Conversation with #inferno at Wed Jan 12 11:54:07 2011 on powerman-asdf@irc.freenode.net (irc)
(11:55:58) powerman-asdf left the room (quit: Ping timeout: 276 seconds).
(13:40:59) GriffenJBS_ [~chatzilla@adsl-8-52-40.mia.bellsouth.net] entered the room.
(13:42:55) GriffenJBS left the room (quit: Ping timeout: 255 seconds).
(13:42:55) GriffenJBS_ is now known as GriffenJBS
(13:56:32) bvalek2 left the room (quit: Ping timeout: 265 seconds).
(14:32:42) bvalek2 [c3e41404@gateway/web/freenode/ip.195.228.20.4] entered the room.
(15:31:18) bvalek2_ [c3e41404@gateway/web/freenode/ip.195.228.20.4] entered the room.
(15:34:10) bvalek2 left the room (quit: Ping timeout: 265 seconds).
(15:37:06) bvalek2_ is now known as bvalek2
(16:07:37) stewarj: uriel: is there an inferno that is more solid, or is this an observation on the state of inferno in general?
(16:12:37) mjl-: inferno-os is the most solid one.  and i don't think it is not solid.
(16:12:42) mjl-: sure, it can be a lot better
(16:14:28) stewarj: that's what i've been experimenting with.
(16:14:57) stewarj: i didn't know what uriel was addressing.
(16:15:20) stewarj: VN's release, the fonts, inferno in general, or something else.
(16:16:14) stewarj: mjl: thanks for the help you've given!
(16:16:41) mjl-: hmm, did i? :)
(16:16:46) mjl-: no problem then :P
(16:17:05) stewarj: (in times past...)
(16:17:49) stewarj: i've been a leech on the group so far... just acknowledging your help!
(16:24:25) mjl-: ok, cool :)
(16:35:23) duper [~super@projects.ext.haxnet.org] entered the room.
(16:38:15) Gegemon left the room (quit: Quit: Leaving.).
(16:45:39) stewarj: brian stuart's book has been very helpful for understanding internals, but i'm only reading it in bits and pieces.
(16:46:58) stewarj: with respect to making inferno "better", is there a "todo" or "help wanted" list anywhere for folks to pitch in?
(16:56:26) mjl-: there's this: http://code.google.com/p/inferno-os/wiki/InfernoAndLimboPlans
(16:57:05) mjl-: but that's more about the core of inferno
(16:57:27) mjl-: inferno could also use more libraries & applications
(16:57:33) mjl-: i've got a personal wish list
(16:57:59) mjl-: some things on that list may already be possible, but i have to find that out
(16:58:14) mjl-: one of those is serving ssl connections (with decent performance of course)
(16:58:34) mjl-: well, including certificate handshakes and such.  ssl alread works
(16:58:57) mjl-: and of course there's the issue list: http://code.google.com/p/inferno-os/issues/list
(17:07:15) mjl-: stewarj: do you have any plans/needs?
(17:07:18) mjl-: for inferno that is
(17:10:07) stewarj: security is an important aspect in my roadmap for what i hope to apply limbo, so despite my total failure at understanding ssl and it's guts, i'd be willing to pitch in if everyone could be important
(17:12:05) stewarj: but back to my interest in inferno, i work on a product that moves cellphone calls onto the IP network. my part is device setup and status reporting (not the radio part) and indeed the OS and hardware are supplied by a third-party vendor.
(17:14:00) mjl-: the ssl/tls rfc's don't seem very big
(17:14:09) mjl-: so if it even needs work, it shouldn't be too hard
(17:14:22) mjl-: luckily crypto is quite easy on inferno (as long as the algorithms are supported ;))
(17:14:50) mjl-: device setup & status reporting.  is that also voip?
(17:14:50) stewarj: i'm not at the stage yet where i've even made the first ssl connection via inferno
(17:15:16) stewarj: femtocell... uses "IuH"
(17:15:31) mjl-: IuH?
