2011-10-06 00:32 <wolfspraul> Steve Jobs has died
2011-10-06 00:32 <kristianpaul> ?
2011-10-06 00:32 <wolfspraul> too early, he will be missed
2011-10-06 00:34 <kristianpaul> phr: aboyt moxie, sure it have, just another intersting project, free of patents? well it seems, also all gcc uclinuc and even rtems port, is amazing how this guy work
2011-10-06 00:34 <kristianpaul> man ..
2011-10-06 00:36 <kristianpaul> wolfspraul: indeed
2011-10-06 00:46 <wolfspraul> kristianpaul: http://en.wikiquote.org/wiki/Steve_Jobs
2011-10-06 00:46 <wolfspraul> when the naysayers get a little too loud, read this as source of inspiration :-)
2011-10-06 01:07 <kristianpaul> Okay and confirmed I'll give a quick intro workshop about writing patches for milkymist the 27th of this month, at the Universidad ICESI, programa Diseño de Medios Interactivos
2011-10-06 01:07 <wolfspraul> nice
2011-10-06 01:08 <kristianpaul> And hopefully but not confirmed yet in a national meeting about free software called JSL, that will be placed San Juan de Pasto, Nariño, thanks to "La Guardia de TUX" LUG.
2011-10-06 01:09 <wolfspraul> just looked it up - Universidad ICESI is in Cali, Colombia http://en.wikipedia.org/wiki/Universidad_Icesi
2011-10-06 01:09 <kristianpaul> yes that one
2011-10-06 01:09 <kristianpaul> he, yes forgot to mention the place :)
2011-10-06 01:10 <wolfspraul> they have ca. 5000 students, maybe you can try to recruit ca. 100 full-time hackers for Milkymist? :-)
2011-10-06 01:10 <kristianpaul> I alredy asked then to buy a M1 the workshop :)
2011-10-06 01:10 <kristianpaul> I doubt the do it, but i tried :)
2011-10-06 01:11 <kristianpaul> (recruit) nah, i'm ont such us person, i mean, i even what to mention tecnical detailts in the work shop
2011-10-06 01:12 <kristianpaul> s/what/whan
2011-10-06 01:12 <kristianpaul> for developers i have a nother plan, but first need to met more hardware hackers, wich are not afraid of modifing the milkymist libc by then self ;)
2011-10-06 01:13 <kristianpaul> (starting with me :))
2011-10-06 01:13 <wolfspraul> sure sure, I was kidding
2011-10-06 01:13 <wolfspraul> it's great that you go there. do you need any marketing material like stickers or brochures?
2011-10-06 01:13 <kristianpaul> Actually cucuta is a nice place for that, but man almost 800km away from here.
2011-10-06 01:13 <wolfspraul> the best I have right now are the Milkymist stickers
2011-10-06 01:14 <kristianpaul> I need to write then again, those guys failed to buy some nanonotes from me, but i dont see other groups more itnerested than then so far
2011-10-06 01:14 <wolfspraul> great design thanks to Christopher Adams, and very good printing quality thanks to Yi Zhang
2011-10-06 01:14 <kristianpaul> ah yes !!
2011-10-06 01:14 <kristianpaul> I forgot i need marketing
2011-10-06 01:14 <wolfspraul> then we have brochures
2011-10-06 01:14 <wolfspraul> but they are heavy
2011-10-06 01:14 <kristianpaul> I have zero brochures aout MM
2011-10-06 01:14 <wolfspraul> the NanoNote and Sharism stickers are far worse technically than the new Milkymist stickers
2011-10-06 01:14 <wolfspraul> in design and print quality
2011-10-06 01:15 <kristianpaul> may be 20 stickers.. and still 200 nanonote flyers from cparty
2011-10-06 01:15 <wolfspraul> maybe I send you 300 Milkymist stickers, 30 brochures?
2011-10-06 01:15 <wolfspraul> ah great, yes
2011-10-06 01:15 <kristianpaul> and some from nanonote, yes please
2011-10-06 01:15 <wolfspraul> do you need more nanonote or sharism stickers?
2011-10-06 01:15 <kristianpaul> wpan had stricker ? ;)
2011-10-06 01:15 <kristianpaul> no, i first need run out of this one i had
2011-10-06 01:16 <kristianpaul> dont you have a Milkymist flyer?
2011-10-06 01:16 <wolfspraul> only brochure
2011-10-06 01:16 <wolfspraul> ah wait
2011-10-06 01:16 <kristianpaul> something i can give away no matter if end on the road :)
2011-10-06 01:16 <wolfspraul> Christophe made this 1-page flyer for the OHS Summit
2011-10-06 01:16 <kristianpaul> Oh, i missed that
2011-10-06 01:16 <wolfspraul> ok, so what do you need now?
