From: Digestifier To: Linux-Misc@senator-bedfellow.mit.edu Reply-To: Linux-Misc@senator-bedfellow.mit.edu Date: Tue, 24 Aug 93 05:13:09 EDT Subject: Linux-Misc Digest #48 Linux-Misc Digest #48, Volume #1 Tue, 24 Aug 93 05:13:09 EDT Contents: Re: Why use linux was Re: Why would I want LINUX? (Charles Hannum) Re: SCSI Performance (Keith Smith) Re: NT versus Linux, the updated and expanded comparison chart (Keith Smith) Re: Wordprocessor under X (Howlin' Bob) Re: Why would I want LINUX? (Tim Villa) *** PLEASE READ THIS BEFORE POSTING *** (misc-2.00) (Ian Jackson) Re: Toshiba 3401 CD-ROM (Keith Smith) Re: Does the Adaptec 1542C have problems? (Drew Eckhardt) Re: Does the Adaptec 1542C have problems? (Drew Eckhardt) TCP/IP traffic/controll Program (Daniel T. Schwager) Re: NT versus Linux (Kevin Brown) Re: NT versus Linux, the updated and expanded comparison chart, version 2 (Kevin Brown) Re: /dev/tty (Whatis it for?) (From c.o.l) (Petter Reinholdtsen) Re: SLIP distance contest. How far have you connected? (Julian Cowley) ---------------------------------------------------------------------------- From: mycroft@trinity.gnu.ai.mit.edu (Charles Hannum) Crossposted-To: comp.os.386bsd.misc Subject: Re: Why use linux was Re: Why would I want LINUX? Date: 24 Aug 1993 03:09:42 GMT Some serious comments on your comparison; no flames intended. In article <1993Aug23.124406.856@finbol.toppoint.de> jschief@finbol.toppoint.de writes: BSD is quite good, but you need more knowledge of unix/bsd I don't really believe this. Linux also requires a fair amount of Unix-specific knowledge to configure. For the most part, the default installation works (as it does in NetBSD also), but it is not enough to actually run a system well. You need to configure the network, sendmail, getty (for serial ports at least), maybe SLIP and/or PPP, etc., etc. for many people Linux is easier to install, The NetBSD installation process is pretty simple, though it's not what I ultimately want. I rather like the SunOS and Ultrix installations, and a similar installation system is being developed for NetBSD. easier in setup, See above. easier in building new kernels, Meaning it's interactive? BSD in general is much easier to configure after you learn how to read and write config files. And it's pretty straightforward. I config'd a SunOS kernel for the first time without consulting any man pages; I just looked at the existing config files and did the obvious thing. easier to get drivers for exotic hardware and ... It is true that there are some pieces of hardware which Linux drivers but not NetBSD drivers exist for. Fortunately most of them aren't very popular (or perhaps that's why we don't have drivers for them yet), so this is not a major drawback. If I had any of them, I would endeavour to rectify the situation. (Consider that a hint to anyone with more money than me.) faster in development (this includes : find & produce bugs) That's not clear. There is a fundamental difference between Linux and NetBSD in this, however; while Linus and others involved in Linux (Peter MacDonald, the authors of various packages, etc.) are willing to make new releases very frequently, there is little organization. In my experience, very few users actually want to update their systems a few times a month (or for that matter, possibly a new version of some program or other every single day). We (the NetBSD group) endeavour to organize more coherent releases. In particular, we like to make sure everything at least minimally works; SLS releases have often contained non-functional programs. For people who want frequent updates, we have our current source tree available by FTP and SUP. The publicly available sources are updated nightly from our working CVS tree, and usually work a bit better than the releases. (The work doesn't ever stop; we've made many changes even since the 0.9 sources were pretty much frozen.) and easier to get (SLS, MCC, SLA....) NetBSD is not as many FTP sites, but it is certainly not hard to get, for anybody on the net, at least. By the time 1.0 is released, there will be a CD-ROM distribution. One serious advantage of NetBSD for me is that it runs on multiple platforms. I now have a HP 370 (68030-based machine) happily running a complete NetBSD system, and it's actually significantly faster than my 386. There are several other ports in progress, some of which are running. Hopefully we will have a few of them in our common source tree for the 1.0 release. ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: Re: SCSI Performance Date: Tue, 24 Aug 93 02:48:55 GMT In article eric@tantalus.nrl.navy.mil (Eric Youngdale) writes: >In article haba@snakemail.hut.fi (Harri "Haba" Suomalainen) writes: >>In article eric@tantalus.nrl.navy.mil (Eric Youngdale) writes: >>> Try using something larger than 512 byte blocks. Your performance will >>>increase significantly if you were to do this. Try using something in the >>>4096-8192 byte range. As I recall, the performance starts to level out >>>somewhere in this area. >> [...] >>Or did you mean one should write/read in larger chunks? I think >>over 1024 bytes sizes make a very small difference even that way.. > > I meant that you shoud run iozone with reads and writes blocked in >something other than 512 bytes. It does make a big difference as you get up >towards 8192 bytes. > > There was a second misconception about iozone posted here in someone >elses note. This had to do with the notion that when you write you merely load >up the buffer cache, so that when you start reading again you are merely >reading back from the buffer cache again. It is ESSENTIAL that you choose the >size of the file that iozone writes so that it is larger than the physical >memory on the machine you are running it on - the rule of thumb is that it >should be 2.5 times the physical memory available, and this will guarantee that >you are not measuring the speed of the buffer cache. Look, run an 'iozone auto' and THEN do the comparison. A 30 Second test in a multi-user OS doesn't show shit. 'iozone auto' Will run files of like 1,4,8,&16MB with block sizes of 512 1K, 2K, 4K, and 8K, back to back. *THEN* you will have something to compare with. -- Keith Smith keith@ksmith.com 5719 Archer Rd. Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201 Somewhere in the Styx of North Carolina ... ------------------------------ Crossposted-To: comp.os.ms-windows.advocacy From: keith@ksmith.com (Keith Smith) Subject: Re: NT versus Linux, the updated and expanded comparison chart Date: Tue, 24 Aug 93 03:00:14 GMT Besides the fact that this is total horse shit ... For Immediate Release: NT versus Linux, a feature comparison ============================================================ Feature NT Linux ======= == ===== >runs DOS apps yes yes, via emulator Mostly > >runs Windows apps yes not currently. Under About 75% SLOWER Than native Windows According to Info-World OS-2 Runs Win APS about 11% slower. So basically if you have 4 times as much time to kill. Oh, And I wonder if this has anything to do with that "MicroKernel" feature? -- Keith Smith keith@ksmith.com 5719 Archer Rd. Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201 Somewhere in the Styx of North Carolina ... ------------------------------ From: gt8134b@prism.gatech.EDU (Howlin' Bob) Subject: Re: Wordprocessor under X Date: 24 Aug 93 03:43:59 GMT sdh@fishmonger.nouucp (Scott D. Heavner) writes: > I tried it and had it running fine, but I can't get it to save >files (this isn't really where I expected the problem to lie). It acts like >it's saving them, but when you load them, they're just garbage. Some >word processor. If you want to do all your work in one sitting, you should >be ok. And you can load files saved under dos (not dosemu). > My config file is rather simple. The imporant options