IRC log started Sat Dec 4 00:00:00 1999 [msg(TUNES)] permlog 1999.1204 -:- SignOff Tril: #TUNES (Tril has no reason) -:- SignOff hcf: #TUNES (Leaving) -:- NetSplit: merril.openprojects.net split from fontana.openprojects.net [12:15am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [merril.openprojects.net] -:- NetSplit: varley.openprojects.net split from bear.openprojects.net [12:55am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net] -:- Netjoined: varley.openprojects.net bear.openprojects.net -:- Fare [fare@quatramaran.ens.fr] has joined #tunes -:- Fufie [stig@tunnel-44-14.vpn.uib.no] has joined #tunes -:- abi [nef@bespin.dhs.org] has joined #tunes -:- zarq [zarq@9dyn25.delft.casema.net] has joined #tunes -:- yoo [ultima@user-38lc69e.dialup.mindspring.com] has joined #Tunes * yoo/#Tunes pondes * Fufie/#tunes yawns * yoo/#Tunes moos :) * yoo/#Tunes is considering implementing an interpreted languageon top of Lisp if you use lisp-syntax you get the parser/reader for free :) -> can concentrate on the language instead of the parser Ah, well, my concern is that 1) I want a language simpler than lisp and 2) I don't want the language to be able to do much 09:10am -:- yoo_ [ultima@user-38lc6fo.dialup.mindspring.com] has joined #Tunes feh feh feh! -:- yoo_ is now known as yoo__ -:- SignOff yoo: #TUNES (Ping timeout for yoo[user-38lc69e.dialup.mindspring.com]) -:- yoo__ is now known as yoo -:- Downix [down@d-gnaps-286.ici.net] has joined #tunes hey-lo all core? core is developing a kernel-less, component-based OS, Clementine, and the XCOM file format (http://xcom.tunes.org) seen core? hmm need to figure out the bot 09:20am seen core hmm abi, seen core core was last seen on IRC 10 days, 10 hours, 3 minutes and 10 seconds ago, saying: air: or another, but i don't remember installing the vga font on any XFree 3.3.5 boxes [Tue Nov 23 23:23:33 1999] heyyoo there we go ok hmm * Downix/#tunes should give him a ring phone call before anyone gets any funny ideas Oh? What's Core Doing? Developing Clementine, which I want to review for my hardware project 09:30am * Downix/#tunes knows the progress of Tunes, which is why Tunes isn't even a consideration. I need something availible by the end of 2000 :) How old is tunes? 5 years? more? 6 by my count Of course you're not much worse than HURD. 15 years to get to 0.4 anyone feel in the mood to pull a linus torvalds and get tunes moving? :) Fufie: Not I, I don't have 10 years for a viable OS * Downix/#tunes remembers the 3 years from 0.01 to 1.0 Linux development sucked but Linux developed yoo: Actually I thought Linux development far superior than the alternative. Coders do better work when millions of their peers will be looking at the code and a committee of one gets things done * Downix/#tunes nods I've got the source to a closed source OS, and OMG is it impossible to figure out Downix: Linux has had one of the worst bug histories of any OS Surpassing even Windows NT yoo: and one of the best bug corrections records of any OS. And the best record for speed of eliminating bugs yoo: It's not the number of bugs that matter, is how fast they're fixed when uncovered 09:40am but was WinNT developed in someone's spare time? Nope, it was a 40 hour a week, 400 person job Linux, in spare time, progressed at a better pace than Windows did Remember, WinNT's core section was originally developed in 1986 for IBM's OS/2 project Downix: Thats a tricky statement Downix: It's not entirely tru yoo: True. There's parts of OpenVMS in there as well Well, the senior designer of WinNT came from Digital (From the OpenVMS team) yoo: Yes he did. (I met him once) My father worked for Digital But there's sections of both in there Hell, WinNT is OS/2 1.x compatable WinNT 3.51 was a decent OS, too bad M$ fsck'd it up witn 4.0 My mother worked for Digital So? Which division/which office? Hmm, in Ma, not sure what office. The one on Route 128? 