2011-10-06 01:16 <wolfspraul> 300 milkymist stickers, 30 brochures?
2011-10-06 01:16 <wolfspraul> those are 4-page brochures
2011-10-06 01:17 <wolfspraul> then nanonote & sharism stickers?
2011-10-06 01:17 <wolfspraul> how many?
2011-10-06 01:17 <kristianpaul> 200 should be okay
2011-10-06 01:17 <wolfspraul> 200 milkymist stickers?
2011-10-06 01:17 <wolfspraul> here is the 1-page flyer http://en.qi-hardware.com/wiki/File:Christophe_M1_flyer.pdf
2011-10-06 01:18 <kristianpaul> no no, nanonote too
2011-10-06 01:18 <kristianpaul> what you already ofered is okay
2011-10-06 01:18 <kristianpaul> i save brochures for REALLY interested people, but i still miss a one page milkymist flyer
2011-10-06 01:19 <kristianpaul> that one from OHS looks nice, i think i can translate i later and print some 100 in Black and White
2011-10-06 01:19 <kristianpaul> bah no translate, to lazy anyway you need good enblgihs to get in to milkymist :)
2011-10-06 01:19 <kristianpaul> s/enblgihs/english
2011-10-06 01:20 <wolfspraul> can we make a complete list? :-)
2011-10-06 01:20 <wolfspraul> 1. 300 milkymist stickers
2011-10-06 01:20 <wolfspraul> 2. 30 milkymist 4-page brochures
2011-10-06 01:20 <wolfspraul> 3. 200 nanonote stickers
2011-10-06 01:20 <wolfspraul> 4. 200 sharism stickers
2011-10-06 01:20 <wolfspraul> like that?
2011-10-06 01:20 <kristianpaul> yes, sure
2011-10-06 01:21 <wolfspraul> great
2011-10-06 01:23 <wolfspraul> kristianpaul: can we get the moxie guy to give a better intro of his project?
2011-10-06 01:23 <wolfspraul> he seems to be aware of Milkymist, so I'm curious to understand how we can work together, what the overlap is, etc.
2011-10-06 01:23 <wolfspraul> what is special about moxie? why did he start the project? why not participate in milkymist? can he reuse anything from milkymist? can we help him in any way? and so on
2011-10-06 01:27 <kristianpaul> sure, i'm interested to read more from him too, (antgreen nick at milkymist), last i know he had a child, and have a was a phone call away from gcc developers ;)
2011-10-06 01:28 <kristianpaul> s/had/have
2011-10-06 01:29 <kristianpaul> also moxie cpu is in early stage,
2011-10-06 01:29 <kristianpaul> but he is talking about a muskoka soc too wich is cool
2011-10-06 01:32 <wolfspraul> which one? muskoka?
2011-10-06 01:32 <kristianpaul> yes
2011-10-06 01:32 <wolfspraul> I'm just trying to understand the motivations, goals, how to collaborate and share, etc.
2011-10-06 01:32 <kristianpaul> sure
2011-10-06 01:35 <kristianpaul> ha, i already asked him, but few information for what you asked me 
2011-10-06 01:35 <kristianpaul> check this http://paste.debian.net/134476/
2011-10-06 01:36 <kristianpaul> yes, i think i got interested because this http://stackoverflow.com/questions/4817950/what-is-the-smallest-simplest-cpu-that-gcc-can-compile-for
2011-10-06 01:37 <kristianpaul> bbl, dinner :)
2011-10-06 01:45 <wolfspraul> kristianpaul: interesting, that's a start [pastebin chat]
2011-10-06 01:58 <wpwrak> (pastebin) that looks old. when was that ?
2011-10-06 02:00 <kristianpaul> wpwrak: jue sep 16 2010
2011-10-06 02:00 <wpwrak> (jobs) hmm, already. a man of many facets to be sure.
2011-10-06 02:00 <wpwrak> (2010) ah, that explains it :)
2011-10-06 02:02 <wolfspraul> didn't look old to me, which shows my confusion
2011-10-06 02:02 <wolfspraul> which parts of that chat are (obviously I guess) outdated now?
2011-10-06 02:03 <wpwrak> sebastien's dismissal of linux and his statement that rtems "kinda works"
2011-10-06 02:04 <kristianpaul> hahah
2011-10-06 02:04 <kristianpaul> :)
2011-10-06 02:05 <wpwrak> i think he's a lot more relaxed about linux vs. rtems now. also because a working linux will get rid of a gazillion of driver issues :)
2011-10-06 02:05 <kristianpaul> so, you'll like this https://github.com/atgreen/linux-moxie
2011-10-06 02:05 <wpwrak> and of course, there have been setbacks with rtems as well. although nothing too nasty so far.
