<wolfspraul>
the status of free software in this is probably not great, but not hopeless either I think
<wolfspraul>
there seem to be quite a few scattered projects already
<wolfspraul>
on Wi-Fi (sige also has chips that atheros and others are using), we decided there is 'too much' between the ICs SiGe offers and what you buy from Atheros, for us to undertake it
<roh>
true. but none were even on the level of sirf3 last time i checked. that was what demotivated me to think about a attempt which does have no 'proper working fw' yet
<wolfspraul>
so the second project in that direction is the 802.15.4 project wpwrak is doing, called ben-wpan
<wolfspraul>
well we work on that
<wolfspraul>
you can watch it here...
<wolfspraul>
:-)
<roh>
that ive seen. am quite interrested about the results (ben-wpan)
<wolfspraul>
great
<wolfspraul>
you should also start to become interested in the SiGe 'driver' project
<wolfspraul>
:-)
<wolfspraul>
but yeah it's a lot of work
<wolfspraul>
basically that would really replace all of what u-blox does
<wolfspraul>
and u-blox worked hard on this, for many years
<roh>
wolfspraul is there and proper working sw for the SiGe yet?
<roh>
like gllin?
<wolfspraul>
the founders came out of college with this project, and started the company
<wolfspraul>
roh: you need to ask kristianpaul
<roh>
(well.. working, not so poor like gllin)
<wolfspraul>
(he is sleeping right now I think)
<roh>
just thinking. its cool to pursue that, of course
<wolfspraul>
we are not trying to get u-blox binary blobs to work, I think that would be a lot of work and it's easier to just go straight to the source, collect the raw data samples, and decode them
<wolfspraul>
I'd say let's give it 6 months and then see where it stands.
<wolfspraul>
kristianpaul only got the SiGe EVB 2 weeks ago or so
<roh>
but i wouldnt start building such hw and doing business with a non-dual way like i explained in the beginning. simply because that would to 'openmoko repeating' business-wise
<roh>
wolfspraul i dont mean 'try use ublox binaries'. but replicate the performance
<roh>
that will take some more work than one person for half a year. in the end, thats where ublox also gets its money made from
<wpwrak>
roh: i think the processing unit is the least of your concerns. what's important is the RF side and the algorithms
<mth>
I've seen the testing that u-blox does on their software GPS... it's the most advanced automated test setup I ever saw
<mth>
and afaik they worked on it for many years, not months
<roh>
wpwrak ack. but the algorithms is the costly part
<wpwrak>
mth: and this is also the stuff they'll never open. so tracking u-blox wouldn't help us much.
<roh>
the rf side is why i'd like to see some working binarie before trusting sige
<mth>
I mean that 6 months will get you nowhere near the same level that they are
<mth>
it might get you at a useful level though; I don't know the material that well
<roh>
its a nice research project anyhow ;)
<roh>
something to thumb-up for sure.
<wpwrak>
mth: i think it would be a success already if the thing can decode anything. that would allow to validate most of the hardware. you can probably analyze rf performance without even decoding anything.
<roh>
just dont expect it works in 6-12month to be sold
<wpwrak>
mth: after that, it's all software. and we're used to software maturing with time :)
<mth>
wpwrak: yes, by all means have a go at it
<mth>
just don't expect results too soon
<wpwrak>
we'll have to ask kristianpaul about the schedule :)
<roh>
does qi-hw have access to a vector-signal-generator which can do rf up to ~1.5ghz?
<roh>
that would be helpful for whoever does the hw validation for sure.
<wpwrak>
roh: does my usrp2 count ?
<roh>
nope.
<wpwrak>
grr :)
<roh>
something with a calibration-seal and a true decibel scale ;)
<wpwrak>
calibration is indeed the weak spot ...
<wolfspraul>
mth: correct, I can totally see that they invest in a (closed of course) automated test system
<wolfspraul>
because it's a software company
<wolfspraul>
if we do this, the automated test system will be gpl as well
<wolfspraul>
I think we are all on the same page about 'realistic schedule'
<wolfspraul>
roh: trusting sige? you are already trusting it if you use (some) u-blox products
<roh>
oh? didnt know that
<roh>
thats in their favour then, of course
<wolfspraul>
the last thing I would worry about in our projet is sige. maybe if we become successful there could be pressure from other sige customers, but that's very hypothetical and I don't worry about that now.