are (maybe?): >keybint off >timint on >RawKeyboard It's really annoying how Linux users have forgotten how to produce good bug reports. THIS OS IS UNDER DEVELOPMENT! YOU ARE NOT USERS, YOU ARE BETA TESTERS! Are you mounting a Linux partition via emufs? Are you trying to write to that partition? Send me the whole of your config file...unless you know dosemu inside and out, you can't possibly guess what options are "important." -- Robert Sanders Georgia Institute of Technology, Atlanta Georgia, 30332 uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt8134b Internet: gt8134b@prism.gatech.edu ------------------------------ From: tvilla@tartarus.uwa.edu.au (Tim Villa) Crossposted-To: comp.os.386bsd.misc,alt.os.bsdi Subject: Re: Why would I want LINUX? Date: 24 Aug 1993 03:15:17 GMT shawn@fenchurch.mit.edu (Shawn F. Mckay) writes: >Thats just not true, Linux makes no effort to run on a 286. If I have >to have a 386 to run, I'll have little trouble choosing. And IDE >drives are CHEAP and plentiful. > - Shawn And it should make no attempt to run on a 286. Apart from the fact that the things are brain dead, they are 16 bit processors which simply cannot support a 32 bit operating system. Tim ------------------------------ From: ijackson@nyx.cs.du.edu (Ian Jackson) Subject: *** PLEASE READ THIS BEFORE POSTING *** (misc-2.00) Date: Tue, 24 Aug 1993 04:23:01 GMT Please do not post questions to comp.os.linux.misc. If you have a question about Linux you should get and read the Linux Frequently Asked Questions with Answers list from sunsite.unc.edu, in /pub/Linux/docs, or from another Linux FTP site. In particular, read the question `You still haven't answered my question!' Then - only after checking the FAQ - should you post to comp.os.linux.help. Note that X Windows related questions should go to comp.windows.x.i386unix. The FAQ for this group is available on rtfm.mit.edu in /pub/usenet/news.answers/Intel-Unix-X-faq. Comments on this posting are welcomed - please email me ! -- Ian Jackson (urgent email: iwj10@phx.cam.ac.uk) 35 Molewood Close, Cambridge, CB4 3SR, England; phone: +44 223 327029 ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: Re: Toshiba 3401 CD-ROM Date: Tue, 24 Aug 93 03:50:59 GMT In article jmk@kksys.com (Jim Kaufman) writes: > >Has a driver been written to use a Toshiba CD-ROM as a CD player. >(Kind of along the lines of what's been done for the Mitsumi?) Uh, the 3401 is a SCSI CD-ROM. Just plug it in on a supported controller and it works just fine. Now the TEXEL is another frigging story. -- Keith Smith keith@ksmith.com 5719 Archer Rd. Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201 Somewhere in the Styx of North Carolina ... ------------------------------ From: drew@caesar.cs.colorado.edu (Drew Eckhardt) Subject: Re: Does the Adaptec 1542C have problems? Date: Tue, 24 Aug 1993 05:03:04 GMT In article <25b0fj$og@crchh327.bnr.ca> minyard@crchh7b9.rich.bnr.ca (Corey Minyard) writes: >I had heard about problems with the 1542C and was wondering if the >problems still exist or are fixed. Some have, some haven't. The initialization code has been fixed to handle the 1542C correctly in most cases. However, some of the enhanced BIOS options scramble things in a way that is incompatable with the Linux driver, and some people have reported problems with Linux after - Using one of the BIOS auto-probe options - Using a BIOS extended translation - Using BIOS support for more than two drives >Is anyone using the 1542C successfully? Many people have been succesfull, as long as they haven't gotten into the more exotic BIOS options. -- Boycott USL/Novell for their absurd anti-BSDI lawsuit. | Condemn Colorado for Amendment Two. | Drew Eckhardt Use Linux, the fast, flexible, and free 386 unix | drew@cs.Colorado.EDU Will administer Unix for food | ------------------------------ From: drew@caesar.cs.colorado.edu (Drew Eckhardt) Subject: Re: Does the Adaptec 1542C have problems? Date: Tue, 24 Aug 1993 05:12:04 GMT In article <25bqdn$1er@indigo.imp.ch> ilg@imp.ch (Philippe Steindl) writes: >Hello, > >I think one should make this clear: > >Ther was NEVER a problem with the Adaptec 1542C. No, there were multiple problems. 