(17:16:18) stewarj: cellphone/telecom acronym
(17:16:35) robot12 left the room (quit: Quit: Leaving.).
(17:17:13) mjl-: femtocells sound like a good idea
(17:17:37) mjl-: based on the first line of the wikipedia page on the subject :P
(17:18:12) stewarj: i think they're a bridging technology. ultimately, i'd love to have wifi everywhere and just use voip from my portable phone/pda/thingy
(17:18:31) stewarj: but in the states at least, 3G will be the norm for a while.
(17:18:51) mjl-: isn't 3g (and newer) supposed to be much more energy efficient?
(17:19:34) mjl-: but, do you need some kind of special connection to a telco if you want to use a femtocell?
(17:20:24) stewarj: 3g is more efficient, but i feel like that won't be the case forever... once again, i'm not a radio guy, so i'm talking out my ass to some degree
(17:20:57) stewarj: femtocell in the states require telco pieces in place.
(17:21:23) stewarj: (you have to get traffic onto their switching systems via IP)
(17:25:01) stewarj: silly irc question here: are the #inferno chats logged anywhere?
(17:25:12) mjl-: not that i know off
(17:25:24) mjl-: perhaps we should start logging...
(17:26:37) stewarj: didn't know if there's any valuable knowledge that gets "lost" here.
(17:26:52) mjl-: nah, not too much
(17:26:59) stewarj: hehe
(17:27:15) mjl-: and i wouldn't want to search in irc logs for tiny bits of knowledge
(17:27:21) mjl-: there's usually too much noise
(17:27:36) mjl-: i find logs are mostly useful to recall something that you know was discussed but forgot the details of
(17:35:41) stewarj: got it.
(17:52:53) stewarj: just to be candid, i need a "safer" language for my platform than straight-up C. java might fill that role, but it still doesn't seem suited to "embedded" (don't flame me for calling a ARM/linux box with 64M RAM / 64M flash embedded) and IMO, it still is crummy at addressing concurrent programming issues (yeah, I'm a total sucker for CSP).
(17:56:19) stewarj: SSL, particularly certificates/CRL are a big piece of the puzzle.
(17:57:10) stewarj: and to be honest, i don't know what's missing if anything from inferno on this part
(18:00:11) stewarj left the room (quit: Quit: stewarj).
(18:01:01) base2design [~base2desi@97-80-161-40.dhcp.gwnt.ga.charter.com] entered the room.
(18:03:47) bvalek2 left the room (quit: ).
(19:48:36) btdn left the room (quit: Ping timeout: 260 seconds).
(20:13:37) j1m_ [~AndChat@213.87.74.188] entered the room.
(20:18:02) KillerX [~anant@nat/mozilla/x-iwiaxrugaurvwrjo] entered the room.
(20:33:35) base2design: anyone here have experience with a commercial inferno license?
(20:44:02) anth_x: maybe? years ago - i don't know if VN's still doing the same terms since the open source releases.
(20:49:13) base2design: is there a per-box license cost if the custom part isn't open-source?
(20:50:12) base2design: that is, if i can't share the source for my part of my inferno installation, do i need to license it, and if so is it a one-time cost or a per-shipped unit cost.
(20:50:38) hotaru2k3 left the room (quit: Ping timeout: 240 seconds).
(21:10:32) anth_x: whether you need to license it or not it a bit more complicated than that - it depends on what you're doing, exactly.
(21:10:54) anth_x: in the past, i believe they've done both types of commercial licenses, with per-seat/unit being more common.
(21:11:12) anth_x: what type of thing are you writing that you can't share source for?
(21:11:22) anth_x: (application, kernel module, port...)
(21:26:05) bvalek2 [598516ea@gateway/web/freenode/ip.89.133.22.234] entered the room.
(21:33:10) j1m_ left the room (quit: Ping timeout: 255 seconds).
(21:46:47) hotaru2k3 [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(22:00:16) Fish- [~Fish@9fans.fr] entered the room.
