<azonenberg> Any of you guys familiar with Glade?
<azonenberg> I don't mean the GDK UI editor
<azonenberg> I mean the GDS layout editor
<xiangfu> aw: Hi. I measure  the temperature of Jtag-serial FTDI chip, it's arount 40 centidegree, just for you info.
<aw> xiangfu, hi morning, me too.
<xiangfu> aw: ok. so far the jtag-serial cable works fine. :)
<rejon> hi lekernel xiangfu i need help
<rejon> what is the latest image to dl?
<rejon> i need to flash my mm1
<rejon> b4 using it tomorrow
<xiangfu> I am not send email to Mailing list. because I am not finish the ChangLog.
<rejon> what is url to flashing instructions?
<kristianpaul> script i think, there is one for that
<xiangfu> rejon: chmod +x reflash_m1.sh
<xiangfu> rejon: ./reflash_m1.sh
<xiangfu> rejon: make sure you have connect the M1 to your PC with usb-jtag-serail cable. :)
<wolfspraul> rejon: on your linux box, do you have urjtag installed? run 'jtag --help'
<wolfspraul> that's the only software you need
<wolfspraul> after that you run reflash_m1.sh and everything will be automatic
<wolfspraul> if you don't have urjtag installed, you need to compile it yourself, apt-get install urjtag will install an outdated version w/o m1 support
<rejon> oi
<rejon> so what
<rejon> fucking shit
<rejon> i don't have that jtag cable
<rejon> lekernel
<rejon> ahhhhh
<rejon> oi
<kristianpaul> hola
<kristianpaul> rejon: jtga cable is a standard usb type b cable btw
<wpwrak> jtag cable = just the usb cable or usb cable + jtag board ?
<kristianpaul> i have same doubt
<kristianpaul> i hope he carry the jtag cable hopefully to brazil :)
<wpwrak> why would he need it there ?
<kristianpaul> just in case?
<kristianpaul> better than nothing..
<wpwrak> hmm, the longer the less one should need that jtag cable, no ?
<kristianpaul> hmm 3 more libs to link..
<wpwrak> i.e., the mm1 ought to be able to update itself, without this tool. lekernel, didn't you say so a while ago ?
<kristianpaul> yes
<kristianpaul> but actually never tried that methof
<wpwrak> (of course, something very similar is true for atusb. which gained this ability just tonight ;-)
<kristianpaul> yeah is in lekernel blog
<wpwrak> or .. re-gained. i had it for months, but with the other mcu
<kristianpaul> will be nice a m1 can boostrap other m1 but thats too expensive ;)
<kristianpaul> guess will be cheaper comparing with atusb :)
<wpwrak> no, how i understood it works is that mm1 boots, then you can change the underlying flash. if for some reason end up with an unbootable device, you still have yet another copy to fall back to
<kristianpaul> yes sure
<kristianpaul> i was making a side comment when mentioned the boostrap thing ;)
<wpwrak> atusb just has two stages. one that's put there at "the factory" (well, tuxbrain hq, i assume :-) and the other that can be loaded/replaced via usb
<kristianpaul> but you need bootloader initallly or this chip have a usboot-like feature?
<wpwrak> to change the first stage, you need the atusb-pgm board (or improvide with UBB)
<wpwrak> the chip may actually come with a first-stage loader, very similar to mine. but it lacks the board-specific initialization, which is a bit unusual in this case, because the mcu doesn't have its own crystal but gets its clock from the transceiver.
<kristianpaul> interesting. well i hope i can preorder an at* soon :)
<kristianpaul> gn8
<lekernel> wpwrak: that $25 piece of trash has less chance of going out than the m1 still it buzzed all over
<lekernel> apparently the slashdot lamers are rejecting my story. fuck.
<wolfspraul> lekernel: what's the headline of your story? url?
<lekernel> slashdot.org/submission/1570602/Consumer-device-with-open-CPU-out-of-beta-soon
<xiangfu> Hi wolfspraul , something wrong with #qi-hardware again. only identify user can login
<aw> lekernel, C202~C207, you mentioned that needs 100pF~200pF.