2011-10-06 02:06 <wpwrak> kristianpaul: nice :)
2011-10-06 02:06 <wpwrak> how does that moxie compare to lm32 performance-wise ?
2011-10-06 02:17 <wolfspraul> gee, I just realize we have to work on a world-class update process for M1
2011-10-06 02:17 <wolfspraul> it has been almost 3 months since the last release, and I don't even think we have a documented testing and release process in place, to avoid bricking units accidentally
2011-10-06 02:18 <kristianpaul> Btw, i will flash my M1 soon, and i was wondering if tha block flash thing will then avoid me to save patches?
2011-10-06 02:18 <wolfspraul> we need a strong release process with proper testing that avoids releasing regressions
2011-10-06 02:18 <wolfspraul> we need to communicate how people add features to a wishlist in one central place
2011-10-06 02:19 <wolfspraul> we need to communicate the changelog back to existing and potential future customers well
2011-10-06 02:19 <kristianpaul> uservoice was a wishlist, dunno if still there
2011-10-06 02:19 <wolfspraul> we need to come out with updates at most every 2 months or so, to keep the flow and to avoid discouraging people from filing bugs and feature requests (because it takes too long for them to be realized anyway)
2011-10-06 02:20 <wolfspraul> a lot of good work :-)
2011-10-06 02:22 <wpwrak> yeah, indeed. you need to clone xiangfu a few times :)
2011-10-06 02:52 <kristianpaul> "When you're young, you look at television and think, There's a conspiracy. The networks have conspired to dumb us down" lol
2011-10-06 02:53 <kristianpaul> oh man what follows.. 
2011-10-06 03:10 <kristianpaul> so the PDA is back, and now speaks, cool :)
2011-10-06 03:17 <kristianpaul> viric: that remind, can use the nanonote spearker/buzzer to remenber to do something?
2011-10-06 04:00 <mac> I am thinking of getting a nanonote and have looked unsuccessfully for estimates of battery life.  I am currently a Zaurus user and would like more than the one day life I get now.  How does the nanonote do?
2011-10-06 05:26 <viric> kristianpaul: I don't use the nn for that, though. It could work.
2011-10-06 05:41 <LunaVorax> Oh well
2011-10-06 05:41 <LunaVorax> A good thing I'm not religious, otherwise I would feel guilty for yesterday
2011-10-06 07:01 <viric> kristianpaul: ah, you mean the program 'remind'?
2011-10-06 07:05 <viric> I'm not that an expert in 'remind'
2011-10-06 07:08 <wpwrak> viric: forgetful ? :)
2011-10-06 07:08 <viric> About that, I'm an expert, yes
2011-10-06 09:29 <qi-bot> The build was successfull, see images here: http://fidelio.qi-hardware.com/~xiangfu/compile-log/openwrt-xburst.full_system-10052011-0521/
2011-10-06 09:58 <antoniodariush> Hi I am trying to test libunistring on nanonote but I get segfaults so there is something wrong with libunistring
2011-10-06 09:59 <antoniodariush> this is the program I use to test libunistring  http://pastebin.com/KqnDyvTL
2011-10-06 09:59 <antoniodariush> after cross compile this I get segfaults
2011-10-06 10:00 <antoniodariush> do you think this is related to libunistring or something else?
2011-10-06 11:08 <kristianpaul> viric: yes remind
2011-10-06 11:38 <viric> antoniodariush: *(strlen(s) + 1)   does this help?
2011-10-06 11:40 <antoniodariush> viric, the strange thing is that the program runs fine on my host machine
2011-10-06 11:50 <kyak> there is not a single program on Ben that uses libunistring; so it can't be proved whether the library segfaults or not
2011-10-06 11:50 <antoniodariush> that's why I wanted to test it 
2011-10-06 11:52 <antoniodariush> I want to compile guile 2.0 which depends on libunistring
2011-10-06 11:53 <antoniodariush> but guile 2.0 fails and I guess is coming from libunistring 
2011-10-06 11:58 <kyak> sounds like a job for gdb.. or strace at least
2011-10-06 12:01 <kyak> btw, do you have the latest image?
2011-10-06 12:02 <viric> antoniodariush: well, that's not that strange
2011-10-06 12:04 <antoniodariush> kyak, I am building my own image with the latest feeds
2011-10-06 12:33 <wpwrak> antoniodariush: maybe try running it under valgrind on the host. that will probably tell you about the insufficient allocation
2011-10-06 13:20 <antoniodariush> wpwrak, ok i'll have a look 
2011-10-06 13:29 <antoniodariush> why is libunistring depending libiconv instead of libiconv-full ? http://projects.qi-hardware.com/index.php/p/openwrt-packages/source/tree/master/libunistring/Makefile
2011-10-06 13:36 <wolfspraul> mac: you had a question earlier but nobody was fast enough to answer I think :-)
2011-10-06 13:38 <wolfspraul> I think it was about battery lifetime? (checking backlogs now)
2011-10-06 13:38 <mac> wolfspraul:  Yes, I was just wondering about typical battery lifetime.
