#qi-hardware IRC log for Wednesday, 2014-03-26

qi-bot[commit] Werner Almesberger: atusb/fw/: support alternative interface names with DFU (master) http://qi-hw.com/p/ben-wpan/f1e689e00:04
DocScrutinizer05whitequark: from that evonik whitepaperabout acrylic handling: >>" Stichsägeblatt mit Geradverzahnun gund einer Zahnteilung von 2,5 mm<< (jigsaw: sawblade with in-line teeth [not one left one right bent tooth] with at toothpitch of 2.5mm)01:59
DocScrutinizer05311-5-tipps-zur-verarbeitung-von-plexiglass-de.pdf02:01
DocScrutinizer05s/Geradverzahnun gund/Geradverzahnung und/02:01
DocScrutinizer05also don't use "Pendelhub" when your jigsaw offers that (blade oscillating sideways)02:03
DocScrutinizer05and I bet water can't hurt02:03
rohnot sideways.. its oscillating in cutting direction: backtilt when down, forward when up02:12
rohwe cut acryllic with a lo2 laser usually02:13
DocScrutinizer05right02:17
DocScrutinizer05and then I dunno if it's deprecated or maybe exactly the right thing for cutting acrylic with jigsaw02:20
wpwrakoscillating sideways would affect the blade's life expectancy in dramatic ways :)02:23
DocScrutinizer05really?02:26
DocScrutinizer05-_-_-_-_-_-_02:26
DocScrutinizer05of course ideally it must also do the forth&back movement roh mentioned, to actually start cutting the overlapping 2nd cut path where it ended the upstroke before last02:29
DocScrutinizer05I don't know details about "Pendelhub" maybe because when I tried to find out, I had a little oopsie that can still be seen on my left digit finger's tip02:31
DocScrutinizer05the throttle of that jigsaw had a broken poti02:32
DocScrutinizer05no, I didn't cut my finger, only contusion between case of jigsaw and blade chuck02:33
wpwraklucky you :)02:40
DocScrutinizer05(MH370) what are those satellite data that the airplane sent to sats?03:59
DocScrutinizer05do airplanes have satellite-based animal-trackers, like whales?04:00
DocScrutinizer05or storks04:00
DocScrutinizer05would make a lot of sense, but then it's quite interesting to learn about a worldwide secret measure initiated by some TLA-agencies04:01
DocScrutinizer05even ETLA maybe, like FIAA?04:02
DocScrutinizer05err, scratch that04:02
DocScrutinizer05FAA04:04
DocScrutinizer05but EASA04:06
wpwrakyou mean ACARS ? that can be lots of things, engine data, other system performance data, text messages (kinda SMS), and so on04:20
wpwrakthe fate of AF447 was discovered largely based on ACARS data, long before they found the wreck with the black boxes04:21
wpwrakin MH370, however, ACARS stopped sending around the time the plane stopped the transponder04:22
wpwrakbut the satellite link was still active, just not sending data. so what they did was use management data from the sat link, which - surprisingly - gets recorded, to try to narrow down the plane's path and last position04:24
wpwrakthe "corridors" come from round-trip time measurements. but they also have some pulse trains that are affected by doppler. these were used in the latest analysis the picked the southern corridor04:30
wpwrakthere's some amount of detail here: http://avherald.com/h?article=4710c69b04:31
DocScrutinizer05http://www.ardmediathek.de/das-erste/report-mainz?documentId=2038547804:31
DocScrutinizer05(mh370) thanks for the info :-)04:32
whitequarkan adapter from HDMI to garden hose: https://pbs.twimg.com/media/BjoqFCqCMAAjogd.jpg:large13:25
whitequarkfor 'streaming video'13:25
larscna, that can't be any good the h2o connector is not gold plated13:28
wpwrakfinally ! i was getting tired of having to make them myself13:28
DocScrutinizer05of course HAMA, haha13:33
DocScrutinizer05good, Gardena is a better standard with more adapters than HDMI13:34
DocScrutinizer05Gardena even has phasers13:35
wpwrakzrafa: i suspect that part of your problem could also be an incorrect pin assignment in the schematics symbol. it's easy to get these wrong. so, if you put your design on a public repo, this can be checked, too :)13:35