(22:12:30) base2design: application
(22:13:21) base2design: any changes to emu or the standard codebase would be folded back into the public release
(22:22:10) anth_x: if all you're writing is an application, i'm pretty sure you don't need a commercial license.
(22:22:48) base2design: sometimes when i read the license i get that impression, other times i'm not sure.
(22:23:20) base2design: at any rate, i'd like to contribute to the well-being of inferno...
(22:23:44) anth_x: officially you should consult a lawyer, i guess, but i'm pretty sure. it's functionally equivalent to doing closed-source linux apps, which happens all the time.
(22:24:08) anth_x: contributing to inferno is a worthwhile cause.
(22:24:28) base2design: i've asked NV via charles, but i'm still not clear.
(22:24:30) base2design: :)
(22:25:17) base2design: from my point of view, it's a better alternative to java on small-ish devices.
(22:25:29) anth_x: oh, very much so.
(22:26:05) anth_x: i wish VN wasn't still selling 3rd edition printed manuals - otherwise i'd suggest that as a good way of supporting them while still getting something out of it.
(22:27:11) base2design: in the cable-tv industry (in the states) there was a big push to put java on settops. it's called ocap and it's quite a ugly/slow mess. a very unscientific observation was that a samsung ocap tv set took 5 minutes to boot up. yikes!
(22:27:21) anth_x: and they don't seem to sell t-shirts any more (although honestly that cost too much to fulfill).
(22:27:36) anth_x: whoa
(22:28:11) anth_x: iirc, the first inferno port was for a set-top box.
(22:28:16) anth_x: that's long gone now, though.
(22:29:13) base2design: i think the fundamental vm wasn't too bad (other than being jvm that is) but the ridiculous collection of library code and it's loading contributed to the lag
(22:29:26) base2design: i'd read about the settop past.
(22:29:39) base2design: i'm not in that space anymore
(22:30:05) base2design: i may spring for a 3rd edition manual just for the sake of contributing
(22:30:21) ***mjl- still has a t-shirt
(22:30:31) mjl-: charles gave away the last ones at the iwp9 in madrid
(22:30:40) mjl-: that must be 4 years ago now
(22:30:57) base2design: and the commercial license isn't too bad... so maybe i just need to fork over the $$$
(22:32:02) base2design: on the subject of t-shirts, my wife has a screen press in the basement. if the artwork is still available, i could do a short run if there's any interest
(22:32:16) anth_x: depending on what your budget and requirements are, VN also does training. my last company used them for that to good effect.
(22:32:31) anth_x: (before the company and project imploded for other reasons)
(22:32:33) base2design: (good info about training)
(22:32:42) base2design: what were you working on?
(22:33:02) anth_x: mobile payments.
(22:34:02) base2design: hmm... with inferno?
(22:34:24) anth_x: among other things. it was an odd architecture.
(22:35:04) base2design: what kind of device did it run on?
(22:35:17) anth_x: we had perl in apache (maybe php, too?) doing a web front-end, python talking to the payments processor and cell networks, and inferno controlling the business logic.
(22:35:24) anth_x: the inferno stuff all ran on OS X servers
(22:35:28) anth_x: (i'll miss the xserve)
(22:35:36) anth_x: the user interface was web or sms.
(22:35:53) base2design: wow, that's quite a collection!
(22:36:07) anth_x: this was... ~4 years ago now, so mobile apps and mobile web was quite immature, comparatively.
(22:36:18) anth_x: it wasn't the greatest architecture. it worked, but was fragile.
(22:36:47) anth_x: it ended up that way because we had to get something done very quickly, and we had one guy who was best in python, one who was best in perl, and so on.
(22:36:50) base2design: it's funny that you mention inferno for biz logic, because that's where i see the biggest benefits in the device i'm working on.
(22:37:07) anth_x: the python stuff absolutely should've and could've been inferno.
(22:37:47) anth_x: the web stuff... well, there's a *lot* of libraries in things like perl and php, and while they're not pretty, they do bring some benefit. that likely would've been the last piece to convert.