09:50am Anyways Linux's development was not what I'd call fantastic, but it worked Pragmatic viewpoint here, it works Sort of. Well yoo, I would note that Linux's bazzar-environment has even surpassed other OS kernel development done in a more structured manner, such as BSD Why do you say that? BSD has a very advanced kernel Yes, but it's behind Linux in many ways the free BSD's for example, just don't go out of their way to advertise. I use FreeBSD and OpenBSD Downix: Unfortunately that's not quite true. FreeBSD still outperforms Linux in networking, filesystem, and MM. So you're saying BSD can handle up to 32 CPU's on an UltraSPARC? performance and advanced are two different things Linux kernel 2.3 has full USB support, BSD's USB support is hideously crude BSD supports 8+gb of memory Does Linux? Yes up to 64GB what I lacked with OpenBSD was documentation and tutorials Agreed Fufie I am basically a developer and not a sysadmin and would've liked some tutorials on how to set up a secure system with OpenBSD.. Fufie: *nods* Of course I'm using OpenBSD for my firewall server and FreeBSD for my web servers, I know their stability FreeBSD is faster and stabler than Linux on equal supported platforms. That's what matters to me. I have a 486 that I want to use as a firewall with OpenBSD, but I don't have time to go through man-pages and mail-acrchives to find docs Fufie: I know how to set up OpenBSD as a firewall but only because I have friends who showed me how However, with the International patch for Linux, Linux's kernel is 99% as secured as OpenBSD's OpenBSD has no documentation because it's surrounded by an elitist atmosphere. Downix: That's a fscking huge 1% Linux's firewall support -blows- yoo: Not in 2.3 And security is not just in the kernel it's in the entire OS and GNU is not secure. yoo: Agreed. yoo: it is maybe elitistic GNU code also tends to be extremely poorly written. GNU code tends to be written Because while it's nice to say there are thousands of coders looking at the code -- it's not true. yoo: Not always true. Most core GNU apps are very nicely written. But in GNU, the best thing is... you can FIX IT! Fufie: Hurd? i think Hurd is the GNU Hurd, an advanced microkernel free OS Downix: My gosh! Downix: So untrue! gcc is crap. glibc 1 and 2 are crap. yoo: I said "core" hell, even emacs is pretty bad. Hurd is a bad example gcc is crap, I'll agree. Glibc needs a lot of help 10:00am are you trolling, yoo? :) Emacs is a victim of it's initial success Hurd is piss-poor, the basic file utilities are extremely poorly written and unstable. agreed with Yoo However EROS is GNU as well I could throw hundreds of examples at you. yoo: And I probably know them already How many -good- pieces of GNU software do you know of? yoo: But I wasn't discussing GNU or Linux for stability, but as a development model As a development model it's poor because of stability Ex: APIs yoo: EROS, Enlightenment, WindowMaker, ACTS, ln Downix: HAHAHAHAH Enlightenment? Windowmaker? your funny =) is WindowMaker GNU? Fufie: Has the GNU licence shipped with it last I knew Enlightenment is one of the most poorly written applications I have ever seen, rivalling even glibc. that's not the same using GPL and being GNU is very different yoo: Which E are you discussing? Fufie: True E the window manager yoo: But WHICH E the window manager? Every version Raster is a poor coder. they dumped their code base 2ce Interesting ideas, but he's a shitty coder. Mandrake is an excellent coder I wouldn't agree there either. and he's done the coding work for 0.16.x and the CVS version, for the most part yoo: Ok, he took a binary of 6 meg and shrank it down to 500k while adding features and making it more stable that is good coding IMHO, and I've looked over his work, it's damned good Fufie: Think that yoo is a bit biased? yes, because I disagree with the way to build stuff yoo: Well, if you know all about OS design, why don't you do a "linus" for the Tunes project? Fufie: I find open and honest is the best way. Let something stand up for it's own merits rather than for what a corporation says about it partly that, but I believe in a model that is open for everyone * Downix/#tunes nods even if you're a 14-year old or a 28-year old whiz who knows everything If I had any improvements for OpenBSD, I couldn't just submit the patch. OpenBSD is controlled by a core group the nice and liberating thing with linux imho is that everyone can participate, even if they suck as programmers Downix: E started as fvwm -- a binary of 100k yoo: Not the new E Downix: And E binaries were so large