#qi-hardware IRC log for Sunday, 2013-01-20

wpwrakworking again. thanks !01:44
viricWhat do you think of this procedure? https://www33.jvckenwood.com/gpl/gpl_form_en.html10:00
larscwell, it is gpl compliant, isn't it?10:02
viricyes :)10:02
viricclever.10:03
viricI wonder whether to ask for it10:03
viricI've no idea how much the postal sending will be10:03
viricok, I asked for it10:06
wpwrakviric: you may want to google for the product name. after all, once anyone receives their CD, they're free to redistribute it (at least the GPL code, in case there's more on it) anywhere they like. in case you're the first one, you should of course do so :)10:36
viricI tried, but I didn't find anything quickly10:43
wpwrakthen you'll be the first to completely liberate the sources :)10:48
virichehe10:48
viricI'd like to get into the device with some shell (it has busybox for sure, and a wifi adapter)10:49
viricbut that's another deal.10:49
Fallenouwpwrak: awesome piece of work, your pattern generator :)10:53
wpwrakthanks :) did you play with it ?10:54
FallenouI don't have any Ben unfortunately10:54
FallenouI might aquire one though one day, it seems a very nice little experimenting toy10:55
wpwrakit is :)10:57
wpwrakgreat typo: <1 << 3)  in a group of similar bit definitions. with the right font, even if the compiler tells you exactly where it is, it takes several minutes of staring to see this one20:16
Fallenounice catch =)20:20
Fallenouit's like special whitespaces which are not "space" that you usually type with a combination of alt|shift|your_cat_on_the_keyboard + space20:21
Fallenouyou can only spot them using hexdump on your source code :)20:21
Fallenouor just removing spaces randomly in your source code and replacing it with .. another space20:22
larsc .....20:26
wpwrakoh, indeed. Alt+Space produces 0xa0. wow.20:27
wpwrakluckily, even vi doesn't show this abomination as just a space20:27
larscwpwrak: try the characters I just pasted20:28
wpwrakpasting already turns them into '20:28
wpwrakoops. '#'20:29
wpwrakand the perverted space simply vanishes on copy & paste20:29
larscthere is even a visible space in unicode: #20:29
wpwrak(alt+spc)20:30
wpwrakit was already pretty daring to start using that 7th bit. why on earth would anyone in their sane judgment want 8 or even more bits per character ?20:32
larscbecause there are more than 256 characters?20:39
whitequarkI'm laterally thinking of what encoding would be a good idea to use in embedded programming20:44
whitequarkit seems to me that UTF-8 is probably optimal, with iteration over 16-bit codepoints20:45
whitequarkthe possible problems are various normalization algorithms, which AFAIK may require large tables, and surrogate pairs20:46
wpwraklarsc: there are the necessary characters. then there is upper/lower case. and then there are transcriptions :)20:50
wpwrakeverything else is just confusion :)20:50
whitequarkwpwrak: you english-speaking chauvinist!20:52
larscwpwrak: tell that to anybody who's not using the latin alphabet ;)20:52
larscwhitequark: hm, funny you can get the same character using different byte encodings, didn't knew that20:52
whitequarklarsc: that stuff is somewhat insane but is unfortunately required20:53
whitequarkthe NFKC/NFKD are especially useful when doing string comparison, e.g. in search queries20:53
larsc"Compatible sequences may be treated the same way in some applications (such as sorting and indexing), but not in others; and may be substituted for each other in some situations, but not in others. Sequences that are canonically equivalent are also compatible, but the opposite is not necessarily true."20:54
larscthat sounds like a lot of fun20:54
whitequarkoh exactly.20:54
whitequarkalso: In using normalization functions, it is important to realize that none of the Normalization Forms are closed under string concatenation.20:54
wpwraklarsc: oh, make kids learn to write in ASCII (let's call it SCII to avoid geographic bias) and in just a few generations you've solved that problem, too20:55
whitequarkwpwrak: it might be helpful to remember that majority of the world's population, in fact, couldn't care less about *SCII, much less enforce it20:56
wpwraklarsc: (confusing characters) it starts with the '. there seems to be a non-ASCII ' in some environments, not sure if it's in some keyboard layouts or a windows thing. very annoying to run into these.20:56
whitequarkand, by extension of the same logic, I would much like to see Americans using proper system, e.g. metric and big-endian dates.20:57
wpwrakwhitequark: yeah, the west missed an opportunity there, when asia was weak and impressionable :)20:57
whitequark(yes, I take this personally. fuck with encodings for a decade and you will as well)20:57