1. The order of initializing things needed to change, meaning Adaptec 1542B drivers didn't work with the C revision. This bug hit Linux, NetBSD, Windows NT, and a large number of other systems. It was rapidly fixed. 2. People still report problems with a few of the BIOS options doing things that keep the Linux drivers from initializing correctly. In any case, a configuration change fixes things. >The signals of the C version has steeper flanks, so you need a better >SCSI cable if connectiong something externally and only if you want to >use a "round" cable. The "flat" SCSI cables used internally and, if you >want, externally can be as cheap as you want it (mine cost 6$). There's nothing that says a ribbon cable isn't going to have the same impedance problem as round cable. Also, a lot of round cable is shielded while most (I've seen shielded ribbon in a few cases) ribbon cable is unshielded, and will be prone to noise hits especially at 10Mhz FAST SCSI-II speeds (I've seen workstations that don't work because of parity errors when running the bus over 8M/sec). -- Boycott USL/Novell for their absurd anti-BSDI lawsuit. | Condemn Colorado for Amendment Two. | Drew Eckhardt Use Linux, the fast, flexible, and free 386 unix | drew@cs.Colorado.EDU Will administer Unix for food | ------------------------------ From: danny@dragon.stgt.sub.org (Daniel T. Schwager) Subject: TCP/IP traffic/controll Program Date: Sat, 21 Aug 1993 18:51:19 GMT Hi all, can somebody tell me, where i can find a program, which measures (log, filter, show, count...) the packets on a) IP Level b) Ethernet Level (raw ethernet frames) Thanks Danny -- ,,, (^ ^) ==================oOO==(_)==OOo======================= Danny ------------------------------ Crossposted-To: comp.os.ms-windows.advocacy From: kevin@frobozz.sccsi.com (Kevin Brown) Subject: Re: NT versus Linux Date: Tue, 24 Aug 1993 01:00:52 GMT In article <1993Aug23.080618.19456@colorado.edu> drew@kinglear.cs.colorado.edu (Drew Eckhardt) writes: >In article kevin@frobozz.sccsi.com (Kevin Brown) writes: >>In article <1993Aug21.013647.21842@colorado.edu> drew@romeo.cs.colorado.edu (Drew Eckhardt) writes: >>>In article <1993Aug20.184709.15303@mnemosyne.cs.du.edu> zevans@nyx.cs.du.edu (Zack Evans) writes: >>>>Are you absolutely sure that you want to do this? [ny] no, but I am going to >>>>anyway :) >>>> >>>>In article <930820201354.23255@world>, >>>>Brian Leary wrote: >>>> >>>>>Q: What does linux do when it runs low on memmory? A: IT FALTERS! The >>>>>only os I have seen that locks up when it runs low in memory. Every >>>>>linux user must constantly watch for memory useage, if linux needs a >>>>>few more kilobytes than are available, it doesn't complain, it just >>>>>crawls to a halt before dying a miserable death, taking down the whole >>>>>system with it. >>> >>>This is a false statement. >> >>From looking at the code, it was true of pl10. Here's the relevant code: > >pl10, two publicly distributed revisions and over half a dozen alphas >behind the current code. If you're allowed to critisize Linux based >on older versions, can I ream NT based on older software, say Windows 3.0? Absolutely. :-) Seriously, though, pl10 isn't really *that* old. But then, things move fast in the Linux universe. :-) [rest deleted] >Boycott USL/Novell for their absurd anti-BSDI lawsuit. | >Condemn Colorado for Amendment Two. | Drew Eckhardt >Use Linux, the fast, flexible, and free 386 unix | drew@cs.Colorado.EDU >Will administer Unix for food | -- Kevin Brown kevin@frobozz.sccsi.com This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end > This is your .signature virus on drugs: <> Any questions? ------------------------------ Crossposted-To: comp.os.ms-windows.advocacy From: kevin@frobozz.sccsi.com (Kevin Brown) Subject: Re: NT versus Linux, the updated and expanded comparison chart, version 2 Date: Tue, 24 Aug 1993 02:05:49 GMT In article buhr@umanitoba.ca (Kevin Andrew Buhr) writes: >In article kevin@frobozz.sccsi.com >(Kevin Brown) writes: >| >| Feature NT Linux >| ------- -- ----- > . > . > . >| kernel+necessary >| extensions to run ~500k+ 200k-500k > >This is why NT runs in a minimum 1 Meg configuration, right? 