DocScrutinizer05a word on multiple GND/VDD pins: Pins are the most "expensive" part of a chip, basically. A chip manuf tends to use all available pins for useful functions. So when they see the need for "redundant" VDD and GND pins, they damn sure *need* them to get connected and powered, since it was the only solution to bring clean power to all places on die where it's needed. Otherwise the chip manuf had found a better use for that pin, or worst 13:42
DocScrutinizer05case simply made it "Not Cpnnected" and save the expense to bond it13:42
DocScrutinizer05IOW: there is zero room to stray from chip manufacturer's application notes which clearly state which pin needs to get connected and which may stay unused13:44
wpwraknaw, there's usually plenty of room :) BUT ... the chip may no longer reach its specified performance at all points13:46
wpwrake.g., additional power pins may be less about distribution than about current13:47
DocScrutinizer05you need *öots* of experience for that, and you're on your own13:47
DocScrutinizer05*lots*13:47
wpwrak*öots* sounds cooler ;-)13:47
DocScrutinizer05and additional VDD pins usually are about additional buffering with low ESR13:49
wpwraka bit like "your arms will hurt because you have to stretch them to type since you can't sit close to the desk for your giant steel balls won't fit under it"13:49
DocScrutinizer05to the point :-P13:49
DocScrutinizer05o/13:49
wpwraklet's put it this way: i agree that it's generally a bad idea to try to take "shortcuts" with such things13:49
wpwrakbut then, you can often get away with it, especially if you're not driving the chip very hard13:50
DocScrutinizer05you generally have no idea about the consequences though13:51
wpwrakwhich makes me wonder whether zrafa is experiencing problems because he didn't connect all the VDD/VSS, or if it's because of something else13:51
DocScrutinizer05everything from electromigration to FDIV bug at <8°C may ensue13:51
wpwrakin fact, unless he missed all of them, i would expect this chip to work for what he's trying to do. i'm less certain about not having decoupling caps, which he seems to have "optimized out" as well13:52
DocScrutinizer05I tend to agree with whitequark who said "or you'll experience very magic behavior"13:52
DocScrutinizer05[2014-03-26 14:49:03] <DocScrutinizer05> and additional VDD pins usually are about additional buffering with low ESR13:53
DocScrutinizer05you might even get away with connecting _only_ a buffer cap to the VDD13:53
DocScrutinizer05charging of that buffer cap would happen through die traces at high ESR, but it might work13:55
wpwrakyeah :) i have something like that in anelok (by accident, noticed it yesterday)13:56
DocScrutinizer05afk13:56
wpwrakthat's also one of the reasons why i believe not connecting all the supply pins isn't zrafa's biggest problem13:56
DocScrutinizer05no buffer capacitors is the problem. afkafkafk13:57
wpwrakpoor kafka, gets blamed for everything13:59
whitequarkDocScrutinizer05: electromigration14:34
whitequark?14:34
whitequarkwpwrak: avrs tend to work through clamp diodes, even without any supply pins at all14:34
whitequarkin fact there's an attiny-based software-defined nfc tag working that way14:35
DocScrutinizer51whitequark: what's unclear about EM?14:49
DocScrutinizer51(clamp diodes) not only AVR14:50
DocScrutinizer51I already built such botch :) Along with other 'outa specs but works great and saves BOM'14:51
DocScrutinizer51like driving LED from GPIO high to GND w/o series R14:52
DocScrutinizer51you *can* do that when you really know what you're doing and are aware of and tolerate the consequences14:53
DocScrutinizer51for the R-less LEDs I had to caclulate total power dissipation in chip and accept that lifetime of that Atmel prolly considerably reduced14:56
DocScrutinizer51it been totally withing specs of that product14:56
DocScrutinizer51within*14:57
DocScrutinizer51~ping14:57
DocScrutinizer51eeew14:57
whitequarkDocScrutinizer51: how about just PWMing it?15:03
whitequarkor you didn't have hw pwm on that pin?15:03
wpwrakDocScrutinizer51: remember GTA02 with a BJT to drive a LED, the transistor not having a base resistor ? :)15:20
larscI remember15:22
larscit required a workaround in the gpio driver15:22
larscbecause the gpios would read back as 0 even though they were set to 115:22
wpwrakah yes :)15:23