<wolfspraul>
roh: well, strange I thought I explained it above
<wolfspraul>
sige is an established company in RF ICs
<wolfspraul>
companies like u-blox use such vendors to build their software on
<wolfspraul>
and so do we now, at least we try
<roh>
wolfspraul now i am confused.. 'such vendors' for sure. but which products of ublox use sige products, and which ones?
<wolfspraul>
I don't know for sure, but I don't think u-blox has any RF competence, and I am not sure whether they want to build it up.
<wolfspraul>
I think they try to stay out of manufacturing, and out of RF. Could be wrong about the RF or they try different things, not sure.
<wolfspraul>
another similar company like sige would be www.maxim-ic.com
<roh>
you are sure about rf competence?
<roh>
i thought they must have atleast some good people there going from what results i see
<roh>
bad hw and good sw doesnt make a good performance either
<wolfspraul>
ahh, well I talk about business model. what is your differentiator. what are you good at. what do you outsource.
<roh>
regardless if they bought the frontend or just used it right and put it on some proper layouted board
<wolfspraul>
if someone sells you something great that doesn't mean they also developed it. maybe they just marketed it, etc.
<roh>
lots of stuff one can do wrong when frequencies are above a few mhz
<wolfspraul>
last one is October 9, that must have been fully automatic already
<wolfspraul>
I noticed that the '16 minutes ago' is relative to whenever that script ran, it's not accurate relative to the person looking at this in a browser
<kyak>
guys, are there instructions about how to build the latest kernel? any patches, what else?
<kyak>
(the idea is to compile the kernel natively on Ben)
<wolfspraul>
kyak: if you want to replicate the currently running uImages, I'd say you start from the openwrt patches. otherwise you can start with kernel.org 2.6.37-rc7 and see how far you get in booting. I doubt it will come up though, not all drivers are there yet.
<kyak>
yes, i'd like to start with the latest kernel
<kyak>
the currently running i already have :)
<kyak>
so it seems like a lot of job.. porting all these openwrt patches to the latest kernel
<kyak>
including drivers
<wolfspraul>
which one are you running now?
<wolfspraul>
you don't need to port 'all' patches, I'm sure. Lars got a large part of the XBurst stuff into 2.6.36.
<wolfspraul>
just not everything, so if you try to compile a vanilla kernel you will run into the missing parts. larsc knows most about it, I am just giving you my best guess.
<kyak>
i'm running the 2.6.32.16 from openwrt, so it makes no sense to try and build it on Ben
<kyak>
ok, i got your point
<kyak>
i'm just trying to understand where to start
<kyak>
i got 2.6.37-rc7 on my desktop, and don't see anything similar to qi/xburst or whatever in make menuconfig
<kyak>
how should i select a target? there is something else in CPu family
<wolfspraul>
hmm, no I also don't know how to configure it. if I run a find . -iname '*4740*' in the 2.6.36-rc7 tree, I do find a lot of files though, so it's definitely there
<wolfspraul>
maybe under mips
<wolfspraul>
I think if you really try to get 2.6.36-rc7 to boot on your Ben, you should wait until you get some advice from Lars. otherwise you may waste a lot of time.
<wolfspraul>
I've heard numbers like "90% should be there" or so.
<wolfspraul>
but for sure you cannot just make install modules_install && reboot
<wolfspraul>
then I don't know whether there are OpenWrt-specific patches you might need, if you want to use this kernel to boot from an OpenWrt generated rootfs
<wolfspraul>
that's a separate question I also cannot answer :-)
<kyak>
ok... thanks for the hints
<kyak>
wolfspraul: btw, i noticed there is a "make yesoldconfig", which might be just what we need instead of yes "" | make oldconfig > /dev/null
<kyak>
need to try
<wolfspraul>
oh really?
<wolfspraul>
will check
<wolfspraul>
in Linux it's called silentoldconfig, so maybe I only looked for that
<kyak>
oh, yes, i'm talking about linux config, maybe that yesoldconfig is not even in openwrt..
<wpwrak>
wolfspraul: (relative date) hmm yes, that's relative to the time the index.html was generated. there's no good way to implement this correctly. i guess i'll remove it then.
<wolfspraul>
I was unable to resist a few more patent mails :-)
<B_Lizzard>
Some functions were renamed, others were removed, their functionality moved elsewhere
<B_Lizzard>
Still working on it
<kyak>
sounds likes a lot of work
<kyak>
mupdf developers could've treated their users better
<B_Lizzard>
Well, my only gripe is the lack of documentation
<tuxbrain>
wolfspraul: have you seen the fedora forbiden items links?