<wolfspraul> xiangfu: ah great, that's probably a result of the qi-bot fiddling this morning
<wolfspraul> let me just kill qi-bot for a moment
<lekernel> aw_: no, keep using the 33pF part
<lekernel> aw_: those are varistors, not capacitors, so there is probably a huge tolerance on them
<lekernel> so most of the capacitance must come from proper capacitors, with a known tolerance
<lekernel> just mount smaller capacitors... like 120pF... and check picture quality
<aw_> ? i confused now...rc 2 has 470pF, i didn't know where 33pF is.
<lekernel> 33pF is the parasitic capacitance of the
<lekernel> varistor
<aw_> so you are meaning that using like 120pf instead of 470pF?
<lekernel> yes
<aw_> um...i see now.
<lekernel> aw_: you have the 33pF variant of the varistors? right?
<aw_> yes,
<lekernel> with a L in the P/N ("'L' suffix is a low capacitance and energy version")
<aw_> V9MLA0402LN
<lekernel> ok
<aw_> yes, i ordered this type (low capacitance)
<aw_> but I think that we should only replace value on C202 ~ C204 with 120pF not include C205~ C207.
<aw_> am I wrong?
<aw_> since they are balanced filters.
<aw_> or we change those 6 pcs capacitor with 120pF directly.
<lekernel> switch all 6 to 120pF
<lekernel> actually I don't know how Altera came up with that 470pF value - thinking about it again, it is weird
<lekernel> so I'd like very much to see what it's doing with a smaller value
<aw_> hmm..sort weird
<roh> where are those caps?
<aw_> roh, nearby the video-in connector.
<lekernel> roh: it's sort of a poor man's anti aliasing filter
<roh> i see
<lekernel> the ADI recommended filter is quite a mess, with at least a dozen parts
<lekernel> anyway, the adv7181 is phasing out, and we'll have to replace it at some point
<roh> quite expensive that video in..  in part cost as in work hours to get it to work properly ;)
<lekernel> and they put the filter on chip in the replacement part
<roh> nice
<wolfspraul> lekernel: tell me more about that phase-out. how long do you expect it to be available?
<wolfspraul> since when is the chip on the market roughly? I'd expect that they are available for years, even after the official EOL announcement.
<lekernel> "Product Status  Production"
<lekernel> "OF NOTE: This product is no longer recommended for new designs."
<lekernel> that's all
<lekernel> since it says "production", I guess it'd still be around for a while
<wolfspraul> they recommend 7180
<wolfspraul> datasheet from 2005
<wolfspraul> hmm, ok
<wolfspraul> yes that part may well be available for another 5 years then
<wolfspraul> or more
<roh> wolfspraul: maybe its replacement is cheaper
<lekernel> the replacement looks pin compatible too
<wolfspraul> well, there will be a solution in due time...
<lekernel> or almost
<wolfspraul> doesn't look like this has any urgency right now
<lekernel> shouldn't be a mess to switch anyway
<wolfspraul> normally parts can still be ordered for years after they disappear from the web
<wolfspraul> once you have the production figured out, these ics are very profitable
<wolfspraul> at least if that's the case (which I would assume for a product available since 2006 or so), I wouldn't be worried
<lekernel> nah... anyway using the 7180 should be a small change
<lekernel> maybe even just drop-in could work :-)
<wolfspraul> moq may go up
<wolfspraul> anyway
<wolfspraul> that seems somewhere on the radar... not urgent...
<lekernel> I think we can have the 7180 on RC4
<lekernel> RC3 has been delayed enough already
<lekernel> checking the datasheet we can certainly manually rework a rc2 or rc3 to install the 7180 for testing
<roh> lekernel: sounds like a good plan. maybe it solves some problem with the filters then too and they could be left out as np (no pcb change)
<wolfspraul> what are the pros/cons of 7180 over 7181?
<lekernel> 7180 has a better on-chip antialiasing filter than the hack we use atm
<roh> wolfspraul: 10bit instead of 9, internal filters
<lekernel> in layman terms, better picture quality
<roh> and pricing says 6-8us$ .. i think you paid much more for the 7181
<wolfspraul> yes 12 or so
<wolfspraul> well then, maybe rc4
<roh> ofcourse that says nothing about real prices...so if you still got stock, use it.
<wolfspraul> for rc3 I only have one goal now: make them asap, and make them a good end-user product
<wolfspraul> ok two goals
<wolfspraul> :-)
<wpwrak> lekernel: ($25 junk vs. mm1) yeah, it's just that you phrased your submission in a rather unfortunate way. if you want to push it now, maybe rephrase it to sound like a real event, and post it to some other site, e.g., dangerous prototypes.
