<kyak>
i think it is needed to have at least 50Mb spare in rootfs
<wolfspraul>
let's not remove things that work
<wolfspraul>
we could move some data files to the data partition?
<wolfspraul>
but we need to be careful that we don't make upgrading and package management harder...
<wolfspraul>
things like the stardict-en dictionary
<wolfspraul>
well first step we can definitely remove the sample music file
<wolfspraul>
it's nice to see that we have so many things that actually work slowly :-)
<wolfspraul>
I'm wondering whether emacs-data can be on the data partition - not sure what it exactly is, when/how it needs to be upgraded, etc.
<wolfspraul>
at least the .ubi is only 505MB, so we're close :-)
<xiangfu>
505MB is already bigger then nanonote ubi rootfs partition. ubifs can not using full 512MB flash.
<xiangfu>
output by 'df' : ubi0:rootfs 471192 471192 0 100%
<rjeffries>
good evening 22"26 in my time zone
<kyak>
i think it's pretty tricky to install something into datafs.. because it can be mounted or not, it can be formatted or not.. how about flashing of Bens in factory? and what about users who upgrade, they could have some files..
<kyak>
..already in data partition, or maybe they have two data partitions...
<wolfspraul>
kyak: yes exactly, but we need a good solid plan/answer for those things, otherwise why do we have the data partition at all?
<rjeffries>
hello wolf
<wolfspraul>
rjeffries: we need investors to take copyleft hardware to the next level fast and aggressively
<wolfspraul>
can you help us find one?
<rjeffries>
I don't have any idea how one would pitch this to investors
<wolfspraul>
I think we should increase speed, and attack on more technical fronts in parallel :-)
<rjeffries>
you almost need to find a guy who shares your passion
<wolfspraul>
no, just a hardware investor
<rjeffries>
the guy who is behind Ubuntu is an interesting case
<wolfspraul>
any investment in any manufacturing business carries the same if not higher risks
<wolfspraul>
you have just never asked yourself what it would take to invest in some manufacturing business :-)
<wolfspraul>
what's the differentiator? what stops your competitors from undercutting you on price? etc.
<rjeffries>
oh I understand thos issues reanably well
<wolfspraul>
it takes much more guts and business swagger for Samsung to invest 30 billion USD this year into their fabs, than for anyone to invest in copyleft hardware
<rjeffries>
at least I think I do
<wolfspraul>
the issue is just to find those venturing individuals and get them onboard :-)
<wolfspraul>
some cold calling necessary I guess...
<rjeffries>
there is an existence proof sorta kinda: Adafruit, Sparkfun, Seed Studio
<rjeffries>
Adafruit has their own pick and place for example (not sure that makese sense, but they do)
<wolfspraul>
what's your point about those companies?
<wolfspraul>
one problem I see is the subsidizing of hardware by content or services companies
<wolfspraul>
at least I'm thinking about it right now
<wolfspraul>
at least they have to make it back somewhere, so we should be ready and market at those places
<rjeffries>
I assume margins in copyleft hardware will be razor thin with modest volume.
<rjeffries>
re: <wolfspraul> what's your point about those companies? [Adafruit etc] simply that they selll low cost open hardware and seem to be doing OK.
<kyak>
xiangfu: hm-hm.. thanks, i'll check it!
<kyak>
xiangfu: i didn't build mplayer with full image, quess it tries to compile in some additional libs/features, which are not explicitely disabled..
<kyak>
xiangfu: could you upload a build log for mplayer?
<xiangfu>
kyak: ok.
<kyak>
(with both stdout and stderr)
<xiangfu>
kyak:Â Â make package/mplayer/{clean,compile} V=99 > mplayer.build.log 2>&1
<wolfspraul>
I think we should tell lekernel so if you need some quick feedback at least he knows when you are there (if it's in the middle of the night I guess we cannot do much...)
<adamw_>
wolfspraul, pm2:30
<adamw_>
the most tasks are to ask them to help & verify if the h/w design is initially reasonable...