<wolfspraul>
I saw it, but not read it
<wolfspraul>
I'm sure they have good stuff there, so I will definitely read it
<wolfspraul>
my passion for legal stuff is limited
<wolfspraul>
life is short
<kristianpaul>
:D
<wolfspraul>
forbidden items? and you want me to read that?
<wolfspraul>
come on! :-)
<kristianpaul>
roh: not yet
<kristianpaul>
wpwrak: yup i need put a schdulle on the wiki
<kristianpaul>
schedule
<kristianpaul>
sadly this weekend i cant work on the hardware side of the board bcae i have a meeting in a town two hours away by ground,
<kristianpaul>
but i think i'll make the first contact with the data next week
<kristianpaul>
i wish early but i dont have some expensive tools at home to take some measures before break soemthing, so i asked a profesor in the local public university to let me use their labs he said yes, but i', not in town in weeks days, so i'll ask for a some free days next weeks to workon it
<wolfspraul>
kristianpaul: wow fantastic!
<wolfspraul>
say thanks to that prof from the Qi community :-)
<wolfspraul>
can you add the name of college and prof to the Community News? I think such information is interesting and valuable.
<wolfspraul>
just dump is here somewhere, I'll make it prettier later
<wolfspraul>
also mention which equipment exactly he gives you access to, one line is enough
<kristianpaul>
roh: performance is clue, i'm more worried about that part besides data acquisition in wich i work right now
<kristianpaul>
roh: that allow me to get a big dump of trusty binaty data so in can upload somwhere for other hack on
<kristianpaul>
roh: Ben is not so powerfull,, may be?? (remenber SIMD is not been explored so depth) so lets see if we stikc to off line correlation or a free correlation damenos for the ben..
<kristianpaul>
so far there is a free sofware projects that beat ubox software i think
<kristianpaul>
is called gpd-sdr (note all this is already documented on the wiki page)
<kristianpaul>
it worked with a  PRIMO board, wich uses a maxin chip, _BUT_ was proven to work to with a demotration board (that just let capture 30s of data !!) is from SiGE to similar chip
<kristianpaul>
so we're not so far
<kristianpaul>
the clue for now is data aquisition
<kristianpaul>
it may sound simple (wich in fact it is) but i just want make sure get data in right way (thats way i'd had been bugging about scopmeter equipment this days)
<kristianpaul>
also i may need a CPLD (wich i have) to implement a three stare buffer to avoid some anoyiance in the future
<kristianpaul>
what stop me? well i jus orderded the JTAG cable some days ago, i think it will take some other days to arrive
<kristianpaul>
what i'm doing now? I'll debug all data output from the SiGE EVB, (as the manual that came with it recomend)
<kristianpaul>
s/so we're not so far/so we're not so away
<kristianpaul>
the missing part is the aquisition data
<kristianpaul>
roh: all this is new for me (in some manner) so if you have recomendation i really will apreciate
<kristianpaul>
ok i must leave now (travel to work)
<kristianpaul>
see ya i some hrs
<kristianpaul>
wpwrak: i added a small schudell i hope update the items on it during the day
<kristianpaul>
schdule*
<tuxbrain>
wolfspraul: yes the topic is bit unfortunate :)Â Â but the text inside seems is real good explanation on patent topics , and reasons of why must be avoided and don't have room in a free project. we can use the less aggressive topic of wpwrak page, just saying that work on that issues is already done and written, and able to use without loosing more time on this. Agree live is short and there is a lot of things more interesting to do :), damn lawyers
<wolfspraul>
tuxbrain: I'm sure Fedora is doing good work on this, I'll read it no worries
<cfy>
hi all,i think i'm in trouble
<cfy>
i reflash the ben nanonote,i want install debian on it
<cfy>
when i finish flash,i can see success on the xterm
<cfy>
then i use the hardware reset  on the button of the device
<cfy>
but it seems does'nt work.
<cfy>
so i want to boot in usb mode
<cfy>
but  the software and the hardware boot mode do not work
<cfy>
i don't know what to do
<cfy>
can some one give me some advices?thanks
<larsc>
take out battery, disconnect usb, wait 5 seconds and repower the device
<cfy>
repower?like normal power?