whitequarkhah!15:29
whitequarkGPIO read on an output pin actually sampled the voltage instead of just returning stored value?15:30
wpwrakmakes the hardware a little bit more efficient, i guess :)15:31
zrafawpwrak: it is not open hardware. we are thinking to be millonaire and we need to write nda so all of you can sign to see the files15:31
zrafawpwrak: after the delirious I need to mention that our board does not have any thing to show :P .. it just exposes the pin vdd and vss (pin 3 and 4), swdclk, swdio, reset_b15:32
zrafawpwrak: then we can plug wires from ubb to test.15:32
wpwrakzrafa: (closed hw) ah good, then we can charge consulting fees ;-)15:33
zrafawpwrak: but of course, it was our hurry try to know if it would be so easy to talk with the mcu15:33
zrafawpwrak: but now I know that we need to work a lot more until swd software can talk with mcu :)15:33
wpwraksometimes it's faster not to take the shortcuts ;-)15:33
zrafawpwrak: however, the electronic engineering student guy did the tiny test board with eagle15:34
wpwrakyou should have insisted that he'd do it with kicad15:34
zrafawpwrak: and he will try to add the other bits to continue (1uF caps), the others vdd/vss, etc. At least to know that we are in some "hello world" stage15:34
zrafawpwrak: and turn on a led :)15:35
zrafavia software15:35
zrafain mcu15:35
zrafa:)15:35
wpwrakwell, since your project is still very simple, this is still a good time to switch to kicad15:35
wpwrakbtw, you can turn on a LED over SWD15:35
zrafaafter that close hw and millonaire. I will move to Cuba to live happy the rest of my life15:35
zrafawith the money of this project15:36
DocScrutinizer51haha, our kicad evangelist again :)15:36
wpwrakwell, one chip, a blinking LED, not much else to show ... yes, i supposed the plan for becoming a millionaire could work :)15:36
whitequarkDocScrutinizer51: what do you prefer? and would switching to kicad be actually worth it?15:37
wpwrakDocScrutinizer51: it's just sick to make yourself depend on closed source for they key piece of your hw development :)15:38
DocScrutinizer51depends on your priorities. every PCB house accepts eagle project files15:38
whitequarkDocScrutinizer51: well, I usually send gerbers15:39
whitequarkwhich is open format and is accepted in 100.0% of houses15:39
DocScrutinizer51wpwrak: one *some* level it a!ways becomes closed. Think chips15:39
DocScrutinizer51whitequark: I know15:39
whitequarkwpwrak: also eagle now has a documented xml format for its project/board files15:39
whitequarkso it at least doesn't lock you in15:39
whitequarkI'm sure some PCB CAD software today uses an equivalent of DRM, but it's not eagle15:40
wpwrakDocScrutinizer51: indeed. and that level is high enough. no need to get any further :)15:40
DocScrutinizer51well, up to anybody's very own priorities. I think kicad not being ready for primetime, for huge projects particularly15:43
DocScrutinizer51sure if you have the preoject requirement 'as FOSS as possible', go for kicad15:44
wpwrakyeah, i do. and it hasn't let me down so far :)15:44
DocScrutinizer51if you have the requirement 'as professional and safe as possible' then you prolly wanna go for a more matrure tool15:45
wpwrakhow would eagle be safer than kicad ?15:46
DocScrutinizer51and particularely when you wanna start learning about all that layout and schematics stuff, you don't pick a tool like kicad that needs a programmer to work around the quirks15:46
wpwrakyou haven't used kicad much, have you ?15:47
DocScrutinizer05whitequark: of course you can send gerbers, and they will accept them. They however usually won't check them and fix the little oopsies you introduced. On eagle otoh they might16:12
whitequarkDocScrutinizer05: I'm not really sure if I *want* them to fix what they consider little oopsies16:12
DocScrutinizer05well, as I already said, it's all up to you16:13
whitequarkthey seem to do some processing on gerbers as well. e.g. when I sent boards to iteadstudio, I did not actually send them a drill rack16:13
whitequarkso they "reverse-engineered" vias from top/bottom layers16:14
DocScrutinizer05unlike wpwrak I won't try to convince you to use any particular ECAD16:14