<wolfspraul>
I saw you uploaded some scope screenshots - anything unusual there?
<adamw_>
suppose that during that period i think no need to like quick support on some parameters changes request from lekernel.
<wolfspraul>
alright then...
<adamw_>
they are not enough yet...keep measuring...i want to make a sequence of timing
<adamw_>
i don't know exactly too much on DONE & INIT_B pins
<kyak>
xiangfu: mplayer has both theora and tremor (vorbis) codecs enabled. .should play the sample file just fine (though i didn't check it).. maybe you can have a look at "mplayer -ac help" and choose the codec manually
<kyak>
so you don't have a time/wish to get money, huh? :)
<Jay7>
kind of :)
<kyak>
would be great to have it on Ben, really
<Jay7>
I should finish all my work before March
<Jay7>
so.. you will have it soon enough ;)
<kyak>
good to know! )
<Jay7>
I've promised to kristianpaul to fork some kind of application launcher from kexecboot code
<wolfspraul>
kyak: no headache, if there's anything I love to care about then it's cultural differences, different characters, writing systems, etc.
<wolfspraul>
as a kid I actually ordered the 2 unicode 1.0 volumes the moment they came out and studied all those interesting characters :-)
<wolfspraul>
I even brought them to my school and told the teacher they should tech this stuff instead of the boring shit we had to learn at the time :-)
<wolfspraul>
yeah well, ...
<kyak>
Jay7: then we will have both boot menu and another application launcher :)
<wolfspraul>
I would have just gone through all those exciting scripts, one by one. every week another one...
<Jay7>
kyak: that is the point :)
<kyak>
wolfspraul: you should've devoted your life to calligraphy :)
<Jay7>
wolfspraul: I've look on chinese calligraphy about week ago :)
<Jay7>
heh
<wolfspraul>
I'm more into really exotic/mystic stuff
<wolfspraul>
the chinese reality is a bit too harsh for me every day to enjoy the beauty much
<wolfspraul>
so I like the georgian alphabet, or nepalese, or mongolian script, or even ancient stuff like maya, egyptian, etc.
<kyak>
do you have tatoos?
<wolfspraul>
ha, no
<kyak>
based on your passion for symbols, i though yo ucould :)
<roh>
wolfspraul: just read a bit of my scrollback buffer... i think we have 2 challenges in 'open hardware'. one is the purely technical, developing stuff others did already but with free licenses. the other being socially, building a community which enables us to solve the first problem.
<roh>
the second part is extremly important since 'money alone' doesnt buy one enough trust and capabilities to get things done. e.g. 'who owns what' 'who has which machines' 'who can help us get $detail solved'(in foss licenses)
<roh>
its hard getting people to write down the knowledge they have for others to 'copy and use', especially if thats what they are living from.
<roh>
wolfspraul: in the end the biggest challenge is getting this all done and working. building a ecosystem which enables us to continue working on foss stuff.
<kristianpaul>
Jay7: :-)
<wpwrak>
roh: money is never a solution :) but it can help make a solution stick.
<roh>
;)
<wpwrak>
roh: e.g., if you find someone capable and that person would have to spend the next few months full-time to solve a major problem you have (like making a ben/ya case), consider option A) that you have money or option B) that you don't.
<wpwrak>
sometimes, B) even works better than A), e.g., if people feel that involving money creates obligations or just somehow wrong at some level. but if someone has to spend a lot of time basically working for you, that's more likely to happen if you can help them in return to not have to worry about existential needs.
<mth>
wpwrak: paying volunteers can be very tricky, but fully paid work has less complications
<kyak>
"existencial needs" may vary from person to person :)
<wolfspraul>
kyak: what's special about this directory? are they all under free licenses?
<wolfspraul>
I am looking for a nice directory of only free fonts, couldn't find it yet
<roh>
wolfspraul: hard work (finding fonts)
<roh>
the one used in the case manual is free
<roh>
somehow
<wolfspraul>
I found dafont.com and fontsquirrel.com, but the free stuff is among a flood of non-free stuff
<roh>
wolfspraul: you really need to meet wetterfrosch some time
<kyak>
wolfspraul: first of all, they are free, and secondly, it would feed your calligraphic passion :)
<wolfspraul>
then there is linuxlibertine.org
<wolfspraul>
roh: did you get that missing package of acrylic sheets?
