**** BEGIN LOGGING AT Thu Feb 25 02:59:58 2010 Feb 25 04:09:40 hi everyone, is there a command to force reinstall all packages in SHR. The graphic acts weird after yesterday upgrade. Feb 25 04:16:01 opkg --force-reinstall ? Feb 25 04:24:01 DocScrutinizer: i found it: opkg -force-reinstall install `opkg list_installed | sed 's/ - .*//'` Feb 25 04:24:35 oops, the sed was new to me Feb 25 04:25:52 otherwise that's basically same as the simplest method to 'upgrade' on low memory Feb 25 04:38:48 here comes more snow Feb 25 04:40:09 keep it with you, please. Almost bearable temperatures here atm Feb 25 04:41:16 wooohooo! forecast says 10C today, even 12C on Sunday :-D Feb 25 04:49:50 0C on Sunday... -10 or so right now. Feb 25 04:50:01 Global warming, apparently. Feb 25 04:50:05 Sigh. Feb 25 04:50:37 I should sue Al Gore and force him to deliver on his promise. My heating bills are insane this year. Feb 25 04:52:04 hmm, from what I seem to remember, I got the impression you burn the cut down trees from your large garden Feb 25 05:01:34 well thefuckingweather.com says it's +2C right now, but my RF-thermometer reads 4.2C Feb 25 05:01:50 downtown bonus ;-) Feb 25 05:33:02 ~0C here Feb 25 05:33:34 Feels surprisingly warm. Too warm to walk in my regular clothes in fact. Feb 25 05:35:51 DocScrutinizer: is it still evening for you? :) Feb 25 05:36:03 yup Feb 25 05:36:45 though I decided to witch to "morning mode" without any intermediate states Feb 25 05:37:01 Good morning then :D Feb 25 05:37:10 good morning :-D Feb 25 05:37:49 time for a coffee Feb 25 05:59:13 PaulFertser: any notion how hard it might be to RE and patch a armel ELF? Feb 25 05:59:50 DocScrutinizer: i guess it all depends on the size and complexity of a particular binary. Feb 25 06:00:47 hmm, yes for *complete* RE that's true, idependant of arch Feb 25 06:01:02 DocScrutinizer: my latest x86 REing attempt ended in me needing help with understanding dynamic relocations. objdump -R solved most of those but other cases of indirect procedure calls still puzzle me. Feb 25 06:01:20 DocScrutinizer: what are you after? Feb 25 06:01:33 yep that's more of the issues I had in mind Feb 25 06:02:08 patchiung maemo5 mce to keep fingers out of LP5523 kbd backlight engine3 Feb 25 06:02:11 for example Feb 25 06:03:16 DocScrutinizer: a perfect opportunity to test StsKeeps's opensourcing system, no? ;) Feb 25 06:03:27 yes, as well Feb 25 06:05:34 DocScrutinizer: if the method to control LP5523 is known patching out access to it shouldn't be that hard. Also you can probably go the other way: LD_PRELOAD tricks or even a SELinux rule ;) Feb 25 06:05:37 look, they're wasting engine3 for stupid silly kbd backlight switch-on/off. For e.g. screen-BL they obviously can do that in sw. So WTF are they blocking one of the three engines, while you got a 3color notification LED Feb 25 06:06:01 hrhrrrhrrrrrr Feb 25 06:06:06 * DocScrutinizer likes those ideas Feb 25 06:06:41 access to LP5523 is via sysfs very likely Feb 25 06:07:40 could SElinux stop access to a specific sysfs node? Feb 25 06:08:29 raster: no CAD files for the case? Why? Feb 25 06:08:40 huh? Feb 25 06:08:40 raster: how on earth does it make any sense? Feb 25 06:09:07 DocScrutinizer: you do not read community ML where raster shares his insights on the business and mobile devices ;) Feb 25 06:09:24 are you kidding Feb 25 06:09:26 ? Feb 25 06:09:32 DocScrutinizer: sure, SELinux can do that. Feb 25 06:09:48 I stopped reading [community] some 15months ago Feb 25 06:10:26 PaulFertser: i dont see that happening. the hw side of makidg gadgets is far from even clooking at "open source" Feb 25 06:10:32 (SE) never used it Feb 25 06:10:33 letalone actually being freidnly to it Feb 25 06:10:35 