2011-10-06 13:38 <wolfspraul> the NanoNote comes with a 850mAh battery by default
2011-10-06 13:39 <mac> Do you usually get a couple of days work out of that?
2011-10-06 13:39 <wolfspraul> that will give you between 4 and 15 hours, that seems to be the consensus
2011-10-06 13:39 <wolfspraul> 15 hours for music playing with screen off
2011-10-06 13:39 <wolfspraul> backlight needs a lot of power
2011-10-06 13:39 <wolfspraul> and 4 hours with video playing
2011-10-06 13:40 <wolfspraul> I think over time the battery may loose some power, after hundreds of recharge cycles. Like any other li-ion cell.
2011-10-06 13:40 <mac> I will be using it for command line based data entry.
2011-10-06 13:40 <wolfspraul> some people are using 1200 mAh batteries in their nanonotes (nokia bl-5c types)
2011-10-06 13:41 <wolfspraul> they will not really fit well, it's about tolerances, but if you squeeze the battery door somewhat they do fit
2011-10-06 13:42 <wolfspraul> with a bl-5c or even bl-6c (but then the battery door will definitely not fit) you will get a lot more hours
2011-10-06 13:42 <wolfspraul> but anyway, most people use the normal bl-4c 850mAh type
2011-10-06 13:42 <wolfspraul> 4-15 hours
2011-10-06 13:42 <wolfspraul> :-)
2011-10-06 13:43 <mac> I have been using a zaurus for this up until now.  I typically turn the unit on and off throughout a 6 hour period.  Does the nanonote go in and out of  sleep mode easily?
2011-10-06 13:44 <mac> Thanks for the responses.  As you can probably guess I am new at the chat thing.
2011-10-06 13:44 <wolfspraul> even better, you are very welcome here
2011-10-06 13:44 <mth> suspend and resume should be less than a second each
2011-10-06 13:44 <mth> with the latest kernel, not sure if the shipped kernel also has suspend to ram
2011-10-06 13:44 <wolfspraul> I don't think sleep (suspend/resume) are well implemented for the end user right now
2011-10-06 13:45 <wolfspraul> they do work in principal, but I doubt it's fully functional in the gui etc.
2011-10-06 13:45 <mth> we have a daemon for system functions on the Dingoo, suspend is power + A there
2011-10-06 13:45 <mth> the daemon could be ported, probably it just needs a different hotkey config
2011-10-06 13:45 <wolfspraul> yes we have triggerhappy on the nanonote, I'm just not *sure* right now it all works
2011-10-06 13:45 <wolfspraul> for suspend/resume
2011-10-06 13:45 <mth> ok
2011-10-06 13:46 <mth> worst case, you could issue the suspend command from the command line
2011-10-06 13:47 <mac> Suspend from command line will work just fine.   A little ZZ script.
2011-10-06 13:47 <wolfspraul> and a request from someone like mac who wants to use a NanoNote for a real life application may be the one last drop that makes us prioritize this to the point that it gets done well :-)
2011-10-06 13:47 <mac> One last:  I use sqlite.  Does that come with the os/ is it in the feed?
2011-10-06 13:48 <wolfspraul> good question, let's see
2011-10-06 13:48 <bartbes> antoniodariush: I tried both, iirc
2011-10-06 13:48 <bartbes> though of course I had no way of seeing which one actually worked
2011-10-06 13:49 <antoniodariush> ok
2011-10-06 13:49 <wolfspraul> mac: here is a first overview of apps/libraries http://en.qi-hardware.com/wiki/Applications
2011-10-06 13:49 <antoniodariush> can we try to use libinconv-full?
2011-10-06 13:49 <antoniodariush> do you have it already so i can try it out?
2011-10-06 13:49 <wolfspraul> libsqlite3 is listed there (but that's a wiki and may be outdated). let's see in the build dirs
2011-10-06 13:51 <wolfspraul> http://downloads.qi-hardware.com/software/images/NanoNote/Ben/latest/config
2011-10-06 13:51 <wolfspraul> CONFIG_PACKAGE_libsqlite3=y
2011-10-06 13:51 <wolfspraul> mac: so yes, pretty sure it's in the latest OpenWrt image
2011-10-06 13:52 <wolfspraul> when you get a NanoNote, it will not have the latest image, so you first need to upgrade to the latest image, 2011-08-28
2011-10-06 13:53 <mac> Thanks.  I ordered last night and look forward to getting started on it.