whitequarkwell, via drill points16:14
whitequarker, s,drill rack,drill layer,16:14
DocScrutinizer05I just say it's worth a closer consideration which tools to use, and you need to know what are your requirements to make an educated choice from the options you got16:15
whitequarkyeah, agreed16:16
whitequarkunrelated: seems like my decision to buy plywood (!) from amazon and then forward it internationally wasn't as dumb as it looks16:16
whitequarkI wasn't able to find thin plywood in either construction stores or on market16:16
whitequarkalso the only place which was selling thin (.5mm) copper was only selling it in ~1 sq.m. sheets, and there is no copper tubing thinner than 6mm too16:17
whitequark(I need a copper anode)16:18
wpwrakplywood seems to have some regional variations. i was also surprised to find that around here, it's not common (if it's used at all)16:18
whitequarkyeah, leroy-merlin has 12mm thick plywood. I don't even have anything to use such humongous logs for16:19
whitequark12+mm16:19
Action: whitequark ponders at http://www.ebay.com/itm/12128789157216:21
whitequarkactually I can probably make it myself, only thing I need is a mesh16:22
whitequarkI'm going to give up on non-screen-printing methods of depositing the silkscreen layer16:22
whitequarkI've more or less exhausted all options I knew16:22
whitequarkit actually seems that mastering the silkscreen technology can considerably help with avoiding noname dry film resist/mask, which you can never know where to buy when your current vendor evaporates16:24
DocScrutinizer05WTF? >>Shipping: May not ship to Germany<<16:32
kyakwalk a mile in my shoes... :)16:35
DocScrutinizer05USA creating a de facto embargo on Germany?16:37
DocScrutinizer05No, we still don't like that free trade treaty that allows US companies to sue Germany for compensation when we don't want your crappy gene manipulated corn food here16:38
DocScrutinizer05whitequark: I don't see how you're going to create any reasonable stencil from that basic kit16:39
DocScrutinizer05seems like made for freehand artwork, not for proper fine technical printing16:40
wpwrakDocScrutinizer05: (embargo) that's what you krauts deserver for complaining about the NSA !17:07
eintopfhi wpwrak 17:09
eintopfwhat's up17:09
whitequarkDocScrutinizer05: do you know how silkscreening for PCBs works?17:10
whitequarkyou spray or pour liquid resist on mesh, then smooth it out with squeegee and dry17:10
wpwrakeintopf: the cloud layer. a bit higher than earlier in the day.17:10
whitequarkthen expose and develop. at end you squeeze yet more resist through the mesh on the PCB17:11
whitequarkbe it etching resist, solder mask or UV-curing ink17:11
eintopfwpwrak: ah ok and that's good?17:12
wpwrakall things considered, it doesn't make much of a difference. well, maybe if you're a pilot17:23
DocScrutinizer05wpwrak: well, where's that liquid resist?17:39
DocScrutinizer05tbh I'd rather try to spray or pour light-curing ink to the PCB, then expose and remove the non-cured ink17:42
DocScrutinizer05with pigments that are not 100% opaque (particularly on the wavelength the resin needs or acrylate needs to cure) that should work out fine17:44
DocScrutinizer05or you even could use thermo-curing when you got pigments that absorb sufficient light to warm the ink up tpo the point where it cures17:45
wpwrakme ? liquid resist ? ain't doing none of that17:49
eintopfwpwrak: but... how are you17:50
eintopfI mean your feeling17:50
wpwrakgood, good. why ? do i seem odd ? i mean odder than usual ? :)17:51
DocScrutinizer05whitequark: ^^^17:51
whitequarkDocScrutinizer05: one word: uniform thickness17:55
whitequarkit's practically impossible to get uniform thickness of liquid resist without a mesh17:56
whitequarkand honestly I'm baffled as to what spray-on resist is useful *at all*17:56
whitequarkit gives you a hugely nonuniform layer and dries for hours17:56
whitequarkew17:56
eintopfwpwrak: no, I like the known the feelings about my friends :/18:06
eintopfwhen we are friends18:06
wpwrakah, alright :) so i can adjust the oddity level up ;-)18:08