DocScrutinizer: http://lists.openmoko.org/pipermail/community/2010-February/060482.html Feb 25 06:10:42 its not going to happen because of the reality of life Feb 25 06:11:00 you have no support - and you actually have no good case to make Feb 25 06:11:06 source for software makes sense Feb 25 06:11:24 for hw design (be it schemtics or case) it makes very little sense Feb 25 06:11:33 raster: unavailability of CAD files doesn't stop chinese manufacturers from producing _numerous_ clones and derivatives of all kinds. Feb 25 06:12:32 raster: i'd guess case designs would make an opportunity for custom niche products with added on functionality and/or ability to adopt a general purpose device to specific needs. Feb 25 06:13:03 * DocScrutinizer moos at raster Feb 25 06:13:53 DocScrutinizer: mooo! Feb 25 06:14:02 PaulFertser: it still doesnt make sense Feb 25 06:14:05 raster: just the same as software: you advocate opennens because of 3rd-party software developers that can produce both mass- and niche-market solutions. Why hardware is not the same? Feb 25 06:14:28 1. open for software makes sense. why? the cost of entry into looking at code. modifying and trying those modificatiosn and distributing them is VERY LOW Feb 25 06:14:31 not so with hw Feb 25 06:14:37 for the schematics a honestly disagree. There's no other documentation about how to write drivers for, and exploit fully the hw Feb 25 06:14:37 be it electronics or case Feb 25 06:14:48 its a totally different world of economies Feb 25 06:14:53 _And_ extending it. Feb 25 06:15:13 opening schemtics AND case design simply hads everything on a platter to chinese knockoffs Feb 25 06:15:18 If one wants to add an i2c magnetometer or altimeter or something, what options will it have with your hardware? Feb 25 06:15:26 i even support companeis not being open on their own apps - themes, icons nand other ip Feb 25 06:15:30 the thnig that makes the product theirs Feb 25 06:15:49 the core software thatrs the same prettymuch everhwhere - thats one thing Feb 25 06:16:13 PaulFertser: you can make that argument all you like. the economies are utterly different Feb 25 06:16:13 raster: that's a joke. chinese cloners buy 3 devices, desolder everything and let 9 blue monkeys check eacxh trace with a voltmeter Feb 25 06:16:25 "ip" doesn't exist in fact and you know that. This term was constructed to lie about the life and the law. Feb 25 06:16:37 *THEY* do _not_ need the schematics Feb 25 06:16:38 where u have 10,000 develoeprs ready to screw with your code, you have 10 ready to build their own i2c attachments Feb 25 06:16:51 DocScrutinizer: sure - but they do it illegally. Feb 25 06:17:07 they simply don't care Feb 25 06:17:13 DocScrutinizer: if it was open - you'd hand that info on a platter WITH a license allowing them to make clones withotu repercussion Feb 25 06:17:13 And who cares? Feb 25 06:17:33 raster: open != copyleft Feb 25 06:17:43 PaulFertser: ip defintiely exists. and i find it offensive for you to state otherwise. Feb 25 06:17:53 nope, copyright on the PCB is no way different to copyright on schematics Feb 25 06:18:02 PaulFertser: if you dont think ip exists then gpl is bunk. ip with open licenses it moot ad ip doesnt existew. Feb 25 06:18:03 raster: where does it? Which law is that? Feb 25 06:18:08 its a free-for-all public domain Feb 25 06:18:45 PaulFertser: open == copyleft when you refer to it in the same breath as open source Feb 25 06:18:59 ie licenses that dont just give code - but allow modification and re-distribution legally. Feb 25 06:19:11 raster: you're spinning away Feb 25 06:19:17 raster: i know people usually use "ip" to talk about utterly different (legally) things: about copyrights, trademarks, software