(22:37:58) base2design: i think the whole internet is held together with gum and toothpicks anyway
(22:38:05) anth_x: heh. yup.
(22:38:52) anth_x: it was fun. closest i've gotten to sticking inferno in production in many years.
(22:38:57) anth_x: but 2011 will be different!
(22:39:07) base2design: half my time is spent writing code that communicates over http "because the security infrastructure requires http/https"
(22:40:22) base2design: indeed our servers sit behind an ssl accelerator that demands http because it inserts stuff into the http headers
(22:41:03) base2design: and it's truly sad that port 80/443 usage "makes things standard"
(22:41:24) base2design: here's to 2011. viva la inferno (to mix metaphors)
(22:42:24) base2design: auth_x, where are you located?
(22:43:09) anth_x: US-OH-CLE+
(22:44:15) base2design: cool.
(22:44:29) base2design: i'm in US-GA-ATL
(22:45:08) base2design: did you make it to IWP9 in athens, ga in '09?
(22:45:25) anth_x: yup
(22:45:51) anth_x: i've made all and only the US ones.
(22:46:09) base2design: i was a lurker there... still trying to make sense of all this stuff.
(22:46:11) anth_x: (not by design; i had a schedule conflict with madrid and a wallet conflict with athens)
(22:47:37) anth_x: it's good stuff. you kinda have to find the right niche to be able to use it, unfortunately: it has non-trivial integration costs.
(22:47:50) anth_x: but once that's overcome, it's such a comfortable and productive way to work.
(22:48:21) base2design: on our embedded device, straight-up c is becoming a liability
(22:48:45) base2design: i won't go on record as to why though...
(22:49:04) base2design: but training people to use a new language is a concern
(22:49:28) base2design: and i'd have to prove a tangible benefit to get sign-off.
(22:50:06) anth_x: depends on the people. i've brought a few people up on limbo from a unix/c background and it's been pretty good.
(22:50:12) base2design: indeed, i've done lots of work with the lua scripting language in the past and it's labelled me as an odd-ball... :P
(22:50:18) anth_x: predictably, the better the programmer, the less learning a new language is an issue.
(22:51:34) base2design: i think go has a wonderful future, but at least on ARM, the runtime footprint is about three times that of a similar inferno bundle... (unless i'm doing something horribly wrong)
(22:51:47) anth_x: wow, i didn't realize that.
(22:51:54) base2design: ymmv
(22:52:18) anth_x: go's great, but it's not the same thing. it fits better (or perhaps just more easily) with existing environments, but that cuts both ways.
(22:52:47) anth_x: i've had to explain this countless times since '97, when the big comparison was java: you get real benefits from being able to virtualize the whole environment.
(22:54:16) base2design: i think my best "gateway drug" at this point is to finish implementing a styx front-end to a few existing apps.
(22:54:48) anth_x: yup. that's a great integration method, too.
(22:54:59) base2design: any feel for that? libixp looks pretty approachable (p9port is a bit big for what i need to do)
(22:55:12) anth_x: in our architecture, the limbo code presented a styx filesystem which the perl scripts used the plan9port tools to read and write.
(22:55:53) base2design: any experience with v9fs in the kernel?... oh you said x-servers didin't you
(22:55:53) anth_x: i've not used any of the smaller libraries
(22:56:09) anth_x: i mean, used them in terms of inside apps that use them, but not really evaluated them.
(22:56:18) anth_x: yeah, no v9fs for us. would've been nice.
(22:56:50) anth_x: something in the 9pfuse+macfuse chain is too fragile. i'm not sure who deserves the blame there, but the effect is that i don't really trust it.
(22:56:57) base2design: i think theres a FUSE interface to 9fs for osx... my dev machine is osx so i peek at that stuff every now and then
(22:57:12) base2design: yuck
(22:58:03) anth_x: yeah, but i don't like it. and the macfuse stuff has stalled (temporarily, i think) for a few months now.
(23:03:31) mjl-: i once tried using v9fs on linux, to mount an ircfs instance
(23:03:32) mjl-: it didn't work back then because the protocol was not the same.  some mode bits were different and caused ircfs to be unusable through v9fs.