<lekernel> yeah if it's not posted on Thursday I'll resubmit it with "the project announced ..." and shorted it a bit, maybe
<lekernel> shorten
<lekernel> what a bunch of lamers, still.
<wpwrak> lekernel: i'm not sure it's a good idea to "hammer" slashdot
<lekernel> fuck them
<wpwrak> lekernel: relax ;-)
<lekernel> don't hammer slashdot => nothing gets posted
<lekernel> hammer slashdot => there is some chance some articles make it through, worst case nothing gets posted
<lekernel> so ...
<wpwrak> lekernel: we already have one event, rejon's conference. they'll probably issue some news statements or get covered too. make sure mm1 gets featured there prominently.
<wpwrak> lekernel: (hammer) or they remember you for the hammering not the greatness of the project and just ignore you because you're annoying
<wolfspraul> wait, please think a bit how this works. slashdot editors wade through hundreds of posts, and randomly pick the ones they like, catching their eyes, etc.
<wolfspraul> that largely depends on the headline
<wolfspraul> you sent a good URL to me the other day
<wolfspraul> you can easily resubmit your story
<wolfspraul> I doubt they will ban you for spam or so (I could be wrong of course)
<wolfspraul> I think the reason David's VGA story was taken is because the headline was great
<wolfspraul> microsd, VGA? abuse?
<wolfspraul> :-)
<wolfspraul> this also largely depends on timing, i.e. what other stuff is around you at the time they look at your post
<wolfspraul> so just don't take it personal
<wpwrak> (headline) yeah, "out of beta soon" somehow doesn't convey a sense of urgency
<wolfspraul> it's just a little text
<wpwrak> something else to consider: what day of the week is actually best for such a submission ?
<wolfspraul> the 25 USD story was written very well, and the guy has credibility thanks to his Elite game
<wolfspraul> hey - I must have spent at least several weeks of my life full-time playing Elite!
<wolfspraul> I forgot the computer that was on, maybe a Commodore or Schneider CPC :-)
<wolfspraul> or even an early Intel 8088? something like that
<wolfspraul> so that's a great hook. the editor may have played Elite himself, and likes the story simply for the fact that the Elite guy is doing this thing now...
<wpwrak> ah, you had a CPC too ;-)
<wolfspraul> no but my best friend did
<wolfspraul> CPC 464
<wolfspraul> when Bill Ray (editor at theregister) took the NanoNote launch story, I asked him what caught his eye in my mail to theregister
<wolfspraul> he said it was the number 3000. I had somehow written that I had to sell 3000 of this thing (NanoNote).
<wpwrak> i had that one too. nice machine. nice upgrade from a vic-20. and the last "home computer" before my first PC :)
<wolfspraul> and he foudn this interesting, because he thought it was more like 300,000
<wolfspraul> that's why he took the story
<wpwrak> lekernel: the reg would be a rather nice place. but harder than slashdot. you need something really catchy.
<wolfspraul> it's so difficult to write something that someone you totally don't know finds interesting right at the moment when they glance at it
<wolfspraul> my bottom line: keep trying, don't hesitate to write a lot of 'tip us' posts
<wolfspraul> that's still in their best interest - they are just picking from stories they think their readers will like
<wolfspraul> unfortunately we all have a language problem because we are not native speakers
<wpwrak> wolfspraul: for the reg, i think "schadenfreude", messing with expectations, or implicit importance (i.e., the regular news) ought to work
<wolfspraul> so most likely the editor needs to go over our text and remove the language issues
<wolfspraul> wpwrak: I basically have a go from thereg that they will take the m1 launch story.
<wpwrak> some editors may think they own the language anyway, so they'll do this with gusto ;-)
<wolfspraul> but I can only do this once I have the rc3 units in stock, otherwise we blow this through and cannot sell and it's a disaster.
<wpwrak> okay, let's keep the reg then. very good to have a foot in the door already.
<wolfspraul> well, I appreciate their editing cleanup.
<wolfspraul> I wrote this column for heise once (in German), and I really went over the text 10 times to clean it up.
<wolfspraul> but when they published it, they still edited at least 10 places in a 2 page article, and all of those were really bad in hindsight!