wpwrakhmm, the "martian chronicles" should be fun. just saw a brief scene. spacecraft lands, astronaut emerges, greets oddly humanoid local, (speaking slowly) "I am from Earth. And this is Mars." local: "Young man, I don't know where you are from but this is Greenville, Illinois." :)18:12
DocScrutinizer05sounds like fun18:34
eintopfno I am level up because I know the author of my old bootloader :O18:46
kyakand your old filesystem, probably18:52
eintopfnever used fat under linux18:54
eintopfbut... I think every fat open source implementation used something code from that18:54
eintopfand indeed some usb sticks has fat :/18:54
whitequarkpretty much all usb sticks do18:55
wpwrakyeah, vfat (the driver) still has fat underneath18:55
wpwrakthe code is even remarkably unchanged :)18:55
kyakhttp://blogs.technet.com/b/microsoft_blog/archive/2014/03/25/microsoft-makes-source-code-for-ms-dos-and-word-for-windows-available-to-public.aspx18:56
kyakjust popped into my mind18:56
kyaknot sure if it was mentioned here before :)18:56
kyakit seems that everybody here is reading the code now :)19:02
wpwraknow us long-haired hippies can finally use a professional word processor to write our applications, and get a proper job19:02
kyakheh19:02
kyaki was amazed that they used some kind of bug tracking tracking system back in 8819:02
kyakif you grep the code for 'bug', you'll see19:02
eintopfwpwrak: do you put some things like backdoors in the fat implementation19:03
eintopfto get root access19:03
kyakeintopf: you didn't know? Just create a file named 'wernerisgod' in root of your fat19:04
eintopf:O19:04
DocScrutinizer05LOL19:06
eintopfwpwrak: did you reverse engineer the filesystem format?19:06
DocScrutinizer05of FAT?19:06
eintopfyes19:07
DocScrutinizer05seems that been public knowledge since very much the beginning19:07
eintopfah, ok19:08
eintopfnot like samba19:08
DocScrutinizer05I think I seen articles in magazines back in the late 80s19:08
DocScrutinizer05I mean, it couldn't get much simpler than FAT19:08
DocScrutinizer05they started with DOS1 which hadn't any folder concept19:09
dos1really? ;o19:09
DocScrutinizer05somewhere around dos2 or dos3 they added folders aka directories19:09
eintopfi don't like folders, i like directories19:09
wpwrakeintopf, DocScrutinizer05: yeah. just looked it up in a book on DOS programming19:10
whitequarkfun fact: in russian, "folder" and "daddy" is the same word19:10
wpwrakDocScrutinizer05: they added a few quirks over time, like different FAT formats, but yes, it's not the most complicated format on earth :)19:11
DocScrutinizer05eventually windoue came up with that botch which mitigated the 8.3 format restrictions19:11
kyakjust wondering, why MS was easy on revealing details about FAT, and so closed about NTFS? It took quite some years to have decent support for NTFS in Linux19:11
whitequark("?0?:0")19:11
wpwrakwhitequark: so "parent directory" is then a double daddy ?19:11
whitequarkwpwrak: exactly19:11
wpwrakgrandpa :)19:11
whitequarkwell, it could be interpreted as "grandfather"19:11
whitequarkthough the grammatical gender is wrong19:12
kyakcross-gender :)19:12
larsckyak: you have to ask wpwrak how easy it was to get the FAT details19:12
DocScrutinizer05kyak: NFS is a *real* filesystem, unlike that FAT nonsense19:12
wpwrakkyak: may be zeitgeist. back then few cared about keeping such things secret19:12
DocScrutinizer05NTFS*19:12
whitequarkdetails? you could get all details you need with a few hours with a hex editor19:12
whitequarkor a peter norton book19:12
whitequarkmy grandmother used to routinely resurrect deleted files with diskedit and some intense glaring at screen19:13
DocScrutinizer05hehe19:14
DocScrutinizer05sounds about right19:14
whitequarkback in the days, DOS used to just write \0 as the first symbol of filename19:14
DocScrutinizer05to delete, yes19:14
whitequarkand FAT directory entries aren't exactly complex19:14
kyakwhitequark: i spotted in comments to source code of msdos that directories have \0 at the end for "easier search" :)19:15
eintopfwas there a tool for fragmentation for fat under linux?19:15
DocScrutinizer05there are a few minor triccky things like secondary FAT which gets used when primary is corrupted19:15