(23:03:32) mjl-: that was a disappointment, may be better nowadays
(23:03:34) base2design: anth_x i think i see you on linkedin... i'm gonna send a request ignore it if you like... i won't be offended.
(23:04:23) anth_x: geez, linkedin. let's see if i remember my login information...
(23:05:32) base2design: hehe
(23:05:47) base2design: double yuck
(23:06:21) anth_x: i don't think i see you.
(23:06:36) anth_x: jesus, my first girlfriend sent me a message on linkedin in november!
(23:06:52) anth_x: wtf? that's not exactly a "professional" relationship.
(23:08:45) anth_x: oh, invitations are handled separately from messages.
(23:10:01) anth_x: um, no, i still don't think i see you.
(23:10:20) anth_x: (because i don't think i have any invites or requests or whatever)
(23:11:14) anth_x: heh. that message from the (long-ago-)ex-girlfriend was her telling me she was getting married (on a date that's now passed). oops.
(23:11:27) base2design: it's going through brantley... might take a while
(23:11:45) anth_x: why on earth would linkedin seem like the smartest way to deliver that information? oh, well.
(23:11:53) base2design: hehe
(23:23:46) sl: does a list exist of the lucent products that used inferno? telecom switches and so forth?
(23:33:08) base2design: i've recently asked this and haven't gotten any feedback.
(23:34:33) base2design: i did notice several lucent manuals (hosted on an avaya site of all places) cite inferno and infernospaces in the license/credit page though.
(23:36:38) mjl-: at least i don't know of a list with lucent's inferno products
(23:38:22) sl: base2design: i believe inferno is used in some telecom switches i've had to work with, but i don't think the inferno bits are reachable from the end user interfaces (mostly, it seems, remote hp/ux x11 sessions)
(23:39:54) Fish-: Lucent VPN Firewall Brick, Lucent Manager Firewall, Lucent PathStar Access Server
(23:40:15) hotaru2k3 left the room (quit: Disconnected by services).
(23:40:19) hotaru2k3_ [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(23:40:19) Fish-: the only three products I know who worked with inferno
(23:40:26) base2design: http://support.avaya.com/css/P8/documents/100011598
(23:40:43) base2design: "MAX administration guide"
(23:40:48) base2design: (whatever that is)
(23:41:11) Fish-: also the Lucent/Philips IS2630
(23:41:17) Fish-: which I have
(23:41:21) Fish-: running Inferno 2.2
(23:44:03) Fish-: base2design: you think it's running inferno ?
(23:45:44) base2design: i don't know but check out the trademark part of that document... it mentions inferno and infernospaces.
(23:46:38) Fish-: yes, I seen
(23:47:15) base2design: i pitched plan9 to a consumer division of philips in the mid 90's... it wasprobably  too researchy at the time but i loved the concepts (and still do)
(23:49:53) base2design: fish: is that a crt or a lcd on the is2630?
(23:50:47) anth_r: lcd. it was a phone.
(23:51:22) base2design: ah... just found a better page. neat.
(23:51:28) Fish-: lcd yes
(23:51:40) Fish-: touchscreen
(23:51:59) Fish-: not a bad one for its age
(23:52:59) Fish-: the major problem with this product is it doesn't have an ethernet interface
(23:53:24) Fish-: only dialup modem
(23:53:41) anth_r: there was a hack to get ethernet on it. it has a pcmcia slot.
(23:53:47) Fish-: yes
(23:53:48) anth_r: but no voip or the like.
(23:53:54) Fish-: there is a hidden pcmcia slot
(23:54:01) Fish-: but I don't want to broke it
(23:54:11) anth_r: i'm not sure the ethernet support was distributed.
(23:54:18) anth_r: is it hidden? hrm.
(23:54:42) hotaru2k3_ left the room (quit: Remote host closed the connection).
(23:54:46) anth_r: no, they're visible, behind a door on the right side.