<wpwrak> (10 times) makes sense :) that's about the number of iterations you need for about any text before it's really clean
<wpwrak> oh ;-)
<wolfspraul> it's amazing what quality language can or cannot have, and if someone goes over it what kind of thing they notice, especially if they are trained
<wolfspraul> well that's their job, they should be really good at this, and they are
<wolfspraul> wpwrak: yes sure, thereg is all I can really count on for the m1 launch.
<wolfspraul> on the tech side
<wolfspraul> but Sebastien trying slashdot now is great, for sure
<wolfspraul> just don't take the reaction personal. open more accounts, submit more stories, have some fun at writing hip and interesting headlines.
<wpwrak> yeah, placing at /. is good for sure. i would just try to avoid sounding too desperate. also, there's no rush at the moment anyway. you have no units to sell, you probably don't want to invite people to come offering nothing more than opinions, you don't seem to be in an unusually bad need for investment these days, etc.
<wolfspraul> yes but you can write stories with our without anything going on in reality
<wolfspraul> in fact if we had a real campaign, we would have a full series of stories planned
<wolfspraul> including things like "Stealth project working on CPU revolution", and god knows what
<wpwrak> actually, do we have numbers for how many people looked at mm1 from the /. submission alone ?
<wpwrak> ("make" news) sure. i'm just saying that not having it in the news right now doesn't hurt so much. of course, there's value in "spreading the word" per se.
<lekernel> hi rejon
<rejon> lekernel hi
<rejon> i'm behind
<rejon> have to delay our this press for a day
<lekernel> so, did you manage to flash your m1 in the end?
<rejon> not yet
<rejon> doing this morning
<lekernel> what version do you have now?
<rejon> let me get back to you in an hour
<rejon> in transit
<wolfspraul> rejon: no worry about delay
<wolfspraul> we can do it on thursday to cap the week
<wolfspraul> the important thing imho is to have real juicy facts and emotions out of Montreal
<wolfspraul> otherwise no release (it's too thin then)
<rejon> yes
<wolfspraul> so you are the front man now
<wpwrak> rejon: have you thought of posting a brief mention of LGM to the qi-hw list ? something along the lines of "it's happening NOW; i'll be there with an MM1; i plan to do this or that with the MM1". plus a link to http://libregraphicsmeeting.org/2011/
<rejon> i should
<rejon> i'm juggling a few releases this week, including lgm itself
<wolfspraul> ah yes, see my mail
<wolfspraul> :-)
<wolfspraul> I couldn't find a high-res lgm 2011 logo
<wolfspraul> the svg links pointed to 2010
<wolfspraul> the only one I found was 250x279 pixels
<wolfspraul> :-)
<wpwrak> rejon: i've tried to get wolfgang to push out conference announcements but he's more stubborn than any mule ;-)
<wolfspraul> I'm probably not the only one...
<rejon> aha
<wpwrak> rejon: i mean an announcement that the thing is happening, for people who are in the area and may want to drop in
<wpwrak> rejon: not a post-mortem
<wolfspraul> because I cannot write a text if I don't really know what's going on, or rather I won't write such a text, as it would be complete fantasy
<wolfspraul> I have never been to lgm, never to montreal, once to Canada (vancouver)
<wolfspraul> not good
<wolfspraul> rejon: yes the links there are bad
<rejon> aha
<wolfspraul> just click on it yourself
<wolfspraul> is pointing to 2010!
<wpwrak> rejon: you need to explain the MM1 angle. that's what mainly makes it relevant for out community. LGM is secondary.
<wolfspraul> Brussels :-)
<wolfspraul> pdf, also 2010 brussels
<wolfspraul> just fyi
<wpwrak> rejon: (explain) and you should do this in a mail to the qi-hw list :) well, unless all that there's to see about the MM1 at LGM has already happened.
<wolfspraul> wpwrak: I'm glad rejon is reading this, he will have 100 different things going on in parallel...
<wolfspraul> that's why I keep saying we have to be patient now and wait and see what snippets we hear from the front line
<wpwrak> wolfspraul: (I have never been to lgm, never to montreal, once to Canada (vancouver)) i've been a bunch of times to canada and i can assure you montreal exists ;-))
<wolfspraul> if there are any usable snippets, I have a draft I can fill around them
<wolfspraul> wpwrak: yes but come on. you know what I mean. I cannot write about the experience from the distance, that's a joke.