due to static linking of libraries and image data there is no fvwm code Downix: THAT is why it was 6mb GNU gets code out the door that works barely but that code never gets fixed because programmers are used to working around bad code and by letting people share code and have review they learn a lot bore than they would've done otherwise and we end up with better hackers Fufie: You often times get the diamond in the rough ex: libc5 dirent bug Several major pieces of software worked around a bug in glibc 1, when that bug was fixed (something like a year after it was introduced), it broke all the application sthat worked around it yoo: yes.. that is a problem.. the source code access lets you patch things easily and may stop people from making good dynamic systems * Downix/#tunes nods GNU has put it's stamp on much lousy code.. like GNU Common Lisp The glibc developers saw that everyone had patched their programs after a month and just decided to let it sit. Caused a much bigger problem then coding it right in the first place. 10:10am agreed Fufie Fufie: Because GNU code is more concerned with getting something piss poor out the door but GNU has made a lot of code available.. that's the important part it's like the Berkeley vs MIT coding debate Get a good product out the door late, or get a poor one out the door early. where would linux or open source or alternatives to Microsoft be if everyone had small elitist groups that hated all other code? Right you want something workable first But open source isn't necessarily an alternative to microsoft because it can not replace microsoft somebody else can use it they add to it there is just not enough -quality- software. someone else adds to it someone rewites.. and so on yoo: I didn't like the way that gcc 2.95.2 handled a lot of optimizations, so I coded in my own fix, now gcc 2.95.2 produces much faster code. rewrites yoo: if you want quality, FIX IT! open source is an alternative to microsoft for many things but not all Downix: Why? Why should I be condemned to picking up after other lazy people? yoo: Because sitting there bitching about it when you can fix it shows that you are as lasy as the bad coders because the alternative is the Microsoft model Downix: Ah, but gnu code is so crappy that I don't want to go within 100 ft of it. yoo: Nah, you're just too lasy downix: To fix a simple bug it can take hours of hacking and trying to understand crappy code. have a hard time getting around? yoo: Takes me only 15 minutes to an hour to fix bugs Downix: Maybe you want to help the f-cpu guys fix their bugs then because they've been working on gcc for days I've been helping them Wow. Then it should have been fixed last week? then it'll take awhile to get done, we're making a whole new architecture for gcc I mean, after all, someone as incredibly brilliant as you, who can consume hundred-thousand-line programs in 10 minutes, wow. yoo: 15 to an hour for a typical gnu program. Since 90% of GNU software is under 2000 lines.... at least 90% I have on my system small things like ls, df, etc. erm, 20,000 Suuure. yoo: Grab the source to gnu binutils sometime I'd rather not go blind. yoo: But again, if you know so much, why not get Tunes actually started? yoo: Tunes is like GNU ya know, making a complete system hmm.. pj harvey's 'is this desire' really rocks Fufie: What is it about? I didn't listen to the text.. pretty tired.. Downix: Because I don't think Tunes is ever going anywhere. yoo: Honestly, I agree with you there. ok because someone needs to pull a linux linus 10:20am and yoo is obviously the best coder :) No...Tunes problem is that its going to be impossible to code * yoo/#Tunes kicks Fufie sorry * Downix/#tunes isn't a coder, he's a bug fixer but it has been a lot of talk and criticism Too much. GNU would be fine if more coders just gave a damn about their code. But they don't. It's more important to get a product out the door soon than to get a good product out the door later. in a way that's true yoo: At least they have a product. but assume you would like some feedback and review.. how should you approach then? "hi, I have a secret project which some of you should have a look at, but don't tell anyone"? Not quite. Just don't rush a product