2011-10-06 13:54 <wolfspraul> wow great
2011-10-06 13:54 <wolfspraul> mac: please stop by here again when you get it, we can most likely make your first steps much easier
2011-10-06 13:55 <wolfspraul> even if something doesn't work, your feedback may make us fix whatever problems you run into
2011-10-06 13:55 <mac> wolfspraul:  I will.  Thanks again.
2011-10-06 14:23 <kristianpaul> I'm really happy with my ben,
2011-10-06 14:24 <kristianpaul> Boot time had improved a lot, so i yday used it (instead of my laptop) to  transfer son logs files from a fsck log
2011-10-06 14:24 <kristianpaul> s/son/some
2011-10-06 14:27 <kristianpaul> I must recall i was in a datacenter, so the faster the better to get out of the place soon :)
2011-10-06 14:29 <wolfspraul> nice
2011-10-06 14:29 <wolfspraul> I think the software for the Ben is on a good track
2011-10-06 14:29 <wolfspraul> the goals remain the same :-) still more and better kernel.org upstream support, finally the same for u-boot, keep a small diff with openwrt upstream
2011-10-06 14:30 <wolfspraul> better testing process for image releases
2011-10-06 14:30 <wolfspraul> better update procedures (not just full reflashing)
2011-10-06 14:30 <kristianpaul> ah,ye, last days i been loading the ben memcard with some content, most of then in html
2011-10-06 14:31 <wolfspraul> better software (music, math, dictionary, etc), better scripting, better ways to download data like maps, music, videos
2011-10-06 14:31 <wolfspraul> and so on :-)
2011-10-06 14:31 <wolfspraul> delay tolerant networks, voip
2011-10-06 14:32 <kristianpaul> there are plenty software now, improve documentatio yes, and a better file manager
2011-10-06 14:33 <wolfspraul> faster boot, suspend/resume, reliable alarm clock, fix bugs in general
2011-10-06 14:33 <wolfspraul> I'm wondering whether Hans Bezemer got his nano back to boot from nand
2011-10-06 14:33 <wolfspraul> it looked like he got a bad block exactly in a place where our software cannot handle it
2011-10-06 14:34 <wolfspraul> unfortunately then it went silent, I hope his nano works. need to followup, that would be an interesting bug to fix.
2011-10-06 14:34 <wolfspraul> oh we need to look at the partitioning again as well
2011-10-06 14:34 <kristianpaul> :)
2011-10-06 14:34 <wolfspraul> maybe find an easy solution to backup all data, and/or to encrypt all data
2011-10-06 14:35 <kristianpaul> at least to keep setting after upgrade
2011-10-06 14:35 <wolfspraul> usb gadgets! make it easy to switch the ben to become a keyboard, mouse, soundcard, usb storage, etc.
2011-10-06 14:35 <kristianpaul> i lost 6 levels from alex4 last time :?
2011-10-06 14:35 <kristianpaul> s/:?/:/
2011-10-06 14:35 <wolfspraul> so there's a lot of good stuff that I hope we get to in time, without creating a maintenance mess, i.e. everything 100% upstream
2011-10-06 14:35 <wolfspraul> or as much as possible
2011-10-06 14:36 <wolfspraul> I think more and easier ways to upgrade software, to backup and encrypt data would be great
2011-10-06 14:36 <wolfspraul> right now it's all too manual and too complicated
2011-10-06 14:36 <kristianpaul> slow :)
2011-10-06 14:37 <kristianpaul> yes,backup make sense too
2011-10-06 14:38 <kristianpaul> may be something like,if you plug a memcard it automatically askyou to backup your /root/ /etc folder on it
2011-10-06 14:38 <antoniodariush> bartbes, compiling libunistring with liconv-full fails http://pastebin.com/2MzM23Yz
2011-10-06 14:41 <kristianpaul> antgreen: hi! 
2011-10-06 14:41 <antgreen> hello
2011-10-06 14:42 <wolfspraul> antgreen: you are the moxie guy, right? a nice project and it's good we finally take more notice of it!
2011-10-06 14:42 <antgreen> yes - that's me! 
2011-10-06 14:42 <wolfspraul> what do you hope to achieve with moxie?
2011-10-06 14:43 <antgreen> is working - have a work related IRC going on right now.  will be slow to respond.
2011-10-06 14:43 <wolfspraul> you are aware of Milkymist - is there overlap? what's different from Milkymist? can you reuse anything from Milkymist or vice versa? what help are you looking for the most wrt moxie right now?
2011-10-06 14:43 <wolfspraul> no problem, this channel is logged and easy to search
2011-10-06 14:43 <wolfspraul> and I think moxie is a search friendly name :-)
2011-10-06 14:44 <wolfspraul> http://en.qi-hardware.com/irclogs/search?q=moxie
2011-10-06 14:44 <wolfspraul> so you can reply slowly, we find it in the backlog or search...