<cfy>
i try it
<cfy>
can someone speak chinese,i know xiangfu can,but he is not here
<cfy>
larsc: it dosn't work
<cfy>
lsusb still not display 601a:4740
<larsc>
press the power button for a few seconds
<cfy>
i press more than 5 seconds,but happen nothing
<cfy>
maybe i should disconnect from the prower for more while
<cfy>
when i plug the battery,and connect the usb ,i can see the led bright
<cfy>
so my nn isn't broken?
<larsc>
i don't think so
<larsc>
it's virtually impossible to brick the nanonote be conventional means
<cfy>
i mean i hope the hardware will not be broken.
<cfy>
i always run openwrt on the nn
<cfy>
but i want to try debian,because i may get more softwares
<kyak>
and more unstability
<kyak>
this is not the first time people complaining about Ben being bricked after debian
<cfy>
be brick?
<kyak>
also, the more software your are talking about is generally just not suitable for Ben
<cfy>
kyak: is there any solution to save the bricked ben?
<cfy>
yes
<kyak>
as larsc said, it is hard to brick it.. so don't worry
<kyak>
just reflash
<cfy>
but i can't boot it in usb mode
<kyak>
then boot it by shortneing pins
<cfy>
oh,i will wait for xiangfu,so i can speak chinese to him:)
<cfy>
thanks
<cfy>
kyak: the hardware usb_boot?
<cfy>
i try,but failed
<cfy>
but i don't try the hardware usb boot before,maybe i do the wrong way
<zear>
afaik you can't physically brick a nanonote, as the system-on-chip allows you to switch to usb boot mode to reflash the nand even if the device is bricked
<zear>
and you can't modify/delete the system-on-chip, so it will be always there
<wpwrak>
cfy: here's a quick process: first without usb_boot:
<wpwrak>
cfy: 1) remove battery and disconnect usb.
<wpwrak>
2) wait 30 seconds.
<wpwrak>
3) press the U key and hold it down.
<wpwrak>
4) while still holding down U, connect USB.
<wpwrak>
5) wait 5 seconds, then release U.
<wpwrak>
6) lsusb on the PC should now show you the xxxx:4740
<cfy>
then done?
<cfy>
ok,i try it
<cfy>
zear: got it
<wpwrak>
then you proceed with the flashing procedure
<cfy>
ok
<wpwrak>
if this doesn't work, try it again, maybe 1-2 times. if it still doesn't work, then your u-boot is broken, too. don't worry, this can be fixed as well
<wpwrak>
but let's first see how this goes.
<cfy>
ok
<cfy>
dosn't work
<cfy>
mabye i try the hardware usb boot
<wpwrak>
okay, there it goes as follows:
<wpwrak>
1) remove battery and disconnect USB.
<wpwrak>
2) wait 30 seconds.
<wpwrak>
3) use the carbonized rubber button (the little rubber thing that came as an acessory with the ben) to short the usb_boot contacts
<wpwrak>
4) hold that button down. then connect usb, while still holding the button. release the button after 5 seconds.
<wpwrak>
5) now you should see xxxx:4740
<cfy>
ok,i try
<wpwrak>
if it doesn't work, retry. it's not easy to get a reliable contact with this rubber button.
<wpwrak>
important: make sure the system is without power for ~30 seconds before you try to start it. if you don't give it enough time, it won't reset.
<cfy>
ok,it works!!!
<kristianpaul>
:)
<wpwrak>
yeah !! congratulations ! :)
<cfy>
i notice you don't tell me to press the power button
<wpwrak>
this is your lucky day :)
<cfy>
maybe this is the problem?
<wpwrak>
"press" was "use to short" :)
<wpwrak>
ah .. no
<wpwrak>
sorry. the power button isn't used.
<cfy>
oh
<cfy>
thank you,and thank all:)
<cfy>
you save me :)
<wpwrak>
the system powers on automatically if you supply power. you need the power button only to start it if it has already had power.
<cfy>
wpwrak: i have try this hardware usb boot before,but it doesn't work,i don't know why this time it works
<cfy>
wpwrak: maybe last times,i always press the power button?
<wpwrak>
cfy: maybe, yes. or you didn't hold down the rubber button all the time. or sometimes it's just bad luck :)
<cfy>
wpwrak: oh,got it,thank you.i finally be lucky:)
<kristianpaul>
hmm Fedex is offering a courier service
<wpwrak>
kristianpaul: breaking news ! ;-)
<kristianpaul>
wpwrak: you thing is good have an customer account with Fedex?