500K for >the system with 500K to spare for apps, eh? ;) Well, since I haven't seen NT run, I can't really say how big the whole thing is when loaded. So I'm being conservative. :-) >The NT "kernel + necessary extensions to run" is presumably somewhere >in the order of 7 or 8 Megs, or it would run comfortably in less than >12 Megs. That's probably true. However, there's data space to consider in the RAM usage. But I basically concur. >Kevin Buhr -- Kevin Brown kevin@frobozz.sccsi.com This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end > This is your .signature virus on drugs: <> Any questions? ------------------------------ From: petterr@stud.cs.uit.no (Petter Reinholdtsen) Subject: Re: /dev/tty (Whatis it for?) (From c.o.l) Date: Tue, 24 Aug 1993 08:31:13 GMT In article , astorey@NeoSoft.com (Alan Storey) writes: > >I believe that /dev/tty is a pseudo device used to open any free >tty (is this correct?). /dev/tty is current tty. -- ##> Petter Reinholdtsen <## | petterr@stud.cs.uit.no Q: What does a computer science graduate say to a humanities graduate? A: I'll have the burger and fries, please. ------------------------------ From: julian@uhunix3.uhcc.Hawaii.Edu (Julian Cowley) Subject: Re: SLIP distance contest. How far have you connected? Date: Tue, 24 Aug 1993 08:53:11 GMT In article ehhchi@staff.tc.umn.edu (Ed H. Chi) writes: >It's not 8000 miles. I was born in Taiwan, and that's only 6000 miles >from Minnesota. Actually little less than 6000 miles. This may be slightly pedantic, but he's right. Using spherical trig gives a distance of ~8900 miles. You can figure it out like this. For a spherical triangle, the law of cosines is: cos c = cos a cos b + sin a sin b cos C where a and b are two sides of the triangle, C is the included angle, and c is the side opposite C. All sides in a spherical triangle are measured by angles. So to cut the story short a bit, using 39 S, 145 E for the coordinates of Melbourne (these are really rough since I just interpolated them from an atlas), and 30 16' N, 97 45' for Austin (from a reference), you get: cos c = cos (90 - (-39)) * cos (90 - 30 16') + sin (90 - (-39)) * sin (90 - 30 16') * cos (-145 - 97 45') which works out to an angle of about 128 degrees. To find miles from this, you have to multiply it by the number of miles in a degree, which is 7926 mi * pi / 360 degrees = 69.17 mi per degree since the diameter of the Earth is ~7926 miles. This produces a result of ~8900 miles. >Australia shouldn't be 8000. Besides, going around the globe is around >12000, so half of the globe is 6000. Why would you go the other way which >is 8000 miles instead of 4000 miles?? Your figures are off by two. If you go around the world 12,000 miles in any direction, you automatically end up at the antipode of your origin, the point on the Earth exactly opposite you. 8900 miles is still hard to beat, though. For me to dial to the other end of the Earth, I'd have to dial up somewhere in Botswana, Africa. Anybody have any SLIP numbers to dial there? :-) -- -=- julian@uhunix.uhcc.hawaii.edu | echo "Every byte is sacred." > /dev/null ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Misc-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.misc) via: Internet: Linux-Misc@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux sunsite.unc.edu pub/Linux End of Linux-Misc Digest ******************************