patents even. And that's all entirely different legislation. Feb 25 06:19:31 PaulFertser: ip == intellectual property Feb 25 06:19:38 ownership of something that is not tangible Feb 25 06:19:48 thats is the entire foundation of copyright Feb 25 06:19:54 yep Feb 25 06:19:56 exactly Feb 25 06:20:05 [2010-02-25 07:17:53] nope, copyright on the PCB is no way different to copyright on schematics Feb 25 06:20:16 and if u dont think ip exists - then you instantly undermined gpl, bsd, apl and everhy open source app and its license Feb 25 06:20:18 raster: are you sure that's written in some law? I really doubt that and i've seen some articles on the topic. Feb 25 06:20:28 it is Feb 25 06:20:29 raster: copyright does exist. Feb 25 06:20:31 it's called copyright Feb 25 06:20:31 nobody thinks IP doesn't exist Feb 25 06:20:40 raster: copyright != ip Feb 25 06:20:42 copyright ISಘಝ Feb 25 06:20:48 err is IP Feb 25 06:21:07 DocScrutinizer: "ip" doesn't exist in fact and you know that. This term was constructed to lie about the life and the law. Feb 25 06:21:11 raster: copyright is a well-established legal term, with all laws and lawyers etc etc attached. IP is a buzzword. Feb 25 06:21:39 but chinese cloners won'T care of any copyright, no matter if it's printed on a schematics or a PCBA Feb 25 06:21:54 PaulFertser: ip is a more geenral term also covering other things like patents Feb 25 06:22:10 and dont tel me patents are not an established method of intellectual property ownership Feb 25 06:22:19 this discussion absolutely misses the point Feb 25 06:22:23 you look like a fool if you do. Feb 25 06:22:34 DocScrutinizer: yes. but you have at least a legal stick to go after them with Feb 25 06:22:35 raster: it has no legal meaning. The patent legislation is not the same as copyright law and iirc has little similarity legally-speaking. Feb 25 06:23:03 raster: same with (C)'d schematics Feb 25 06:23:35 DocScrutinizer: it doesnt change thew fact that the economics of software and hw are entirely different Feb 25 06:23:48 it doesnt chaneg the fact that opening up schematics makes it easier for cloners Feb 25 06:23:50 it does Feb 25 06:23:51 nobody claimed that Feb 25 06:23:54 raster: also i'm not sure copyright is technically "property" at all. Feb 25 06:24:02 and there is no BENEFIT to the companies Feb 25 06:24:05 yes Feb 25 06:24:20 yolu can say "hey me - i can read a chematic and find that you have hooked compnent a to component b badly" Feb 25 06:24:24 as above Feb 25 06:24:29 for every 1 person that can and wil ldo that Feb 25 06:24:37 there are 10000 software enigners who will do it with software Feb 25 06:24:37 raster: The effort to RE the PCBA is negligable for cloners Feb 25 06:24:51 there is an economic benefit in giving away soruce in return for that Feb 25 06:25:00 there is no benefit for hw as u just dont have the same scales Feb 25 06:25:07 and as i said in email Feb 25 06:25:16 softwre side is MUCH MORe friendly to open soruce already Feb 25 06:25:21 its a battle that is winnable Feb 25 06:25:41 [2010-02-25 07:14:36] for the schematics a honestly disagree. There's no other documentation about how to write drivers for, and exploit fully the hw Feb 25 06:25:44 tryng to fight for open hw is like trying to fight the sun from coming up each day Feb 25 06:25:48 pointless battle Feb 25 06:26:45 heh, I made the sun stop coming up for one day then :-D Feb 25 06:26:47 raster: how to hook up extensions then? Feb 25 06:27:23 the hw and software divisions run as separate groups Feb 25 06:27:30 withs eparate management and cultures Feb 25 06:27:41 and i can say that they dont give a rats arse about open Feb 25 06:27:46 PaulFertser: It's absolutely pointless to focus on extensions. See