<wpwrak>
kyak: (the trouble with debian) it's indeed interesting why people have so many problems with that. i wonder what the install instructions for debian say.
<cfy>
wpwrak yes,i retry,your instructions is better
<cfy>
From my mobile
<cfy>
Oh,the debian cannot work,maybe i should return to openwrt
<kristianpaul>
wpwrak: ah i tought, you're right documentation dont need mention POWER button in those cases
<wpwrak>
kristianpaul: the POWER button is quite misleading because it suggests an entirely different start mechanism. if you apply power and only then hold usb_boot/U, then it'll never work
<kristianpaul>
hmm thats true usb_boot is not same when hardware
<zear>
B_Lizzard, what does that need? just sdl and mixer/
<qbject>
B_Lizzard, thanks.
<B_Lizzard>
I also just finished porting the puzzles sdl thing
<methril>
jlime could be full supporteb by q-hw if they get rid of pantent stuff
<B_Lizzard>
sdl, sdl-image and sdl-mixer
<zear>
B_Lizzard, the puzzle collection?
<methril>
s/pantent/patent/
<B_Lizzard>
Yeah
<zear>
B_Lizzard, it rocks
<B_Lizzard>
Which was originally done for GP2x
<B_Lizzard>
Works like a charm
<zear>
B_Lizzard, using my source, or the gp2x one?
<B_Lizzard>
gp2x
<methril>
zear, you are doing really amazing games work on MIPS :)
<zear>
heh :P
<B_Lizzard>
For googlecode
<methril>
thsi is nice
<zear>
methril, thanks, though i am simply finalizing other people work
<zear>
the original program authors should get all the kudos
<zear>
i only recompile stuff
<methril>
zear, never said tha recompile and packaging is simple stuff
<zear>
B_Lizzard, just so you know, there is a new version of STPPC out
<zear>
if you used the old gp2x sources, they're probably outdated
<zear>
methril, well, it is
<zear>
it's just time consuming
<methril>
sometimes
<methril>
well, time is gold!!! ;)
<zear>
but it's nothing you need proper knowledge for
<B_Lizzard>
Nah, i used svn
<zear>
B_Lizzard, ah, good
<methril>
i've no much time to spent on my projects.... so.... for me.... time is gold
<zear>
:P
<zear>
methril, any stuff that could be ported to the mips?
<methril>
:)
<zear>
for 320x240, that is
<qbject>
zear: What methril said. I can barely compile hello.h
<zear>
qbject, believe or not, but i don't know much besides that ;)
<qbject>
zear: nevertheless, you can make it go. I have other skills, but that ain't one of them.
<zear>
i have a very limited c knowledge, only enough to modify the little fragments of the code responsible for input/video output, stuff you need to optimize for small devices
<methril>
well, i would like to help more. I started an effort to port RTAI (Real Time Application Interface) to MIPS
<zear>
but i'm unable to write anything from scratch on my own
<qbject>
zear: OpenTyrian, static-linked.
<methril>
zear, write from scratch is hard, but understand others code is harder (sometimes) ;)
<B_Lizzard>
Yeap
<zear>
methril, i don't really understand their code, just the little fragments, and there's a lot of trial and error
<zear>
but recently i started learning proper C knowledge
<zear>
from the very beginning
<zear>
and it already helped me a lot as i finally understand pointers and their relation with arrays and i fully understand the compiling/linking process, and the latter is really useful for porting stuff
<methril>
zear, if you need help i could give you some points (i express myself better inspanish, but i could try to do it in english)
<methril>
i've been programming in C for more than 6 years now
<zear>
methril, don't worry, i'm not english native myself
<zear>
so far i got stucked on pointers, as the course began referring to it's exercise part (while at the start they said the exercise part wasn't mandatory)
<zear>
so basically now i have to start reading the exercise part from the 1st lesson, which will take me a couple of days to get to the pointers section
<methril>
well, pointers are easy when you understand them
<methril>
later you love them
<methril>
and you have most of the sw control
<methril>
you could point pointers to hw registers, to hw memory maps,,......