eintopfhow many copies of the superblock do fat save?19:16
eintopf:\19:16
whitequarkI think 2?19:16
eintopfmhh19:17
DocScrutinizer05and I thnk before there been any FAT fs driver, there already been dostools19:17
eintopfthat's not many19:17
wpwrakmtools19:19
wpwraksuperblock ? one19:19
DocScrutinizer05ntools19:19
DocScrutinizer05mtools, yeah19:19
DocScrutinizer05mcopy19:20
DocScrutinizer05lol19:20
DocScrutinizer05A: B:19:20
DocScrutinizer05made me pull my hair19:20
DocScrutinizer05http://privatepaste.com/729c9280a319:21
kyakinteresting.. why not just mount it and then access?19:23
kyakor is it just legacy?19:23
wpwrakkyak: yup, mtools < kernel driver19:23
DocScrutinizer05kyak: when mtools were invented, there wasn't a FAT fs-driver to mount a FAT fs19:26
DocScrutinizer05mtools access directly the raw physical device, iirc19:27
DocScrutinizer05which can become really funny when you access a *non*FAT that way19:27
DocScrutinizer05wpwrak: did you sign for mtools too?19:28
DocScrutinizer05I can't find a (C) in man mtools19:29
kyakso werner saved people much typing :)19:30
kyakanyway, i briefly tried to run that "uncovered" msdos in virtualbox..19:31
kyaki then realized that i need IO.SYS19:31
kyaki found one from msdos 3.x, but it didn't work, and then i gave up :)19:32
kyakbut i think it would be interesting, especially if one would be able to build it from source19:32
kyak(there are both object and source files in archive)19:33
DocScrutinizer05http://distro.ibiblio.org/knoppix/docs/knoppix_tutorial/english/mtools.html19:33
DocScrutinizer05200419:33
DocScrutinizer05meh, msdos is sooooo out, I refused to use it when it been in, I preferred drdos or err PCdos?19:34
DocScrutinizer05I think there are FOSS dos implementations since ages19:35
DocScrutinizer05honestly nobody needs or just wants MSdos anymore19:36
kyakDocScrutinizer05: that's solely for historical reasons.. Frankly speaking i don't even know what i would do with that or any of FOSS implementations19:36
larscDocScrutinizer05: just you watch soon MSDOS will replace Linux as the leading opensource operating system19:40
wpwrakDocScrutinizer05: as far as i recall, i only contributed some small changes to mtools. i think it was a more sane configuration methos19:42
DocScrutinizer05wpwrak: on the lovely topic of "what means `A:´"?19:43
wpwrakyeah ,that kind of thing19:44
wpwraki think it was originally hardcoded or something like this19:44
DocScrutinizer05my hero!19:44
DocScrutinizer05yeah, was iirc. Now figure the fun when your boot HDD is second partition of hdb19:46
DocScrutinizer05or similar non-standard stuff you frequently see on dualboot/tripleboot systems of hackers like me19:47
DocScrutinizer05GOD I hated mtools19:47
DocScrutinizer05back in that time you could do funny stuff in linux, like `cd /x/y; mv .. x-foobar´  -  then you have to deal with a floppy and suddenly A: and B: and C: comes up19:49
DocScrutinizer05and now it all returns, the unbearable messing around with /dev/fd360 /dev/fd720ds and whatnot19:58
DocScrutinizer05*shudder*19:58
DocScrutinizer05ROTFL!!! >> The  VFAT  file system allows to store the case of a filename in the attribute byte, if all letters of the filename are the same case,<<20:04
wpwrakfd0H1400 and such ... i also freed the linux world from that :)20:06
DocScrutinizer05I owe you a fine bottle of vodka!20:08
DocScrutinizer05again LOL > >These formats are supported by numerous DOS shareware utilities such as fdformat and vgacopy. In his infinite hubris, Bill Gate$ believed that he invented  this,  and  called  it       `DMF disks', or `Windows formatted disks'. But in reality, it has already existed years before! Mtools supports these formats on Linux, on SunOS and on the DELL Unix PC.<<20:08
wpwrak(vodka) hehe :)  see "man setfdprm" for the gory details (or google for "setfdprm", in the likely event that you don't have it installed)20:10
eintopf:/20:10
wpwrakinterestingly enough, nobody seemed to care sufficiently to improve that man page :)20:11
DocScrutinizer05mshowfat  :-o20:14
eintopfwe should write a fat 2.020:15
whitequarkit's called ext220:15
DocScrutinizer05wpwrak: hey, you been at ETHZ in the early 90s?20:15