hostmode of N900 for a killer example Feb 25 06:27:48 it doesnt even register on the interest scale Feb 25 06:27:50 Also in fact schematics are usually widely available. I see no harm making CAD files availabl as well. Feb 25 06:27:59 the software side sees a benefit and is thus willing to give it a go Feb 25 06:28:06 fight the battles you can win. Feb 25 06:28:12 but.. feel free to fighting your hw battle Feb 25 06:28:17 you'll be fighting it for decades. Feb 25 06:28:28 raster: the software side can't succeed without schematics Feb 25 06:28:34 of course it can Feb 25 06:28:41 thats just wrong Feb 25 06:28:43 DocScrutinizer: i still think hostmode of n900 can work. It's just nobody from Nokia and TI want to help understand that fucking testmode register. Feb 25 06:28:57 raster: if ever you were to write a kernel driver, you'r change your mind Feb 25 06:29:15 DocScrutinizer: no one will be making you write the drivers from scratch Feb 25 06:29:26 u will get them already written for that hw Feb 25 06:29:30 so its moot Feb 25 06:29:31 so it's no FOSS system Feb 25 06:29:36 period Feb 25 06:29:38 neither is a pc Feb 25 06:29:43 has it mattered? Feb 25 06:29:51 its totally moot Feb 25 06:30:00 your points have been proven wrong for decades in the pc world Feb 25 06:30:10 nonsense Feb 25 06:30:16 foss systems exist minus schematics or any cad designs Feb 25 06:30:20 Now kick me for spamming: Feb 25 06:30:21 Now kick me for spamming: Richard Stallman argues that, although the term intellectual property is in wide use, it should be Feb 25 06:30:24 rejected altogether, because it "systematically distorts and confuses these issues, and its use Feb 25 06:30:27 was and is promoted by those who gain from this confusion." He claims that the term "operates as a Feb 25 06:30:30 catch-all to lump together disparate laws [which] originated separately, evolved differently, Feb 25 06:30:33 cover different activities, have different rules, and raise different public policy issues."^[18] Feb 25 06:30:34 every laptop, desktop and server shipping wiht linux or having it installed on it Feb 25 06:30:36 Stallman advocates referring to copyrights, patents and trademarks in the singular and warns Feb 25 06:30:39 against abstracting disparate laws into a collective term. Feb 25 06:30:40 you can write drivers Feb 25 06:30:41 and make them work Feb 25 06:30:45 and everything else Feb 25 06:31:38 raster: uhuh Feb 25 06:31:54 you can RE windoof drivers Feb 25 06:32:03 you can Re wince drivers Feb 25 06:32:07 or android drivers.. Feb 25 06:32:09 he.ll u have the src Feb 25 06:32:22 for WAHT? Feb 25 06:32:34 the drivers? Feb 25 06:32:40 its moot if the device already COMES with linux and open drivers Feb 25 06:32:48 so pointless. Feb 25 06:32:54 yep, thats a valid point Feb 25 06:33:07 sure Feb 25 06:33:10 there may be bugs Feb 25 06:33:14 and some features are not leveraged Feb 25 06:33:16 it happens Feb 25 06:33:22 hap-pesn on closed drivers too Feb 25 06:33:23 in windows Feb 25 06:33:28 and anywhere else Feb 25 06:33:36 BUT Feb 25 06:33:43 who cares about windoze?? Feb 25 06:33:45 your freedoms have been preserved Feb 25 06:33:53 you are free to modify them - audit them and do what u like Feb 25 06:34:01 and the entry barrier is low to do that Feb 25 06:34:11 modifying hw - entry barrier is very high Feb 25 06:34:12 Without the datasheets it's not that enjoyable Feb 25 06:34:17 and it gets higher all the time Feb 25 06:34:40 PaulFertser: reality is on most of these devices u dont have datasheets Feb 25 06:34:42 so what? Feb 25 06:34:43 or very buggy incomplete ones Feb 25 06:34:51 raster: that's what i say Feb 25 06:34:52 you have hw put together on the fly Feb 25 06:34:52 I never said I want to