wpwrak(metric) maybe we could work out a deal between asia and the US ;-)20:57
whitequarkif only asia didn't use its own version of measurement system, which is way more fucked up than imperial units20:58
whitequark*versions20:58
wpwrakrussia is close enough to ASCII that the switch would be trivial :)20:58
wpwrakah, do they ?20:58
whitequarkwpwrak: well, from what I know there is quite a bit of these systems, all fragmented and empirically defined20:59
whitequarki.e. units derived from "buckets" and similar ad-hoc means of measurement20:59
whitequarkbut due to their quantity, it again seems that they do settle on metric mostly20:59
wpwraksounds familiar .. "foot", etc.20:59
whitequarkyeah yeah20:59
whitequarkwe had one in Russia too. which was, ironically, derived from imperial units20:59
wpwrakyeah, the US seems to be the last bastion of widespread use of non-metric21:00
whitequarkif I remember my history classes correctly, it emerged in early 1700s at the reign of Peter the Great and lasted until the communists took over21:00
wpwrakthere are plenty of funny measurements that all go straight to the romans (and who knows where they stole them)21:01
whitequarkas a side effect, 1U server has the height of >48= 25@H>:21:01
wpwraki hope that's an irrational number :)21:01
whitequarkwhich sounds reeeeally arcane and antique in Russian21:01
whitequark1U is an integral number of inches21:01
whitequarknot sure which exactly21:02
wpwraki meant ">48= 25@H>:" :)21:02
wpwrakOAIH BEPWOK. see, it's easy :) no, what do you cook in a bepwok ?21:02
whitequarkwell as I've said, it simply equals to the inches. it's basically a translation of "inch" dating back to 1700s21:03
wpwrakah wait, that was a reversed N, so OANH21:03
whitequarkpronounced somewhat like "odin vershok"21:03
whitequarkyeah, the google translate's spoken version of that is close enough to be understood21:04
wpwrakah, odin after sharing a few barrels of mead with thor ;-)21:04
whitequarkoh, unicode has "security considerations" http://www.unicode.org/reports/tr36/21:05
whitequarkbecause google.com and g>>gle.com are different things!21:06
larscunicode in domainnames is lots of fun21:06
whitequarki'm not exactly sure what purpose IDNs serve21:07
wpwrakjust don't do it :) most languages have quite tolerable transcriptions. and the chinese like numbers anyway, so no problem there either21:07
whitequarkthey look very weird, and you have to switch to latin to type http:// and whatever follows anyway21:07
whitequarkand anyone without a cyrillic layout cannot type them, etc21:07
wpwraki think the main purpose is to enable interesting new scams21:08
whitequarkas a russian, I would be happier (or as happy) without IDNs21:08
whitequarkoh rtl scripts interrupted by ltr sequences in domain names god why21:10
whitequarkwho on earth could think this would be a good idea in any circumstances?!21:10
wpwrakbecause there are rtl scripts ? :)21:11
wpwrakand the scammers like it, too: www.<rtl>knab<lrt>.com/login21:12
wpwraks/lrt/ltr/21:12
whitequarkyeah yeah21:14
larschello21:15
larschm, doesn't work on my terminal21:15
larsc21:15
larscwpwrak: see that's why you need more than 256 characters21:17
wpwrakall i see there is a tiny grey squiggle :) may be \^o. or an abstract portrait of tux ?21:19
whitequarka snowman21:20
larscit's a snowman21:20
larschttp://shapecatcher.com/21:22
wpwrakhmm, doesn't work here. probably too hot.21:29
larsc§21:29
wpwrakthis one is nice: http://news.ycombinator.com/item?id=203592621:29
larsc¨21:29
larsc©21:29
larsctry to use that in your c program ;)21:30
whitequarklarsc: am I stupid? see http://en.wikipedia.org/wiki/UTF-1621:30
whitequarkwhere's the U+FFFF..U+FFFFF range?21:31
whitequarkoh I am.21:31
wpwrakUnicode hexadecimal: 0x1f61c "Face with stuck-out tongue and winking eye". now, how do i add tilde and cedilla ?21:32
wpwraklarsc: probably more typing than BEGIN in Pascal ;-)21:32
Action: whitequark hands wpwrak a COMBINING TILDE U+030321:33
wpwrakargh21:33
wpwraki should have known ...21:33
whitequarkand a COMBINING CEDILLA U+032721:33
whitequarkI also have no clue what a cedilla is, and my font rendering engine neither21:33
larsca'21:34
whitequarkoh21:34
wpwrakand of course, they also have things like "Greek small letter upsilon with dasia and oxia"21:34
larscit's something at the bottom of the character21:34
whitequarkwpwrak: the job of unicode is to make every single other encoding obsolete21:34
whitequarkthus, they just plain include everything imaginable21:34
wpwrakcedilla looks a bit like a comma. in french it's below some cs ("c-cedille")21:35
whitequarkLinear Scripts A and B :D21:35
whitequarkthis is actually quite wise. first, because unicode actually reduces a huge PITA with encodings to a significantly less PITA with one encoding21:35