<methril>
it's the most flexible tool you havein C
<zear>
i know the tehory now
<zear>
but as i said, in the examples of the code they started referring to the exercise part of the course, so i simply don't know what's going on in the code anymore. So i can't really learn pointers on the examples anymore, must read the exercise part of the course now
<zear>
*theory
<zear>
methril, i will remember about you and in case i have some problems i'll ask for help ;)
<methril>
zear, i'm here as methril or methril_work
<zear>
ok
<methril>
i'm glad to help you
<methril>
:)
<wpwrak>
methril: "you could point pointers to hw registers" ... but not to CPU registers :) (register int foo; ... &foo ...)
<methril>
wpwrak, it depends on the compiler and the cpu ;)
<methril>
i deal with compiler where you could do almost everything :)
<B_Lizzard>
A good book is a lifesaver
<wpwrak>
methril: that should be generally disallowed in ANSI C. some systems may have memory-mapped CPU registers elsewhere, of course
<B_Lizzard>
I got "C Programming: A Modern" approach which is really good
<B_Lizzard>
I read C Primer Plus and Practical C Programming before that and they weren't as good
<methril>
wpwrak, off course, but some crappy compilers are not ANSI C ;)
<wpwrak>
K&R is pretty nice, too
<B_Lizzard>
Yeah, K&R is great.
<B_Lizzard>
Dives deep very fast
<methril>
i never read C book, but i read some C++ books
<methril>
and others
<methril>
i love to read the C compiler manual
<methril>
it helps so much+
<B_Lizzard>
Well, I stopped reading for a month and I've forgotten everything :/
<wpwrak>
methril: the appendix of K&R is particularly useful for the finer points :)
<B_Lizzard>
manpages manpages manpages
<methril>
maybe i 've to take a look
<methril>
i'm still learning, but... C is my favorite languange
<methril>
C++ is the second
<methril>
and later on... Python
<wpwrak>
and then there's Zlotnick for POSIX. compact and to the point.
<B_Lizzard>
I was thinking of learning Go, that Google language...
<B_Lizzard>
But C comes first
<methril>
in my defense, i've tosaid i learned Pascal and Java first (in that order)
<B_Lizzard>
:)
<methril>
i think every language adapts is needs
<methril>
s/is/it's
<mth>
methril: learning Python would be more useful I think, since it is less like C than C++ is
<methril>
but C is really well adapted for SoC's and low level (that is where i like to live)
<methril>
well, lera
<methril>
sorry
<mth>
so you'd have a broader spectrum of languages to choose from when you want to solve some problem
<methril>
learning VErilog HDL is my point now
<methril>
to mix with C and C++
<methril>
to get the things working
<methril>
i like some C++ APIs, but i love to have the control i could have with C
<methril>
maybe someday i'm able to mix only C and python, but for now, i choose the language it adapts better for my needs :)
<methril>
i also know other languages (Perl, Shell Script, Java, Visual Basic (also .NET versions), C# , ....
<methril>
and i've my own preferences
<methril>
i also programmed in some pseudo languages
<methril>
C+Pascal+VisualBasic, Some Pseudo C, .VHL, ...... etc.
<methril>
s/VHL/VHDL
<methril>
wpwrak, did you see my lik, it's an interesting amateur radio hw
<methril>
sorry, i misunderstand the project, it's for HAM and radio amateurs
<wpwrak>
0 to 1 GHz is nice :)
<wolfspraul>
read this from Chitlesh about FEL: "We operate like a EDA/CAD group of a company but for the
<wolfspraul>
opensource hardware community."
<wolfspraul>
very nice!
<wolfspraul>
I'll keep it in mind when talking about Qi Hardware...
<wpwrak>
wolfspraul: (community news) seems that the next one will be delivered in five volumes and will be read only by scholarly researchers ;-)
<wpwrak>
wolfspraul: (i.e., there's already a lot of stuff and there are still 2.5 weeks to go)
<wolfspraul>
maybe I move some to 12/1
<wpwrak>
why not make one, say, next week ? if the rate doesn't drop, 12/1 will be even worse :)
<wpwrak>
and then you have a lot of MM to report about
<wolfspraul>
right now 11/1 is still a mess
<wolfspraul>
maybe I find an hour or two today
<kristianpaul>
methril: nice link about HPSDT, have you got one of those?
<kristianpaul>
Open Hardware License - Non-Commercial License Btw
<kristianpaul>
i not sure wich license uses the USRP(2)
<kristianpaul>
But sadlt i think those Endless Open Hardware Licenses consortiums, and blah blah blah, just focused on non-comercial afaik (i hope i'm wrong)
<kristianpaul>
just getting excuses that hardware cant be theated like sofware lets make a new licencese ! (suspicious)
<kristianpaul>
wich qi had proven is NOT true
<kristianpaul>
methril: are ahcking around SDR recently btw?