mod hw Feb 25 06:35:05 and as an issue is fixed some kernel driver dev is sent a mail and told to fix something Feb 25 06:35:20 releasing schematics is extra overhead that will cost money and thus raise unit prices Feb 25 06:35:24 ok - so there is a cost Feb 25 06:35:31 what is the BENEFIT to releasing them? Feb 25 06:35:44 will the benefit balance the cost.. the answer i would say is a firm no Feb 25 06:35:52 and i can tell you that thats how the companies think Feb 25 06:36:03 better drivers -> full exloiting the hw Feb 25 06:36:12 you have no case to make until you can show convincingly that there is gerater benefit than cost Feb 25 06:36:32 and i'd say all the facts to this day squarely land in the camp of the hw makers beliefs that there is not Feb 25 06:36:41 cost is about, err... ZERO? Feb 25 06:36:44 DocScrutinizer: not signficantly enough to bother Feb 25 06:36:50 DocScrutinizer: its not. see above Feb 25 06:36:58 PaulFertser: reality is on most of these devices u dont have datasheets Feb 25 06:37:07 or very buggy incomplete ones Feb 25 06:37:10 ... Feb 25 06:37:20 someone needs to put them together Feb 25 06:37:22 maintain them Feb 25 06:37:28 make them presentable for the outside world Feb 25 06:37:47 raster: what does it mean? How the drivers were developed in the first place if there was no documentation? Feb 25 06:38:05 And if there was, what makes it hard to release it as is? Feb 25 06:38:08 raster: I humbly disagree. I got datasheets of almost every chip I was able to spot inside N900 Feb 25 06:38:16 PaulFertser: first 90% of it is soc documentation Feb 25 06:38:22 and even that has errors Feb 25 06:38:22 And? Feb 25 06:38:59 the soc docs are issues by the chip vendors Feb 25 06:38:59 ...and OM as well didn't publish a complete set of chip datasheets Feb 25 06:39:06 not the oem's Feb 25 06:39:11 so u'll need that the be public Feb 25 06:39:43 as such thats an argument with another group - the same argument u make to imgtyec for open drivers etc. etc. Feb 25 06:40:23 I guess i shouldn't have asked, i still do not like and can not accept your reality :( Feb 25 06:40:42 * DocScrutinizer finds the sudden switch from night-mode to morning-mode hasn't done good effect on his mood Feb 25 06:40:43 o know u dont like it Feb 25 06:40:47 but thats reality Feb 25 06:40:55 the sooner you accept reality and work WITH it Feb 25 06:41:00 the sooner you will get something done Feb 25 06:41:04 and u'll accomplish something Feb 25 06:41:16 Me? Accomplish? muhaha Feb 25 06:41:22 until then you are the mosquito on the elephants back. Feb 25 06:41:31 that's apologetic Feb 25 06:41:36 that hide is pretty damned thick Feb 25 06:41:37 :) Feb 25 06:43:40 right now there's a big battle on the software side Feb 25 06:43:57 there are tough bits to fight over there - the imgtec's of this world and their closed drivers Feb 25 06:44:17 the oem's have literally not even the choice to be open or closed there Feb 25 06:44:26 that's the event horizon of raster's world obviously :-) Feb 25 06:45:47 making all the core - kernel, core libs, drivers, middleware, toolkit etc. open is still a battle to fight - and it'll be years. Feb 25 06:46:05 companies are not going to jump in and open everything up at once Feb 25 06:46:12 they are going to dip a toe in and start bit by bit Feb 25 06:46:17 but if I'd be interested in the question "can N900 WOL from WiFi?" and maybe write an augmented driver for the friggin chip, I'd clearly be on the wet side without schematics Feb 25 06:46:22 hw will be the last thing - if at all that opens up Feb 25 06:46:39 WOL? Feb 25 06:46:42 ooh wake on lan Feb 25 06:46:48 Wake On LAn Feb 25 06:47:07 same for accels Feb 25 06:47:09 i wonder if wol is viable for wifi