<wolfspraul> I don't even know how many people go there. 20? 500? 5000?
<wolfspraul> no really, I cannot
<wpwrak> wolfspraul: nobody expects you to write about the experience
<wpwrak> wolfspraul: no, you got that all wrong. all you have to say is that there is this conference and what qi-hw-related thing(s) will happen there.
<wpwrak> wolfspraul: then people can make their own decision if they want to go
<rejon> fixed
<wpwrak> wolfspraul: maybe the find LGM utterly uninteresting but want to see an MM1. there's also always the risk that they can't see it for some reason. people going to conferences are usually perfectly aware that there can be good and bad surprises.
<wolfspraul> perfect!
<wolfspraul> rejon: svg too?
<wolfspraul> if the pdf is postscript I can render a high-res png from that... very good.
<wolfspraul> I meant if the pdf is made out of vectors not pixels... (checking)
<wpwrak> wolfspraul: e.g., if you go to amsterdam for that flickernoise workshop and lekernel's plane was delayed and he's missing his talk. you would be disappointed but you wouldn't blame the person telling you about the event for it.
<wpwrak> lekernel: i think you better take care of announcing your two events on qi-hw yourself as well :)
<wolfspraul> if lekernel writes nice blog posts somewhere, even if short, that's material that can be reused
<wolfspraul> I think that's most efficient
<wolfspraul> if he cannot do that, we need to wait a little until more people have a really good feel for m1 and will do it instead, Jon is close, and I'm also getting there. same for kristianpaul and Yann and others
<wpwrak> wolfspraul: all i'm talking about is telling our community where there are events they specifically may find interesting.
<wpwrak> wolfspraul: conference reports and all this are a completely different story
<lekernel> Fallenou: did you find your lm32-rtems-gcc macports sources again?
<lekernel> wpwrak: I announced them to more specific lists already
<wpwrak> wolfspraul: the reason why i'm pestering you with this is because you're the principal spokesperson for qi-hw and the main channel for any kinds of announcements. so it should be part of your role to process such things. also, people may be too shy to announce some minor event or a minor role at an event (e.g., kristianpaul), but if you bundle the announcement with other events, then it's quite okay to have some minor items in there. so
<wpwrak> the aggregation also helps to properly bring minor events to the audience.
<wpwrak> lekernel: do you consider the events too specific for the broader qi-hw list ?
<Fallenou> lekernel: forgot to have a look sorry
<lekernel> maybe... it's more VJ specific stuff now
<lekernel> wpwrak: if you want you can copy and paste the nimk.nl announcement into the list :)
<wpwrak> lekernel: c'mon. you of all people can speak for yourself ;-))
<lekernel> also my feeling is that mailing about events on the project list themselves has little effect... most people are too far and won't come
<lekernel> so it's mostly noise
<lekernel> I did it in the beginning, now I'm a bit tired to be honest
<wolfspraul> lekernel: I agree. the most important is first hand reporting. it doesn't matter where you do that, it can be picked up from there.
<wpwrak> lekernel: (noise) yes, that's usually a problem. that's why an aggregator is useful. one mail per minor event is noise to 99% of all people. one mail with ten events is relevant news to 10%, which is probably better than many other posts. and that's why i'm pestering wolfgang ;-)
<lekernel> if the event has enough outreach to be worth going to, people would hear about it through other channels
<wolfspraul> cool it's on the twitter.com/milkymistvj
<wolfspraul> I will start to watch that from now on
<wolfspraul> and if there is something there, I will pick it up. we can also connect twitter.com/qihardware then it goes right through there
<wpwrak> lekernel: there are lots of events. not everybody is scanning all the news channels where they could possibly find out. e.g., people may be intested in local events but they may not actively track local news sources specializing in the area.
<wolfspraul> wpwrak: there it is http://twitter.com/milkymistvj
<wpwrak> lekernel: also, they may come to the event specifically because there's one item they want to see. they may find lots of other interesting items there, but there will typically be a small number of things they expect to find that justifies attending. (unless they just want someplace to hang out; happens too)
<wpwrak> lekernel: e.g., if some local free software group has its weekly meeting, i wouldn't care about these news and i probably wouldn't bother tracking any channels where such news would appear. but if they had a meeting and you'd attend and demo mm1, then it would all of a sudden become very relevant and interesting.