2011-10-06 14:51 <wolfspraul> this may be a stupid question, but hey, have to dare to ask, right? :-) so...
2011-10-06 14:51 <wolfspraul> there is always this talk of adding a MMU to lm32
2011-10-06 14:51 <antgreen> yes, I've known about milkymist for a long time, and lekernel and I have traded a few mails and chatted on IRC over the past couple of years
2011-10-06 14:51 <wolfspraul> if we would do that, would that mmu be lm32-specific? does moxie have an mmu? if not, can a yet-to-be-written lm32 mmu be ported over to moxie one day, or is it closely tied to the core so that that makes no sense
2011-10-06 15:14 <bartbes> antoniodariush: that might be why it used libiconv, it compiled and didn't complain
2011-10-06 15:14 <bartbes> :P
2011-10-06 15:15 <antgreen> re: "what do you hope to achieve with moxie?".  right now it is a learning platform for me, but I certainly hope that it becomes a useful and competitive soft-core with a solid software stack.
2011-10-06 15:15 <antoniodariush> bartbes, it went fine now with liconv-full but now I cannot build that simple program
2011-10-06 15:16 <antgreen> I come from a dev tools background, and have been involved in many many GNU tools ports (including the original NIOS GCC port, which was my first exposure to FPGAs)
2011-10-06 15:16 <antoniodariush> i get this error : http://pastebin.com/hRJVMbtK
2011-10-06 15:16 <antgreen> and I was surprised at how many compiler-unfriendly processors there were out there
2011-10-06 15:17 <antgreen> so creating my own compiler friendly ISA has a been an interest for a long time,
2011-10-06 15:17 <antgreen> and definitely possible with all of the great free software these days.
2011-10-06 15:17 <bartbes> antoniodariush: sounds like a missing -l flag
2011-10-06 15:18 <antgreen> wolfspraul: there is no MMU for moxie yet, although somebody sent me notes on a possible MMU design 
2011-10-06 15:18 <antgreen> I would love to collaborate on an MMU, but am concerned about licensing.
2011-10-06 15:19 <antgreen> I currently use GPL, but eventually will switch to something more liberal (less Free), like MIT or BSD
2011-10-06 15:20 <wolfspraul> curious - why?
2011-10-06 15:20 <antgreen> So I would like an MIT or BSD licensed MMU
2011-10-06 15:20 <wolfspraul> the core is the only part in Milkymist that is not GPL :-)
2011-10-06 15:21 <antgreen> GPL doesn't map cleanly to hardware descriptions, also, I want moxie to be maximally useful to people, and there is still a lot of resistance around GPL, especially in embedded
2011-10-06 15:21 <antoniodariush> bart
2011-10-06 15:21 <antoniodariush> bartbes, * where about?
2011-10-06 15:21 <antgreen> I love the GPL, don't get me wrong.
2011-10-06 15:22 <bartbes> antoniodariush: it just looks like it's not linking with libiconv
2011-10-06 15:22 <antgreen> I wrote/maintain a library called libffi that is very popular and gets lots of contributions, but I think the BSD license is a big part of that
2011-10-06 15:22 <wolfspraul> no worries, we are all relaxed here and try to understand
2011-10-06 15:23 <wolfspraul> I would agree about the resistance in the industry, but you can say the same about patents. to most companies, unless you stuff is patented it hardly exists, definitely not 'yours', etc.
2011-10-06 15:23 <wolfspraul> so if you have unpatented bsd stuff, you will still see the same end result, people saying "legally unsafe" (because unpatented :-))
2011-10-06 15:23 <antgreen> bah - don't get me started
2011-10-06 15:23 <wolfspraul> then you will start to file patents as well?
2011-10-06 15:24 <antgreen> I'm not sure.
2011-10-06 15:24 <wolfspraul> I'm just saying those in embedded that have a problem with the GPL probably also have a problem with unpatented tech
2011-10-06 15:24 <antgreen> Yes, I'm sure there's some overlap.
2011-10-06 15:24 <wolfspraul> it's just so natural to patent for them, if you don't that means something is wrong with your stuff
2011-10-06 15:25 <wolfspraul> but feel free to choose whatever you like and let us know whether you feel it changes anything
2011-10-06 15:25 <antgreen> they are mostly useful to raise money from investors
2011-10-06 15:26 <antgreen> what would be really great is a license that speaks directly to processor IP.
2011-10-06 15:26 <wolfspraul> scary, isn't it :-) the herd...
2011-10-06 15:27 <antgreen> like I said, GPL doesn't map very cleanly, so I don't think people will immedialy know what their responsibilities are for GPL HDL
2011-10-06 15:27 <antgreen> LGPL makes even less sense IMO
2011-10-06 15:27 <wolfspraul> so you mean an mmu could well be designed and work for both moxie and lm32? the same mmu?