<roh>
yeah
<roh>
4 hours later
<roh>
from a veeeery sorry delivery guy
<wolfspraul>
well good then, happy ending
<roh>
yes. still.. next time i will pay somebody i know personally to fetch it. much too slow and expensive for the bad quality of service
<wolfspraul>
there is also typedia.org, but I'm not sure whether they focus on free fonts or also will get buried under thousands of non-free fonts later
<wpwrak>
kyak: (existential needs) absolutely. some people are so horrified about the idea of having to file a tax declaration that they'd rather not have any income. of course, someone must be covering their existential needs somehow. maybe their parents.
<wpwrak>
kyak: or maybe seniors. i'm actually a bit surprised that we don't see more of them in the free culture world. there ought to be a lot of excellent engineers with their juices still flowing at retirement age. who is it that virtually none of them seem to enter the free software world ? or free hardware for that matter ?
<roh>
wpwrak: true. maybe the way to get people to commit their work to opensouce is finding a way to get them live from it ;)
<kyak>
in my country, 99% of retired people don't stop working. Otherwise, they won't provide their own "existencial needs"
<kyak>
they stop working when they physically can't work, and therefor not interesting for potential employer
<wpwrak>
kyak: yeah, you have a good system :) but in many of the richer countries, a lot of people just fall off a cliff at 65. all of a sudden their job is gone. their whole life changes. some manage to give it new purpose, find something to do. a lot don't and just switch into "wait for death" mode.
<wpwrak>
roh: indeed ;-)
<kyak>
wpwrak: i know old people from USA and Germany travel a lot..
<kyak>
means that they not only can work, but they have money!
<kyak>
ideal employee :) no need to pay
<wpwrak>
kyak: exactly !
<kyak>
wpwrak: btw, how old are you? :)
<wpwrak>
kyak: 43, still a bit short of retirement age :)
<kyak>
what's the age in your country?
<kyak>
65?
<wpwrak>
kyak: i don't even know :) in argentina, i think it's about as relevant as in russia. you can retire when you have enough to last to your projected end of life. else, you better keep on working.
<kyak>
but there's still an official retirement age - and it is a very important milestone for everyone
<wpwrak>
kyak: yeah, that's the age when they lose their regular job :)
<kyak>
what do old people do in Argentina?
<wpwrak>
kyak: some earn enough money that that they can live comfortably with it. a lot get by with savings and retirement money. quite a lot have to work. it all depends a bit on the social class as well.
<kyak>
i mean, those who work - what do they do?
<kyak>
here they usually work as janitors, nurses... men can work as guards (yes)
<wpwrak>
kyak: yup, very popular.
<kyak>
if you were not kidding, it's hard to imagine those people working on some modern open source project
<wpwrak>
kyak: the guards and such, probably not. but whose who worked as engineers, probably living off their savings. where are they ?
<wpwrak>
kyak: does it not appeal to them ? are they too disconnected technologically ? do they dislike the communities ?
<kyak>
yeah, these are good reasons.. they are probably way too out of date
<kyak>
some of them are not very keen with computers
<kyak>
most of them
<wpwrak>
kyak: maybe it's still too early for sw engineers. by now, only the first cobol grunts have reached retirement age :)
<kyak>
they may be very cool respectful professors, but this is another reasons why they can also think that they are always right
<wpwrak>
kyak: hardware should be easier, though. transistors haven't changed all that much in recent decades.
<kyak>
i bet some teachers from my univercity have seen lamps...
<kyak>
wpwrak: you are right, perhaps it's too early. I wonder what would happen when i become old. Would i be out of date?
<kyak>
"what's that? ah, a teleporter.. What th ehell, i prefer driving a gasoline car!"