when a system is off Feb 25 06:47:12 eg wpa Feb 25 06:47:16 wep - maybe Feb 25 06:47:17 same for USB hostmode Feb 25 06:47:44 WOL is a lower layer than crypto Feb 25 06:47:54 it's a very simple IP packet Feb 25 06:51:06 for accel LIS302 which is same in N900 as is in FR, I have *no* idea if I can use the IRQ options of that chip to kick SoC out of 0clock, as I simply got no schematics to tell what the wiring is Feb 25 06:52:57 USB hostmode: Nokia decided we don't need drivers that could do that. Is it just because that's been a business decision while sw-devels were short on time, or had they actually left out the power supply making N900 like neo1973? Feb 25 06:53:28 I can't tell without either disassembling and REing the device, or get the fsckng schematics Feb 25 06:55:55 and again: cost for publishing the FR schematics was roundabout 4 mandays of me reviewing the schem, and checking some NDA issues Feb 25 06:56:32 how could the ammount of work be any higher for an arbitrary mobile of same complexity? Feb 25 06:58:01 ther _is_ _no_ reasonable rationale *not* to disclose the schematics, except from "we never did that before" Feb 25 06:59:59 so please don't tell me why we don't need schem, and why we'll never achieve to get schem. And why OEMs can not do that. Feb 25 07:00:22 If you tell me "I simply don't care" that's fine for me Feb 25 07:03:25 DocScrutinizer++ Feb 25 07:16:42 btw each cloner would be catious to use officially published schem (well, any schem) as the oricinal author might have introduced errors. A good EE could even introduce errors in such a way it would neither have any impact on repair service (which usually is a good source for schematics of "closed" hw), nor on the device driver devels' understanding of the hardware. A very good EE would introduce errors that bite the cloner after they Feb 25 07:16:44 manufactured and sold 10k devices ;-P Feb 25 07:18:46 (and you don't know if *I* am a poor, good, or very good EE X-P ) Feb 25 07:20:38 so if you were to clone FR, you'd RE the physical device anyway, no matter if the schem is published or not Feb 25 07:48:19 hmm interesting discussion before morning coffee Feb 25 07:48:38 cofffeeeeee, my cup is empty Feb 25 07:48:49 and I was wondering wtf is wrong here Feb 25 07:50:46 DocScrutinizer: if it's only empty cup, than you're lucky guy Feb 25 07:51:23 well to the other nasties around here I got used by now and easily spot them :-P Feb 25 08:09:21 DocScrutinizer: u dont need schematics to know that... 0clock wouldnt work at all if the chip couldnt wake up the cpu Feb 25 08:09:30 sshing in via wifi wouldnt work... for sure Feb 25 08:09:32 but it does Feb 25 08:09:47 also a quick look at the driver code would tell you Feb 25 08:09:56 schematics wont help u with how mt kae the chip turn on wol etc. Feb 25 08:10:02 how to make Feb 25 08:10:55 raster: nope. For LIS302 you got a master at CPU and a slave at accel chip. The only way for the chip to wake SoC is via IRQ. Usual mode is poll though Feb 25 08:11:39 (WOL) I didn't say normal operation mode of N900 could use any WOL Feb 25 08:11:55 oh sorry Feb 25 08:11:57 it was the first and worst example that came to mind Feb 25 08:12:01 was thinking of wifi wol - not accel Feb 25 08:12:40 WiFi WOL was useful only if somebody had the weird idea to send the system to suspend Feb 25 08:12:56 as such if irq lines were hooked up from accel to soc - driver shoudl reflect that with ircq's being handled for the accelerometer Feb 25 08:12:58 which btw works on N810 Feb 25 08:13:02 as opposed to only having poll mode Feb 25 08:13:05 if done right Feb 25 08:13:27 that's the pont: if done right Feb 25 08:14:01 it should be Feb 25 08:14:34 open source makes sense for companies