 13:49 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has joined #code |
| 13:49 | | mode/#code [+o celticminstrel] by ChanServ |
| 13:52 | | celticminstrel [celticminst@Nightstar-4r2i8l.dsl.bell.ca] has quit [[NS] Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] |
| 14:21 | | Kindamoody|afk is now known as Kindamoody |
| 14:52 | | Jessikat [Jessikat@Nightstar-m5i562.dab.02.net] has joined #code |
| 15:23 | | Degi [Degi@Nightstar-va3g87.dyn.telefonica.de] has joined #code |
| 17:56 | | Jessikat` [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 18:48 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 18:48 | | mode/#code [+o mac] by ChanServ |
| 18:51 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 19:58 | <&McMartin> | Re: Inno: I haven't had a chance to give it a fair shake yet. |
| 19:59 | <&McMartin> | It's possible this makes me a holdout due to accident of history. |
| 19:59 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 21:04 | | Jessikat`` [Jessikat@Nightstar-bgv74f.dab.02.net] has joined #code |
| 21:07 | | Jessikat [Jessikat@Nightstar-m5i562.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 22:11 | < ToxicFrog> | McMartin: well, the packages are from Nix, NixOS just extends that to "let's have build rules for things like "the entirety of /etc" and generate an entire immutable system as a build output" |
| 23:21 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds] |
| 23:47 | | Kindamoody is now known as Kindamoody[zZz] |
| --- Log closed Wed Aug 09 00:00:56 2017 |