out the door. Take your time when your coding. yoo: Not always possible. They needed to fill a hole, planning on fixing it later. But later never came. one way to get that is to get out something workable and let the "market"-forces find new uses and bugs for you andf get the feedback from the 2-3 people out of 100 If you are faced with a decision like 'implement this routine quickly in 15 minutes hackishly' or 'implement it in an hour beautifully', take the latter. Later never comes! yoo: Right, that's what a lot of people don't get If you don't do it now, you will never do it. Yes. They think 'I will do it in 15 minutes now, and later i will do it right' yoo: See why I've spend 27 months on my hardware, and have at least another year if not 2 before release? Downix: Bravo, then. yoo: But now I'm at the point where VC's want me to rush something out the door, which is a pain in the ass usually it is 15 minutes now or 3 days Fufie: if not a week. Thinking something completely through takes time Hell, implimenting the USB bus on my hardware took 2 months A *simple* interface by any stretch But you can be almsot certain that the piece of software you design will work well. yoo: You can with hardware too, if you think it through Yes. yoo: Rushing either hardware or software is bad. Look at the Pentium as an example And its worth it IMHO, because you won't have to put up with strange bugs caused by a lot of faulty code. know the number of bugs that damned chip has? Way too many. yoo: I agree there. yoo: But now I'm at the point that I need an OS for my hardware, so here I am Hmm. I thought the whole point of GNU should have been 'there is no pointy haired boss standing over your shoulder, so do what YOU want and do it RIGHT!' yoo: That is the point of GNU. The problem seems to be ego People who code products refuse to admit that they coded stuff badly, and noone else is willing to step on their toes Emacs is an ideal example I figured out once it would only take 8 months to completely re-code Emacs into a smooth machine Because whenever one threatens to step on their toes, others step in and push the 'stepper' back. right yoo 'Oh but he's doing it for the community, he doesn't have to' That's the problem There needs to be a level of responsibility associated with volunteer work, and people don't acknolwedge it. * Downix/#tunes nods 10:30am I personally would pay for someone to code up the GNU products correctly 'I'm doing it for free therefore I am entitled to do a half-ass job' Well, anyways * Downix/#tunes does need to get on his own subject Yeah. Your board is MIPS, is it not? Correct Hm. I want to get my hands on an entry-level MIPS protoboard. We will not be ready for awhile And the MIPS is inside of our chipset, not an external CPU not a normal protoboard by any means -:- SignOff yoo: #TUNES (Ping timeout for yoo[user-38lc6fo.dialup.mindspring.com]) Was he here just for arguements sake? *shrug* oh well -:- yoo [ultima@user-38lcn6l.dialup.mindspring.com] has joined #Tunes doh * Downix/#tunes hates that he can't find an OS adequate Your board does not use MIPS as an external CPU? What are you doing? VM machine? hey wow that was redundant We put the CPU inside of our control chips Plus allowed for the addition of external CPU's Ah So, how will it operate? 10:40am that's a hard question to answer, in what way do you mean? * Fufie/#tunes yawns.. maybe someone is doing something fun on the telly Well, if the CPU is internal? yoo: Do you mean how does the architecture work, how will the API function? Howsabout D) all of the above It sounds like your making a nifty VM or something Ok, it will be using the external CPU's as targets, not main processing centers *nods* Are the external CPU's mips? MIPS, but Alpha's and Athlons would not be a problem to support on the external bus with a few minor modifications hmm What is your target market? businesses looking to produce media-empowered machines Hey wow That's nice and vague =) like companies looking for an Internet Appliance that is not crippled, or a PDA that delivers on the primices made for palm units, or companies looking for media production machines, like the Amiga Also you can use the same chipset in a peripheral, like a video card Why the