<wpwrak> lekernel: also, people who are in the area of an event may spread the word via channels you don't even know.
<lekernel> the only channels that really work are the big ones. theregister, slashdot, engadget, ...
<wolfspraul> agreed
<wolfspraul> 100%
<wolfspraul> most people are only 'syndicating' (=copying)
<lekernel> you could spend a life manually scanning the bottom of the pyramid with less results that what that '$25 computer' guy did in two days
<wolfspraul> I am pretty sure about that since I do this kind of thing for the NanoNote since February 2010
<wolfspraul> non-English publications are totally on the edge - nobody will pick them up
<wpwrak> the big channels only announce big conferences. and they often don't announce much about the program. maybe one or two highlights. like steve job doing the opening keynote.
<wolfspraul> and on the English ones, you need to focus on the top of the pyramid, then it will be picked up all over
<lestat> lekernel: thanks je mate ça
<wolfspraul> I'll add it to the community news. good changelogs are great.
<wolfspraul> I'm not even sure about slashdot being a good place for syndication anymore
<wolfspraul> from my experience, it's engadget and/or theregister
<wolfspraul> in tech
<wolfspraul> don't know about news.com
<wolfspraul> and then of course the big reviewers like Mossberg, Pogue (for product reviews)
<wolfspraul> was the microsd/vga story picked up in a lot of places? let me check the qi stats...
<wpwrak> (non-en) i think heise is also good. broad coverage and it seems that germans tend to be more interested in participating in free culture things than people in many other countries.
<wpwrak> (ubb-vga) dangerousprototypes.com and electronics-lab.com
<wolfspraul> yes, just saw it
<wolfspraul> well slashdot brought 8000+ viewers, dangerousprototypes 80
<wolfspraul> heise will only have reach inside Germany
<wolfspraul> a little bit in Poland
<wolfspraul> and a tiny tiny little bit in the english/global world via h.open
<wpwrak> yeah :) of course, the active interests of the average DP viewer are probably a hundred times closer to what we're doing than those of the average /. user :)
<wolfspraul> this one http://www.h-online.com/
<wolfspraul> but it has zero power
<wolfspraul> Germans are indeed interested in free culture, so are French, people in the Netherlands, etc. but it is very hard to reach them all effectively
<wolfspraul> see how long we struggle with one well written text
<wolfspraul> so the best way to reach them is via engadget :-)
<wpwrak> still, germany is big. and lots of people who seem to like to get involved. e.g., in openmoko .de was just as strong if not stronger than .us (of course, it helped that much of the workforce came from there, and with ccc connections)
<wolfspraul> because local blogs in Germany, Poland, Czech, France, and so on will all pick this up and neatly translate it
<wpwrak> yeah
<wolfspraul> xiangfu: how much did the boot time increase?
<wolfspraul> why? how to bring it down?
<xiangfu> ~35 secs
<xiangfu> the DHCP, and 'Simple Mode' needs a lot time.
<xiangfu> compile a lot patches I guess iin Simple Mode.
<lekernel> for DHCP you can solve it by using MDIO and checking that the cable is there
<lekernel> I posted two issues regarding those, you can have a look at them
<wolfspraul> we compile them all before entering rendering?
<lekernel> but first please finish that input.c ...
<wolfspraul> what is this 'compile' anyway? can we preload compiled versions?
<lekernel> yes but we'd need a cache and then keep that cache in sync
<wolfspraul> then only recompile them if the date of the .fnp is newer than the cached compiled version
<lekernel> there's no RTC so your date thing won't work
<xiangfu> without DHCP and Simple Mode it's needs ~9 sends
<wolfspraul> ok, then a simple hash, either inside the compiled file or maybe the filename or so?
<wolfspraul> well anyway
<wolfspraul> I got it already
<lekernel> maybe hash would work, yes
<lekernel> in either case this needs some work
<wolfspraul> xiangfu: at some point we should add this precompiled thing
<wolfspraul> 35 seconds is bad, but no disaster. I hope this is temporary though. at least it's still usable.
<kristianpaul> not bad (brochure)
<mwalle> good evening
<kristianpaul> evening
<lekernel> hi mwalle