2011-10-06 15:28 <antgreen> opencores has really promoted LGPL for hardware IP, and I think that's sad
2011-10-06 15:28 <wolfspraul> would that make it sub-optimal for both?
2011-10-06 15:28 <antgreen> I don't think so
2011-10-06 15:28 <antgreen> I mean, I think the same MMU could work for both, and not be sub-optimal
2011-10-06 15:29 <wolfspraul> interesting
2011-10-06 15:29 <wolfspraul> what are the highest quality opencores projects in your opinion
2011-10-06 15:29 <wolfspraul> and in which products are opencores projects used?
2011-10-06 15:30 <antgreen> I don't know enough to judge
2011-10-06 15:32 <wolfspraul> sorry gotta go soon, this is all very helpful that you explain a bit about moxie! cool!
2011-10-06 15:32 <wolfspraul> one more question:
2011-10-06 15:32 <LunaVorax> is BSD licensing wrong?
2011-10-06 15:32 <wolfspraul> what are the next most important features/bugs you want to work on in moxie?
2011-10-06 15:33 <wolfspraul> I have the impression that in the long run, gpl licensed projects tend to have a more balanced multi-party contributor structure (the ones that are successful and have any active contributors at all)
2011-10-06 15:34 <wolfspraul> when I see a BSD project, I tend to be a little cautious on that, and check whether it's actually mainly controlled by 1 party (which may or may not be a problem, just saying in response to your gpl/bsd question)
2011-10-06 15:34 <wolfspraul> but no, I don't think there is anything wrong with BSD
2011-10-06 15:35 <aisa> There was a fascinating article, some long while ago, researching user communities around BSD/GPL/&c licenses.
2011-10-06 15:35 <wolfspraul> for data, I'm even leaning towards public domain lately, cc-zero etc. nobody follows those multi-page agreements anyway
2011-10-06 15:35 <aisa> IIRC, GPL was an amazing bootstrap, but that some communities would/could/did benefit, after becoming mature, in switching to the BSD license.
2011-10-06 15:36 <wolfspraul> can you find the url?
2011-10-06 15:36 <bartbes> antgreen: you made libffi? awesome!
2011-10-06 15:37 <antgreen> bartbes: thanks!
2011-10-06 15:37 <aisa> I'm looking.  I'm finding conversation, but not convinced yet that I've found exactly what I read.
2011-10-06 15:40 <aisa> this is the only thing I've found:
2011-10-06 15:40 <aisa> http://www.softpanorama.org/Copyright/License_classification/index.shtml
2011-10-06 15:40 <antgreen> wolfspraul: I agree that GPL is a great choice for the reason you describe above.  I've been dealing with GPL-in-embedded for many years now (since '95), and I think it's the wrong choice for semiconductor IP today.
2011-10-06 15:40 <aisa> I'm not sure if that is what I read, though.
2011-10-06 15:40 <antgreen> GPLv3 is particularly bad.
2011-10-06 15:41 <antgreen> Well, let me step back a second...
2011-10-06 15:41 <antgreen> If you want to maximize your user base, the GPL isn't the first license I would choose.
2011-10-06 15:43 <aisa> The article I read was strictly a research paper, and didn't have a horse in the race, as it where.
2011-10-06 15:43 <aisa> so if my above linked article takes a stand in conclusion, it's the wrong one :-)
2011-10-06 15:45 <antgreen> wolfspraul: re "what are the next most important features/bugs you want to work
2011-10-06 15:45 <antgreen>      on in moxie?": there's so much to do!  First, I want it to run code!  I don't know if you've been following my blog, but I basically got it to execute its first few instructions the other day.  
2011-10-06 15:45 <antgreen> Now I need to access RAM, and am trying to figure out how to best do that.
2011-10-06 15:46 <antgreen> It looks like I will use separate busses for instruction and data memory
2011-10-06 15:47 <antgreen> but I'm not sure yet how to handle reads and writes on the same data bus.
2011-10-06 15:47 <antgreen> this is computer architecture 101 stuff
2011-10-06 15:47 <antgreen> I'm starting at the beginning!
2011-10-06 15:48 <antgreen> it will (and has been) a long haul
2011-10-06 15:49 <antgreen> maybe I'll be ready for Milkymist Three
2011-10-06 15:49 <antgreen> :-)
2011-10-06 16:57 <wpwrak> aisa: (changing license) if the project has a mixed group of contributors and doesn't require copyright assignments for contributions or similar, it's usually not feasible to switch the license at a later point in time
2011-10-06 16:58 <aisa> I was trying to quote the researcher, you're correct that the problem contains nuance.
2011-10-06 16:58 <aisa> IIRC, and I may not, the researcher was saying that socially, being able to do such a thing would result in a net-benefit.