<kyak>
it's all changin so fast
<wpwrak>
kyak: well, one problem with hardware is failing vision. that small stuff seems to get really bad once you cross the 50. i even notice it beginning for me already when wearing lenses. luckily, i'm near-sighted, so i can still focus just fine after removing correction.
<kyak>
wpwrak: you could find yourself a young pupil. He could be your hands and eyes :) He will continue your work after your death (at least it worked like that couple of centuries ago)
<wpwrak>
kyak: that's one of the good things in the free software world. your work is likely to outlast you. i think it really does take some psychological burden off you :)
<kyak>
do you imagine how many old sites, repos, projects there will be on the Internet in 70 years? :)
<kyak>
old= abandoned, whose owbers are long dead
<kyak>
*owners
<kyak>
and nobody knows the password
<kyak>
and then intersted people start to fork it, and then they are dead, too
<kyak>
intersting picture :)
<wpwrak>
kyak: well, ihave my stuff in the linux kernel. i don't think there's much of a risk that people will read their ancient fat file systems in 50 years with code that's substantially different from what's been around for the last 15 years. likewise for initrd. it has evolved slightly and become initramfs. still the same principle. even my copyright is still valid. killall and pstree - anyone in a hurry to replace these ? of course not. s
<wpwrak>
o yes, i'm pretty sure that my work will outlast me :) i've planted my tree. and can worry about other things now ;-)
<roh>
hrhr
<kyak>
linux kernel, huh? i don't know if it still will be around in 50 years ;) Anyway, it seems like a good place to plant your tree
<wpwrak>
anyway ... let's see how my toys are doing. ah, still waiting for customs to finish the morning coffee.
<wpwrak>
argh. i just realized that the prototypes for adam will reach him just around CNY :-(
<kyak>
xiangfu: i had an idea. could you make {clean,compile} another qt app? NanoMap, for instance
<kyak>
and see if it works
<kyak>
xiangfu: for the mplayer, it's my bad, there was ac=mad in /root/.mplayer/config.. i changed it to tremor now, and ogg files should play fine.. however you could have a look at "mplayer -ac help" for a list of codecs and choose..
<kyak>
..the one appropriate for your file
<kyak>
xiangfu: if NanoMap is broken after rebuild, then this might indicate some problem in your toolchain, perhaps some libs discrepancy
<xiangfu>
kyak: will try rebuild NanoMap, (I guess it will works fine). now I start rebuild NanoMap
<kyak>
xiangfu: thanks
<xiangfu>
kyak: ok. finish rebuild NanoMap. works just fine.
<xiangfu>
kyak: there may something relate with 32bit 64 bit system,
<xiangfu>
kyak: there : move :Â Â nethack, qt4-demos, stardict-dit-en-cn, qt4-examples gcc-mips make binutiles xz, totally about 37MB.
<xiangfu>
<xiangfu> those will build as modules, not inside ubi rootfs image.
<xiangfu>
<wolfspraul> ouch, I like them
<xiangfu>
<wolfspraul> well if it's between you and your image, go ahead you have the power :-)
<xiangfu>
<wolfspraul> are the qt4-demos and qt4-examples actually working? last time I checked they weren't
<xiangfu>
<Jay7> btw, I've found 3 good looking console mp3 players - cmus, gamp and mp3blaster
<xiangfu>
* panda|x201 has quit (Ping timeout: 260 seconds)
<xiangfu>
<xiangfu> kyak: after install libggi, libgii, libaa, svgalib, giflib, fribidi, libspeex, libtheora, the mplayer working.
<xiangfu>
<xiangfu> but can not play the sample audio file in the nanonote.
<xiangfu>
<xiangfu>
* Loaded log from Tue Jan 25 17:45:24 2011
<wpwrak>
hmm, the photos i took of board flaws were useful. some of them, i couldn't spot with the naked eye and had to look them up in the picture ;-)
<wpwrak>
zrafa: (speaking ben) nice ! that was with the ben's speaker ?