fsck am I so good at losing important things? hrm? 10:50am The problem is we made a chipset that is too flexible We can't hope to market it ourselves to consumers so we're instead hoping to licence the architecture and sell the chips to companies that know their consumer marketplace too bad suits are sometimes needed I know yoo: what kind of projects do you work on? ok, have to be going Well I work on finding important addresses on post-it notes I lost. a worthy project -:- SignOff yoo: #TUNES (Read error to yoo[user-38lcn6l.dialup.mindspring.com]: Connection reset by peer) Anyways 11:00am Anyways, the split between OS and app CPU's is the barrier that i have to work with * Fufie/#tunes yawns I can't believe that someone just suggested what they just did in #f-cpu that I shouldn't go for performance (seperate OS and App CPU) I should instead go SMP, to make sure that incase one CPU fails the others can take over I keep wondering... how often does his CPU fail on him? 11:10am -:- AlonzoTG [Alonzo@207-172-49-209.s209.tnt7.lnhva.md.dialup.rcn.com] has joined #tunes Hey Alonzo! om? pentiums don't boot ;) nope then the other one can fail to boot too at least according to balbasar they don't balance, symmetry and everything :) Personally I don't think SMP is worth it, master/slave makes more sence -:- lar1 [LarMan@dialup-209.245.136.70.SanJose1.Level3.net] has joined #tunes Hey -:- SignOff Downix: #TUNES (BitchX: now with 38 percent fewer calories!) -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) -:- SignOff lar1: #TUNES (Leaving) 11:20am -:- Downix [down@d-gnaps-286.ici.net] has joined #tunes boring out there in RL? no, got my lunch 11:50am -:- yoo [ultima@user-38lc6b5.dialup.mindspring.com] has joined #Tunes -:- SignOff Downix: #TUNES (BitchX: more nutritious than a six-pack.) -:- SignOff zarq: #TUNES (Ping timeout for zarq[9dyn25.delft.casema.net]) -:- eihrul [lee@usr5-ppp73.lvdi.net] has joined #tunes -:- hcf [nef@me-portland-us327.javanet.com] has joined #tunes -:- rinky [remdutta@giascla.vsnl.net.in] has joined #tunes hi there hi is everybody here into OS development? hello? no.. I like language crafting better oh....ok :) maybe you can explain lex & yacc to me sometime..... always wanted to know something about them lex and yacc are for implementing, not crafting ahhh.. there are other people here -:- RUMPOLE [rumpole1@212.49.245.65] has joined #tunes lex and yacc aren't my favourites so I might guide you in another direction though no....just making smalltalk... :) hi anyone know where i can get hold of a certain mp3? read the topic please... is anybody here _into_ OS develop? lots... lots what -:- RUMPOLE [rumpole1@212.49.245.65] has left #tunes [] bright chap hello? fufie: what exactly is language "crafting"? making/extending/designing languages anybody wonder why bash _still_ doesn't have an "except"? Fufie? can't you write one? um.....nope! i've never seen its source actually and if it's so easy then why hasn't it been done? maybe because it is easy? 01:40pm i don't get you what is "except" anyway? so you asked me if I could write one without knowing what "one" was? except (*.sys) dir yes, that was appropriate like that 4dos has had it for years so what does except do? the same as 'grep -v' ? -:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes -:- mibin [mibin@a-ii2-57.tin.it] has joined #tunes whats grep -v ? except EXCEPTS grep -v shows the lines that do not match echo l|grep -v l no... except works on whatever command except (*.exe) del *.* get it? hmm, i dont it works on what? what is that except command the output or the input or both? output of what? it only works on filenames read it as english: except (*.exe) del *.* i seems we are having an hard time understanding each other so it works on filename-expansion only? mibin: u must be thick heh fufie: actually come to think of it,thats all *I've* ever used it for but I'm sure it has more uses but whatever lets leave this dicussion aside let me ask what i came here to ask rinky: well.. do the expansion through find and then use egrep or some fancy regexp to remove matches.. a function basically what do all of you know about writing bootsectors? that is to say bootstrap loaders rinky: dd bs=512 count=1 /dev/hda :P