2011-10-06 16:58 <aisa> but it has been many years.
2011-10-06 16:59 <wpwrak> what the GPL does very nicely is to give people who contribute the safety that they're not being exploited. it may also keep some of the nastier characters away from the project, which also helps.
2011-10-06 17:01 <wpwrak> for companies, there's probably not such a big difference between contributing to BSD-licensed and contributing to GPL-licensed (at least for GPLv2). few really have the kind of long-term plans where the difference would matter.
2011-10-06 17:01 <aisa> I guess you'll have to find the researcher in question and ask him about it. :-)
2011-10-06 17:02 <phr> i tend to stay away from volunteer development for BSD projects... i'm happy to work on them if i'm getting paid
2011-10-06 17:03 <aisa> I certainly have my own horse in this race, but I'm not entertaining a change of opinion on the matter, myself.  I was hoping, in my mentioning it, to cite data, whether it matched my preference or not.  Which I find always interesting even when it challenges my own assumptions.
2011-10-06 17:03 <wpwrak> (net benefit) hmm, i see it as a good flame bait if you do it without including it in the roadmap from very early on ;-) and if people know it's coming, it may actually prevent contributions even before the switch. kinda like the osborne effect.
2011-10-06 17:04 <wpwrak> aisa: yeah, if you can dig out the reference, that would be interesting
2011-10-06 17:04 <aisa> I posted above what it *might* have been, alas I was not able to confirm this was what I read.
2011-10-06 17:04 <aisa> It was a good number of years ago, and a lot of electrons have been inconvenienced on the topic since thene,
2011-10-06 17:05 <wpwrak> licensing discussions tend to have that effect, yes ;-)
2011-10-06 17:05 <aisa> then*, and I wasn't sufficiently able to articulate to google that I didn't want the results it was returning.
2011-10-06 17:06 <aisa> all of the links I'm finding are advocacy, alas.
2011-10-06 17:08 <kristianpaul> antgreen: You are making your own wishbone variant,or all upstream as posible including IP cores?
2011-10-06 17:09 <kristianpaul> antgreen: btw so you moxie cpu is ready, just missing how to put the SoC towork together?
2011-10-06 17:10 <aisa> ok, it was very much this link: http://www.softpanorama.org/Copyright/License_classification/index.shtml
2011-10-06 17:10 <aisa> it has been greatly expanded since I last visited it.
2011-10-06 17:10 <aisa> and my memory of it may not be consistent with it's current incarnation.
2011-10-06 17:10 <aisa> but it may be!
2011-10-06 17:12 <aisa> wpwrak: in case you didn't see, I did post the url referring to my memory, exactly for what it is worth.  ;-)
2011-10-06 17:21 <antgreen> kristianpaul: I want to reuse as much surrounding logic as possible, so standard wishbone.
2011-10-06 17:22 <antgreen> the moxie core isn't even ready yet
2011-10-06 17:22 <antgreen> it's under slow-but-active development
2011-10-06 17:23 <antgreen> I have never even run anything on hardware yet.  Just iverilog and gtkwave.
2011-10-06 17:23 <antgreen> I have a Altera DE2 board waiting, though,
2011-10-06 17:23 <antgreen> and I'm synthesizing with altera's tools.
2011-10-06 17:25 <kristianpaul> antgreen: but the gcc port?
2011-10-06 17:25 <antgreen> done
2011-10-06 17:25 <kristianpaul> I meanso i get updated as soon you mvoe
2011-10-06 17:25 <kristianpaul> ok
2011-10-06 17:25 <kristianpaul> s/i/it
2011-10-06 17:25 <antgreen> it's all upstream in FSF's gcc tre
2011-10-06 17:25 <antgreen> e.  As well as gdb and binutils.
2011-10-06 17:26 <kristianpaul> qemu?
2011-10-06 17:26 <antgreen> I have a linux kernel that boots on a sim, and a qemu port
2011-10-06 17:26 <antgreen> and RTEMS and u-boot ports
2011-10-06 17:26 <antgreen> and uClibc
2011-10-06 17:26 <antgreen> and busybox
2011-10-06 17:26 <antgreen> but only the gnu tools work are upstream
2011-10-06 17:27 <antgreen> the rest are in moxiedev: http://github.com/atgreen/moxiedev
2011-10-06 17:27 <kristianpaul> yeah,your repo is quite impresive ;)
2011-10-06 17:27 <antgreen> thanks!
2011-10-06 17:28 <bartbes> regarding licensing: copyleft is fine for 'products', but imo a library(/utility?) or similar really benefits from copyfree, since copyleft severely limits your users
2011-10-06 17:29 <bartbes> even for 'products', if you plan on letting other people work with your code, copyleft is kind of a dick move
2011-10-06 17:33 <wpwrak> aisa: (link) thanks !