DocScrutinizer05you met Wirth? Maier, Boese?20:16
eintopfhe met my current professor for operating systems20:19
DocScrutinizer05err s/boese/vogt/?20:20
whitequarkDijkstra?20:20
DocScrutinizer05A&L Mayer&Vogt20:20
whitequark(he was visiting EHTZ in '94)20:20
eintopftanenbaum20:20
DocScrutinizer05hehe tanenbaum20:20
DocScrutinizer05tried to use his minix before linux came out20:21
eintopfi think tanenbaum never was at eth20:21
eintopfoh20:21
DocScrutinizer05alas it needed virtual memory which my machine didn't offer back when20:22
DocScrutinizer05MMU20:22
DocScrutinizer05Amiga was a nice machine, but alas had no MMU20:23
wpwrakDocScrutinizer05: ETHZ, 87-92, yup. crossed wirth a few times in the corridor. and no, i never met tanenbaum20:26
DocScrutinizer05mhm, then I undersatnd why you don't like Pascal and the successors ;-P20:27
DocScrutinizer05prolly they tortured you with pascal in your ETH courses. And honestly, plain Pascal is wirthless ;-P20:30
whitequarkyeah, it's no better than C20:31
whitequarkwho would voluntarily use such idiotic language, I don't even know20:31
DocScrutinizer05It lacked a few elementary vital things, like runtime file handles20:31
wpwraknaw, had to suffer MODULA, which is worse. and then a bit of oberon.20:31
whitequarkor closures and proper lexical scoping20:31
DocScrutinizer05oberon is terrible, Modula is rather nice I think20:31
wpwraki'd say it's just the opposite :)20:32
DocScrutinizer05actually turbo pascal been 90% modula and 10% pascal20:32
DocScrutinizer05hmm, maybe I never bothered about oberon much20:33
wpwraki hated modula. but oberon was less heavy. alas, wirth decided to keep it proprietary. well, his loss.20:33
DocScrutinizer05heavy as in "redundant"?20:33
DocScrutinizer05begin; end; instead of { }20:34
DocScrutinizer05or runtime heavy?20:34
DocScrutinizer05for runtime, the original wirth languages/dialects always been unusable20:35
wpwrakdunno about runtime efficiency. but yes, lots of typing. felt almost as sluggish as cobol20:35
DocScrutinizer05Turbo OTOH been the killer20:35
DocScrutinizer05oooh COBOL, there's nothing as terse a s cobol DOT20:35
DocScrutinizer05.20:35
wpwrakturbo pascal's been great. that's where i really started to learn programming.20:35
DocScrutinizer05I had the "joy" to use M-PASCAL (you know, of those siemens BS-M M80 monsters)20:36
DocScrutinizer05err PASCAL-M20:36
DocScrutinizer05had to typecast the IO handles, and write my own open() close()20:37
wpwrak;-))20:38
DocScrutinizer05you can't write recursive flile handling code in plain pascal20:38
DocScrutinizer05max 8 file handles, listed in first line of "PROGRAM foobar (in, out)"20:38
DocScrutinizer05or 4? can't recall20:39
wpwrakwho has more than 4 card readers anyway ?20:43
larscwho _needs_20:44
DocScrutinizer05right20:45
DocScrutinizer05BS-M had a flat file system device:filename with device something like "PLSK11" and filename 6chars20:47
DocScrutinizer05some genius at Siemens(?) invented "libraries" which mapped max 625 files from ASM.AA to ASM.ZZ to a thing like a dir, like ASM/7CHARFN20:49
DocScrutinizer05ASM.AA was the directory of that lib20:49
DocScrutinizer05those were funny times20:50
DocScrutinizer05if you wonder: Plattenspeicher Lesen Schreiben Kombiniert20:50
larscyea, when I started programming in 2002 or so things were much more boring20:51
DocScrutinizer05. NOTE DOT (end-if).20:53
DocScrutinizer05only way to make COBOL listings digestible20:54
DocScrutinizer05at least in COBOL-68(?)20:54
DocScrutinizer05sigh, COBOL. computed gosubs, no stack at all, real fun20:57
DocScrutinizer05I really wonder if the concept of a stack got invented after COBOL, or they simply were too stupid and ignorant to implement it in COBOL20:58
whitequarkCOBOL is the first high-level language ever20:59
DocScrutinizer05yep, I know, together with FORTRAN21:00
DocScrutinizer05but honestly, storing return addr in a storage cell at end of the subroutine? how stupid is THAT?21:00
whitequarkPDP-7 had this as the standard procedure call instruction in its ISA21:01
whitequarkwell, at start of the subroutine21:01
DocScrutinizer05aaah21:01
DocScrutinizer05oh21:01
