<mark_weaver>I've forgotten the details of your board. is it an fpga board? <nalaginrut>well, I plan to provide the performance between Arduino and Rpi, plus wifi, and cheap <nalaginrut>but we have alternative, GD, maybe you never heard <nalaginrut>the software part, I'll provide a way to let Lispers play hardware <nalaginrut>mark_weaver: it's similar to STM32, I can hardly say it's copy-cat, but you know what I mean <nalaginrut>mark_weaver: you know, from market perspective, a processor provides some clock rate, but sometimes, their choice is not so reasonable for us <nalaginrut>and the registers define and pins are same with stm32 <nalaginrut>mark_weaver: do you know picobit? it's one of the solution to run Scheme on bare metal <mark_weaver>you said you'll "provide a way to let Lispers play hardware". what kind of software will you put on the device for this? <mark_wea`>sorry, losing my connection a lot since moving to a different ISP. <nalaginrut>for all things above, and I'm in Shenzhen, the holy land of hardware manufacturing, and I have a hackspace/community here, so I happens to have a chance to try it <mark_wea`>indeed, you are well located for hardware hacking :) <mark_wea`>it's okay, I can understand you well enough :) <nalaginrut>if you want to learn real computer science, don't play any given OS ;-) <mark_weaver>well, for most computer science, the choice of OS is mostly irrelevant. <mark_weaver>but hacking at the low level is fun. I've done some of that, long ago. <nalaginrut>yeah, computer science is all about math, so it'd be related to engineering <nalaginrut>but anyway, I hope to use lisp/scheme to play hardware <nalaginrut>at least, I don't have to worry about integer overflow in 99.99% possibility, since the number could be very large <nalaginrut>although for a limited hardware, the number can't be unlimited <mark_weaver>the number can't be unlimited for any finite hardware <mark_weaver>but yeah, for embedded use, bignums are probably a bit silly :) <nalaginrut>last year, I'm worry about that I can't provide wifi in current solution <nalaginrut>mark_weaver: few months ago, there's a startup Chinese chipset company, they provides very cheap wifi chip <nalaginrut>unfortunately, I can't get the source code of firmware <nalaginrut>tonight we'll have regular meeting in hackspace, so we'll discuss and analysis it tonight <mark_wea`>my new ISP is terrible. drops my IRC and SSH connections very frequently. <nalaginrut>last year, I'm worry about that I can't provide wifi in current solution <nalaginrut><nalaginrut> mark_weaver: few months ago, there's a startup Chinese chipset company, they provides very cheap wifi chip <nalaginrut><nalaginrut> unfortunately, I can't get the source code of firmware <nalaginrut>I have to push the project anyway, because the firmware is the only closed part <nalaginrut>and we don't know the market in the future, if the companies see the big success of open, they may change their mind <nalaginrut>moreover, if we can sell more products, maybe we have enough chip to talk to the chip company ;-D <mark_wea`>we can only make it happen by refusing to buy devices that require non-free software to run <nalaginrut>mark_weaver: if you do so, the companies will treat you as enemy, and you lose the chance to talk <mark_weaver>I think the kind of talk they'll understand best is "I'm going to buy from Atheros instead" <nalaginrut>mark_weaver: yes, I've bet Atheros in the past, but ar6003 driver for bare metal is not so easy <nalaginrut>I prepared two hands, if the board sell well, I'll negotiate with the wifi chip company <nalaginrut>if not so nice, but earned some money, I can find people to port the driver <mark_weaver>do you really think you'll sell enough units to persuade them? <mark_weaver>I might have bought a unit from you if not for this, fwiw. just tonight I was talking with some friends about a project, and this would have been perfect, if not for that. <mark_weaver>not that I expect my one purchase to be significant, but just fyi. <nalaginrut>it's open hardware, I'll try my best to let the hardware be fine, and you may hack it if you think it's not so cool <nalaginrut>but anyway, I don't imagine I can persuade them, it depends how cool stuff I can provide ;-) <mark_weaver>the market is totally flooded with hardware devices like this, and some of them don't require any non-free software to run. being able to run without blobs is a significant draw for many of us. you'll be missing out on that market segment, fwiw. <nalaginrut>mark_weaver: I do want to provide pure free software, but there's always compromise when you are in a business, although I hate compromise, I have to push it rather than do nothing ;-D *nalaginrut hold optimistic opinion <mark_weaver>well, it'll be interesting to see if you'll be able to stand out from the crowd enough to be successful. <nalaginrut>mark_weaver: even another way, the final aim is rsf, for me ***eMBee is now known as no-one
***no-one is now known as eMBee
<amirouche>sneek: later tell davexunit have you pusblished your bootstrap theme for texinfo ? <davexunit>amirouche: got your message via sneek in #guix. I haven't made a real bootstrap theme for texinfo. I just passed a flag to makeinfo that includes the bootstrap.css base stylesheet on every page. <davexunit>I don't yet know how to change the template it uses to generate pages. I would like to change that, too. <linas>so I've got a guile-2.2 buglet, unless I am using it wrong <linas>I say this: (spawn-server (make-tcp-server-socket #:port 18001)) <linas>Guile comes with ABSOLUTELY NO WARRANTY; for details type `,show w'. <linas>This program is free software, and you are welcome to redistribute it <linas>under certain conditions; type `,show c' for details. <linas>ERROR: In procedure string->utf8: Wrong type argument in position 1 (expecting string): socket <linas>Guile is distributed WITHOUT ANY WARRANTY. The following <linas>sections from the GNU General Public License, version 3, should <linas>so the socket mostly works, but something about compiling is unhappy <linas>this is from code built from igt about a week ago. <linas>Should I just open a bug report for this, or is my code insane? <davexunit>linas: I can't spot what's going on there. I haven't telnet'd to my REPL in some time, but when I have it has worked. if no one else has any input, I would say just open a bug about it. <linas>its low urgency, as I don't use it that way, anyway. <linas>I guess maybe my question was this: does it work for you? <linas>silly me, I have a differrent way of testing this anyway <mark_weaver>linas: can you send a message to bug-guile@gnu.org about the problem? <mark_weaver>wingo: in master, evaluating (+ 2 2) in a REPL server doesn't work. Turns out, the reason is the (port-filename <socket>) returns the symbol 'socket', and assembler.scm:2188 assumes it's a string. <mark_weaver>I'm thinking that maybe we should change (port-filename <socket>) to return #f. WDYT? <mark_weaver>or alternatively we could make assembler.scm robust against non-string "filenames". thoughts? <wingo>mark_weaver: humm, interesting! <wingo>btw i looked into the goops bug yesterday for a few hours -- it's gnarly :) <wingo>will see if i can get a fix soon <davexunit>ArneBab: thanks :) someone very kindly made it for me. <ArneBab>congrats: Sly obviously already inspires people. <ArneBab>davexunit: I had to fudge a py2guile logo myself ☺ *davexunit heads home now <dsmith-work>ArneBab: Ooo. There was something someone wrote thawas similar. "A pythonisa in the land of Scheme", or something like that. <ArneBab>dsmith-work: yes - that got me hooked ☺