with shit engineers Feb 25 08:14:42 so they dont do things right and hope u fix it for them for free :) Feb 25 08:14:50 ack Feb 25 08:15:04 irf the company has top engineers that dot all the i's anxd cross all the t's Feb 25 08:15:08 the soruce will have everything u need Feb 25 08:15:10 either implemented Feb 25 08:15:16 but in the end that's the whole point of FOSS for a company Feb 25 08:15:23 or code being there and dsiabled for whatever production reasons needed it to be disabled Feb 25 08:15:27 and a comment as to why Feb 25 08:33:56 raster: see all the whining and bitching about meego and Nokia alegedly abandoning maemo5 and the N900 users. Why is that? Simply because everybody is aware you can't build a *free* system just as there are no FOSS drivers for some parts. Making the drivers FOSS would not even change the situation radically, as you still have to rely on all of them being bugfree and comprehensive, or Nokia promising to fix any flaw or shortcoming up to and Feb 25 08:33:58 open end in the future Feb 25 08:36:43 but the killer argument still is, was, and remains: publishing schematics is basically *free* for the company Feb 25 08:36:53 both in effort and risk Feb 25 08:37:04 chmm few kernel builds and Load average: 13.47 13.57 9.03 Feb 25 08:38:59 DocScrutinizer: all the schematics in the world wot help u - without open drivers Feb 25 08:39:01 or middleware Feb 25 08:39:14 and iuts not free Feb 25 08:39:23 they need schems they can publish to start with Feb 25 08:39:36 that dont contain ip they dont own... Feb 25 08:39:39 like any software Feb 25 08:39:39 I'd think they *got* schem Feb 25 08:39:45 parts of the schems may not belong to them Feb 25 08:40:04 so that's what I checked in 4 days for OM schem Feb 25 08:40:37 what's FOUR mandays for a company like a mobile device OEM? Feb 25 08:41:45 I guess their webmaster will invest more time into making up a nice frontpage note out of the schem publishing Feb 25 08:43:34 no including management approvals, legal approval and vetting etc. etc. Feb 25 08:43:46 (open drivers and middleware) sorry I don't get the point there Feb 25 08:44:33 of course I can write my own driver, once I got the schem Feb 25 08:44:47 and middleware... err, hum?? Feb 25 08:46:28 DocScrutinizer: no chance of writign a driver from a shem Feb 25 08:46:30 zero Feb 25 08:46:32 schem Feb 25 08:46:40 u have no idew what memory ranges are mapped to what Feb 25 08:46:41 uhuh Feb 25 08:46:46 what each memory range is and what registers they are Feb 25 08:46:48 tyhe bits in the register Feb 25 08:46:50 what they mean Feb 25 08:46:53 behavior etc. etc. Feb 25 08:46:55 zero Feb 25 08:46:58 thats not schematic Feb 25 08:47:07 thats chipset docs Feb 25 08:47:08 I got the datasheets for that shit Feb 25 08:47:11 yes Feb 25 08:47:13 thats not schem Feb 25 08:47:42 [2010-02-25 07:38:08] raster: I humbly disagree. I got datasheets of almost every chip I was able to spot inside N900 Feb 25 08:47:44 right now u donteven have the chip docs (datasheets) for a lot of things Feb 25 08:47:47 schem wont help u Feb 25 08:48:24 well then - u should have no problems Feb 25 08:48:42 you can happily make an open source battery charging logic driver then :) Feb 25 08:48:55 I have *one* problem: missing schematics Feb 25 08:52:06 raster: this type of backward conclusion won't work on me: "you're missing xy of party Z (which I actually don't), so why do you think you'll get yz from OEM" Feb 25 08:52:43 i'd say the problem isnt missing schematics. as a quick probe of memory and whats mapped where will tell you whats going on. Feb 25 08:53:19 but in the end if the driver was open - the whole thing would be moot. Feb 25 08:53:22 it's not sufficient, you should know