DocScrutinizer05ok at start21:02
whitequarkmaybe they took inspiration from there21:02
whitequarkdon't ask how I know about PDP-721:02
DocScrutinizer05definitely21:02
whitequarkI briefly considered porting LLVM to it21:02
Action: DocScrutinizer05 can't help thinking about bread crumbs in deep forest21:02
whitequarkbut it has 12-bit bytes, and the amount of places LLVM assumes bytes = octets is so massive I gave up rather quickly21:03
DocScrutinizer05"when I went by here, I came from 7373175"21:03
whitequarkalso, PDP-7 had self-modifying code as a norm rather than exception21:03
DocScrutinizer05:nod:21:03
whitequarke.g. you were supposed to write most loops that way or something21:04
whitequarkI don't quite follow their broken logic21:04
DocScrutinizer05never touched PDP-7 only PDP-821:04
DocScrutinizer05which already been much saner, iirc21:04
whitequarkyep21:04
DocScrutinizer05PDP-7 must be older than me21:05
DocScrutinizer05:-o21:05
DocScrutinizer05nope, 5 years younger21:06
whitequark"you're older than PDP-7" sounds almost insulting21:07
larsc"you are so old, you used a hexeditor to un-delete files!"21:09
DocScrutinizer05life is insulting21:09
whitequarklarsc: that's not really old. I mean, it was still happening in late 90s21:09
whitequarkI could conceivably do that if I was a few years older21:10
DocScrutinizer05I'm so old, we used a shredder to delete files21:10
whitequarkhehe21:10
DocScrutinizer05or matches21:10
DocScrutinizer05punch paper tape is funny to burn21:11
DocScrutinizer05and you definitely can't undelete them with a hex editor21:11
DocScrutinizer05honestly though, when I started working in a IT job, we already used punch cards as notice papers21:12
DocScrutinizer05nobody punched them anymore, and only few of my colleagues knew how to read them21:13
DocScrutinizer05actually the first programmable device I touched in my life been a HP-2521:16
DocScrutinizer05errr nope, there even been some cheap cheesy crap several years before, a calculator with some 19 program steps or sth, no conditionals though, no loops21:17
DocScrutinizer05and actually I used the HP-65 before I had access to HP-2521:19
Action: whitequark used MK-61 a lot21:20
whitequarkhttp://en.wikipedia.org/wiki/Elektronika_MK-6121:20
DocScrutinizer05hah, nice21:27
whitequarkit eventually broke for some obscure reason21:29
whitequarkseems I could pick one up for $10 but I've already played enough with it21:29
DocScrutinizer05WOOOOW http://en.wikipedia.org/wiki/File:MK-61.gif  :-D21:30
DocScrutinizer05yeah, those nasty vacuum fluorescent tubes21:31
DocScrutinizer05eventually the electronics always died, prolly due to the high voltages needed to drive the tube21:32
whitequarkI think I simply dropped it unluckily enough21:34
whitequarktube didn't shatter but I guess there was a microfracture21:34
DocScrutinizer05either that or the heating/cathode wire broke21:45
DocScrutinizer05I prefer nixie tubes over those (usually green) fluorescent tubes any day21:47
DocScrutinizer05but for calculators they are even less suited than the green ones21:48
DocScrutinizer05http://en.wikipedia.org/wiki/File:Nixie_Wozniak.jpg  wow22:02
whitequarkDocScrutinizer05: enjoy: http://wathifi.tumblr.com/22:58
DocScrutinizer05The Halo CD1 player - LOL23:42
DocScrutinizer05I definitely should start selling such crap. I guess however that you first need to pay a lot of money to buy a positive article in one of those religious magazines23:43
DocScrutinizer05whitequark: facebook buys oculus23:50
whitequarkI know23:53
DocScrutinizer05I could use garden hose and fill it with mercury, to make a nice cable for audi-fanatics X-P23:56
wpwrakwhy can't they spend their money more sensibly and, say, buy qi-hw for a few hundred M$ ?23:56
whitequarkdoes qi-hw even still exist as an entity?23:56
wpwrakwe could then divide the loot by, say, IRC activity23:57
DocScrutinizer05haha23:57
DocScrutinizer05so qi-hw == $THIS ?23:57
wpwrakyeah. i'd get about 1/4, extrapolating from "top10"23:59
whitequarkwhoa23:59
DocScrutinizer05meh!23:59
wpwrakwhitequark: i think you'd agree with such a split, too :)23:59
--- Thu Mar 27 201400:00

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