(23:54:49) anth_r: (just checked mine)
(23:55:12) base2design: i may have an old pcmcia network card. i'll give it to you if you want it.
(23:55:48) anth_r: inferno only supported a specific few; the phone likely fewer.
(23:55:57) hotaru2k3 [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(23:56:11) anth_r: the dev work was done with a 3com... 3c589? that's from memory.
(23:56:52) Fish-: you also need to hack it to make it support 5V cards
(23:57:11) base2design: whoops, i lied
(23:57:23) base2design: i have a pcmcia fax modem :P
(23:58:03) base2design: i must have given the ethernet adapter away with my old newton 2100 ;)
(00:00:57) anth_r left the room (quit: Ping timeout: 272 seconds).
(00:01:29) hotaru2k3 left the room (quit: Remote host closed the connection).
(00:04:46) hotaru2k3 [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(00:11:13) hotaru2k3 left the room (quit: Remote host closed the connection).
(00:12:21) hotaru2k3 [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(00:17:06) base2design: anyone ever hear of inferno on blackfin?
(00:17:30) base2design: google didn't uncover anything obvious to me at least
(00:20:54) mjl-: i've heard the name blackfin, but that's all...
(00:21:01) mjl-: now for some weirdness:  http://www.ueber.net/who/mjl/tmp/q.b
(00:21:04) mjl-: and qmk.b
(00:21:28) mjl-: "statically linked .dis programs" ;)
(00:29:26) hotaru2k3 left the room (quit: Ping timeout: 240 seconds).
(00:33:24) hotaru2k3 [debian-tor@gateway/tor-sasl/hotaru2k3] entered the room.
(00:33:52) base2design: that's pretty darn cool!
(00:34:23) base2design: very valuable for embedded stuff (i would think for systems that don't really have filesystems too)
(00:35:32) mjl-: normally the idea is to modify your kernel config (e.g. emu/Linux/emu), and add the files you need to the built-in root file system
(00:35:47) base2design: blackfin = 16/32 bit processors from analog devices (dsp) doesn't have a mmu
(00:36:09) base2design: nice work.
(00:36:22) base2design: (as usual)
(00:37:01) base2design: btw, i started work on an audio device for osx a while back... anyone interested in making fun of it when it's done?
(00:37:33) base2design: (inferno audio device that is)
(00:38:35) base2design: q/qmk is kinda like an initrd in the linux world, right?
(00:39:21) mjl-: yeah, i think so (about initrd)
(00:39:36) mjl-: i don't have any osx, so not sure if i can be of help
(00:40:02) mjl-: base2design: if you configure files into emu, you can see them at kernel device #/
(00:40:08) mjl-: but that requires compiling them into the kernel
(00:40:23) mjl-: i want to be able to take any emu.exe, and run a q.dis that doesn't need any more files
(00:42:55) Fish- left the room (quit: Quit: So Long, and Thanks for All the Fish).
(01:09:02) bvalek2 left the room (quit: ).
(01:12:59) anth_x: base2design: i'd be interested in audio on os x.
(01:19:39) base2design: anth_x: ok... btw, brantley forwarded the linkenin thingy
(01:25:34) base2design: btw, i've done more limbo programming in the last 3 days than i've done in the 10 years i've been aware of it. tuples and channels and alts are very nice. lots of fun.
(01:25:44) btdn [~btdn@99-189-120-112.lightspeed.milwwi.sbcglobal.net] entered the room.
(01:31:11) base2design: mjl... what's the significance of ''Q' in your program? does it stand for anything?
(01:34:38) base2design: btw, i did a dirty dirty trick with lua years back where i allowed scripts to be appended to the end of the binary so you could just do "cat lua script > newlua" and make a self contained runnable version of it. stuff like that keeps people from mis-installing your stuff.
(02:16:43) GriffenJBS: is there anyone left active with inferno-ds?
(02:19:22) base2design: not I... sorry.
(02:52:21) The account has disconnected and you are no longer in this chat. You will be automatically rejoined in the chat when the account reconnects.