<zrafa>
wpwrak: no, I plugged it to external speakers.. but no much more quality with external.. those suck. With ben speaker it has very similar sound output
<zrafa>
as video.
<zrafa>
wpwrak: I needed to use external speakers to put those near to the cam, because it was a bit far from nn (i used zoom to record)
<wpwrak>
zrafa: ah, i thought it sounded surprisingly good for an internal speaker. and yes, particularly because cameras aren't good at catching sounds from that angle. nice solution :)
<zrafa>
wpwrak: well, the sound is very similar from nn, it is surprisingly a good speaker. :)
<wolfspraul>
xiangfu: we need a better strategy for the data partition
<wolfspraul>
first I think the rootfs should always try to mount it when booting
<wolfspraul>
if it can be done in the background/in parallel with other booting, that may (or may not) speed up the entire boot process, but in any case I think we should try to auto-mount it
<wolfspraul>
can we just add it to fstab?
<xiangfu>
wolfspraul: no. can not add to fstab. I have searched a lot.
<wolfspraul>
ok then we just add a little mount script maybe?
<wolfspraul>
and where should we mount it?
<xiangfu>
how about /root/data ?
<kristianpaul>
a init script?
<xiangfu>
there is one problem is we need format the data first.
<xiangfu>
and create volume on the data partition.
<kristianpaul>
or just mount it after gmen2x..
<xiangfu>
the ubifs support multi volumes.
<xiangfu>
like we can create data1: 400MÂ Â data2: 1G
<xiangfu>
we should have default format script in /usr/bin/
<kristianpaul>
you mean format the data first when?
<xiangfu>
kristianpaul: that is the problem. we let user manually format it or what , I am not sure.
<kristianpaul>
xburst tools can help?
<wolfspraul>
xiangfu: not even more partitions
<xiangfu>
xburst tools can format it by flash an empty ubi image to nand.
<kristianpaul>
well something like a script just executed at first power up and the delete it self
<wolfspraul>
one partition is enough
<wolfspraul>
/root/data is long, I suggest /data or /home
<kristianpaul>
no data partition? one partition for all data  + rootfs you mean?
<kristianpaul>
ah i see
<wolfspraul>
no I mean not more partitions inside that partition
<kristianpaul>
oh, of course not !
<wolfspraul>
kristianpaul: yeah, but xiangfu just suggested data1 and data2 :-)
<kristianpaul>
no no !! xiangfu :-)
<kristianpaul>
just one data, to keep us sane :-)
<wolfspraul>
even the 1 data partition is a headache, but let's move as carefully as possible now
<xiangfu>
wolfspraul: no I am not suggest, I just saying it support :)
<wolfspraul>
ok, but 'no thanks'
<wolfspraul>
I think don't worry about formatting now, first we should add an auto-mount into the boot process
<wolfspraul>
then either it will be ubi-formatted out of the factory (it may be already, I forgot), or it's a one-time step one needs to do for now
<wolfspraul>
but the important thing is that the boot process should try to auto-mount, and fail gracefully if there is a problem
<xiangfu>
wolfspraul: yes.
<wolfspraul>
my vote is for /home
<wolfspraul>
hmm, or maybe not maybe the non-standard /data is better so people don't think too much about the normal use of /home
<wolfspraul>
ok, maybe /data
<wolfspraul>
kristianpaul: what do you think?
<kristianpaul>
I think it should point /home and make sure the apps ie, gmu point that too for data :-)
<kristianpaul>
point to that*
<wolfspraul>
oh sure, that's the second step
<xiangfu>
ping ping
<wolfspraul>
so /home ?
<kristianpaul>
pong xiangfu
<kristianpaul>
i wonder is ther is a /home in open wrt
<xiangfu>
kristianpaul: no
<kristianpaul>
as i checked last time it was /root
<kristianpaul>
the only home-like folder
<kristianpaul>
and if some apps already point to /root/.the_app_config, then choose root better
<wolfspraul>
is there a /home in jlime?
<wolfspraul>
maybe /data is better for jlime compatibility