<kristianpaul>
[five volumes] lol :)
<kristianpaul>
wolfspraul: http://en.qi-hardware.com/wiki/Piwik_stats an the end of the top chart why always the line go down after current wich in think is not true, is this a piwik bug?
<kristianpaul>
That why you moved that graph for qi main page?
<wolfspraul>
kristianpaul: that's the last/current day, right?
<wolfspraul>
is that what you mean?
<kristianpaul>
yup
<wolfspraul>
because today is not over yet, the numbers for today are still growing
<kristianpaul>
but it seems falling
<kristianpaul>
just kid of funny
<wolfspraul>
if you think that doesn't 'look good', we just end the chart with yesterday
<kristianpaul>
or may be not so ./
<kristianpaul>
i think so
<kristianpaul>
just a falling pointer never look good in those graphs ;)
<wolfspraul>
ok, point taken
<wolfspraul>
but you understand why it is so, right?
<wolfspraul>
if only 10 hr of a 24h day have passed...
<kristianpaul>
ahh
<wolfspraul>
I like to 'deflate' those numbers anyway, since everybody else is inflating.
<kristianpaul>
but it evetully will change if i look it at 23:59 ?
<wolfspraul>
yes correct
<wolfspraul>
right now we take out traffic from admins
<kristianpaul>
to admins?
<wolfspraul>
but also this is 'visitors' but there is a smaller number 'unique visitors' that I will switch to
<kristianpaul>
ok
<wolfspraul>
basically I set all switches and settings I can find to come out with lower numbers
<wolfspraul>
I really don't like these inflated numbers.
<wolfspraul>
it would be easy to make the graph show 2000 visitors, trust me :-)
<kristianpaul>
;)
<wolfspraul>
about homepage
<wolfspraul>
yes, it was at the bottom of the homepage for a few weeks
<wolfspraul>
but there are 2 problems
<wolfspraul>
one is, the chart is using flash
<wolfspraul>
who likes flash?
<kristianpaul>
well i was not pointendhihg visit rates either just worried about now clarifying a bit non-complented data
<kristianpaul>
well i dont
<kristianpaul>
but is there..
<wolfspraul>
and another one is, querying the chart is creating a high load on the server
<wolfspraul>
and MediaWiki cashing doesn't help because it's all javascript driven
<kristianpaul>
mysql queries?
<wolfspraul>
well it's not just one png that is served
<wolfspraul>
it's an endless series of data points queried via javascript
<wolfspraul>
and then flash
<kristianpaul>
piwiki dint have a RRD-like out put like in munin?
<wolfspraul>
so... one sec...
<kristianpaul>
hmm
<wolfspraul>
I'm already on it
<kristianpaul>
ok
<wolfspraul>
there is a plugin to render charts as png
<kristianpaul>
go ahead :)
<wolfspraul>
but it's buggy, and it doesn't support unique visitors yet
<wolfspraul>
he said he would come out with support for unique visitors in 'a few days'
<wolfspraul>
that was 4 days ago
<kristianpaul>
ah unique your dont wan to track us :)
<wolfspraul>
so let's just give it some time
<wolfspraul>
piwik is running
<wolfspraul>
/wiki/Piwik_stats works
<wolfspraul>
I removed it from the homepage 'for now' to solve the flash and unnecessary load problem. a new plugin is 'in the making'
<wolfspraul>
I'd say give it a few weeks and a nice PNG unique visitors chart is back on the homepage.
<kristianpaul>
ok
<wolfspraul>
also I plan a rotation of small, thumbnail sized interesting pictures on the homepage
<kristianpaul>
oh
<wolfspraul>
say we put a picture in a new category, Homepage_rotation
<wolfspraul>
then we have a small band, like a newsticker, very small thumbnails
<wolfspraul>
just with some fun/interesting pics
<wolfspraul>
anyway just an idea
<wolfspraul>
no time right now...
<kristianpaul>
wpwrak: the SiGE manual said minimum output drive (7.5pF load) but the Ingenic Datasheet said that VDDIO pins have a typical 5pF, so may be, i should be so scare about getting triangled waveform, after all
<kristianpaul>
anyway i'll do the test
<kristianpaul>
s/test/debug
<wolfspraul>
kristianpaul: I start the chart with yesterday