that Feb 25 08:53:35 nokia dont want an open driver - so they're not going tohelp you out. Feb 25 08:53:44 it is sufficient Feb 25 08:53:55 i spent years writign driver-level code for my amiga Feb 25 08:53:55 that's a totally different argument Feb 25 08:54:00 i never once read a schematic Feb 25 08:54:07 they had them at the back of the manual that came with it Feb 25 08:54:15 nevr once used them Feb 25 08:54:27 why? the "datasheets" were complete and good enough Feb 25 08:54:59 nah, I realize we agree to disagree here Feb 25 08:55:07 and u didnt ahave drivers unless u worked at the very highest levesl (workbench land) if u did games u dropped right down to swizzling the fdd controller, blitter, audio chipset etc. etc. yourself Feb 25 08:55:24 didnt need a schem - i had register mappings and docs on what they did Feb 25 08:57:42 when I wrote the driver for the RF-RTC to hok up to mouseport-2, without messing with the standard functions so each usual joystick still would work, you bet I used the schematics to check I didn't fry anything Feb 25 08:58:17 frying is fun! Feb 25 08:58:18 :) Feb 25 08:59:17 on my A1000 yes, as it had decent protection for VDD on joystick ports. If you got a A500 though, you'll probably want to think twice Feb 25 09:00:58 nayway. This leads to nowhere. Gonna get me a breakfast at a very unusual time - in the morning. :-P cya Feb 25 09:42:24 * Weiss thinks it'd be nice if all driver-writing consisted of MMIO poking Feb 25 09:42:53 and if all hardware registers were read/write and gave sensible values on read Feb 25 10:03:09 Russian news: "a non-sober passer-by helped a drunk driver to beat up a road policeman" :-/ Feb 25 10:09:55 :/ Feb 25 10:10:22 HEHE Feb 25 10:17:23 In fact we rarely trust our policemen, too many criminals and assholes among them. Feb 25 10:18:40 * Weiss remembers watching a tourism TV programme about somewhere in Russia. they mentioned how, if you're caught without ID, it might cost you a hefty *bribe* (not fine..) Feb 25 10:20:30 Weiss: sure, for minor offences it's more common to pay a bribe and not fine. Feb 25 10:21:36 the bribe is usually less? Feb 25 10:22:15 how do you know that they're not going to haul you up for bribing the police, and ask for an even bigger fine (or even a bigger bribe..) Feb 25 10:25:30 (although, I suppose, how do you know that in the UK the police officer isn't just going to beat you up :S) Feb 25 10:26:09 Weiss: the bribe is less and faster, yes. And punishment for bribing is really rare anyway... Feb 25 10:36:03 anyway I heard they gave a decent sentence to that policeman shooting several people in a moskow shopping mall Feb 25 10:36:20 which seemed to amaze everyone Feb 25 10:37:03 as usually police are not under legislation for whatever they do Feb 25 10:37:21 Bruce Schneier has a few things to say about that Feb 25 10:37:48 the countries where police powers are most regulated are often the safest Feb 25 10:38:01 DocScrutinizer: yeah, life inprisonment sounded like something surprisignly fair. Feb 25 10:38:58 and screwed all the police, as they need to rethink their notion of their job Feb 25 10:39:41 "duh, no more protection / unconditional support from court?" Feb 25 10:41:23 you bett all policemen will elect putin next time Feb 25 10:42:17 if that's a direct vote at all Feb 25 10:44:13 Voting doesn't matter without freedom of mass media. And even with it often sucks. Recent ukrain elections fully confirmed that :( Feb 25 10:44:42 yep :-/ Feb 25 10:45:20 * DocScrutinizer tends to think elections never really matter Feb 25 10:46:23 some wise man said "if elections could change anything, they'd be forbidden" Feb 25 10:56:26 often attributed to Emma Goldman (google says) Feb 25 10:57:27