whitequarkand second, because languages die, and we don't want them to be buried with their respective encoding21:36
wpwrak"Canadian syllabics nwii" from "Unified Canadian Aboriginal Syllabics Extended". i think they forgot the name of the person in whose handwriting it looks exactly like that21:36
wpwrakwell, as some point you have more codepoints than reasonably distinguishable glyphs :)21:37
whitequarkwpwrak: this was actually a hot political topic21:39
whitequarkso-called Han unification21:39
wpwrakheh :)21:39
whitequarkthe unicode consortium has unified the Chinese hieroglyphs and corresponding Japanese derived ones21:39
wpwraki can imagine that unicode is a good place for political stakeholders21:39
whitequarknot only they have some subtle, but practically important, differences in rendering21:39
wpwrakand if the differences are invisible, they're even more important :)21:40
whitequarkbut also for both of these nations, unificating their language with the language of their oppressor has an immediate danger of someone going into a rage21:40
whitequarkthis is also the reason Ruby, instead of just settling on Unicode, includes support for all imaginable encodings21:41
whitequarkthe important part is that they feature Shift-JIS and friends :)21:41
whitequarkalso that made the encoding system a total clusterfuck impossible to reimplement in a compatible way21:41
wpwrakthe more i hear about unicode, the more i like my ascii unification idea :) at least people from all but a very small number of countries would have to give up some of their priced cultural heritage. so it's unfair to almost all. let's strike some nearly redundant letter and also the rest will be pissed off. how about the K ? there should be a great howl of agony in some places in the southern US.21:42
whitequarkfortunately, ruby 2.0 defaults on utf-8 everywhere and maybe they'll do something with the encodings...21:43
wpwrakjust strip and 8th bit and peace is restored21:43
wpwraks/and/the/q21:43
wpwrakhey, that was a "quit after the first substitution" !21:44
larscy can go aswell21:44
wpwrakjeah21:45
wpwrakthough it would be better to begin with things like I=1, O=0, S=5, etc.21:45
larscyea, we can actually get rid of 0-9 anyway21:46
wpwrakso instead of s/y//g how about s/i/y/g ?21:46
wpwrakcorrect. the romans had it pretty much figured out. back to the basics.21:46
wpwraklet's use Z for 0.21:47
whitequarkyou have invented the typewriter alphabet, congrats21:47
wpwrakthat one usually has numbers21:48
whitequarkhm21:49
wpwrakmaybe if you add a reverse button ... O, I, Z, rev-E, rev-h, S, G, rev-L, semi-rev-B, rev-G. yeah, that could work.21:49
wpwrakbut better to just switch to roman numerals. proven technology.21:50
whitequarkMany older typewriters did not include a separate key for the numeral 1 or the exclamation point, and some even older ones also lack the numeral zero. Typists who trained on these machines learned the habit of using the lowercase letter l ("ell") for the digit 1, and the uppercase O for the zero.21:50
larscjust use A-I instead of 0-921:50
whitequarklarsc: or ADGJMPTWZ21:51
larscno!21:51
wpwrakwho so inefficient and not A-Z ? 21:51
larsc;)21:51
larscbecause we only have IA fingers21:52
whitequarklarsc: i.e. 90?21:52
wpwrakby the time you've convinced everyone, genetic engineering will have a solution for that, too21:52
larscBA21:52
larscbee-te a21:53
qi-bot[commit] Werner Almesberger: ubb-patgen/timeline.fig: timeline of pattern generation (description in README) (master) http://qi-hw.com/p/ben-blinkenlights/a31351522:21
qi-bot[commit] Werner Almesberger: libubb/include/ubb/regs4740.h: add symbolic bit/field definitions for MSC_* (master) http://qi-hw.com/p/ben-blinkenlights/8509c1f22:21
qi-bot[commit] Werner Almesberger: libubb/mmcclk.c: use symbolic bit/field definitions for MSC_* (master) http://qi-hw.com/p/ben-blinkenlights/4cb68b722:21
qi-bot[commit] Werner Almesberger: ubb-patgen/ubb-patgen.c: use symbolic bit/field definitions for MSC_* (master) http://qi-hw.com/p/ben-blinkenlights/486471e22:21
qi-bot[commit] Werner Almesberger: libubb/include/ubb/regs4740.h: add symbolic bit/field definitions for DMA (master) http://qi-hw.com/p/ben-blinkenlights/5efafd622:21
qi-bot[commit] Werner Almesberger: ubb-patgen/ubb-patgen.c: use symbolic bit/field definitions for DMA (master) http://qi-hw.com/p/ben-blinkenlights/5025f3622:21
wpwrakhah ! now microsoft also have their competition for google's "project glass" (the somewhat mysterious VR glasses): https://www.youtube.com/watch?v=ZwModZmOzDs22:31
larscgood to see they still have some money left22:31
larscoh it's not real :/22:32
wpwraklooks good, though :)22:33
--- Mon Jan 21 201300:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!