From: Digestifier To: Linux-Activists@senator-bedfellow.mit.edu Reply-To: Linux-Activists@senator-bedfellow.mit.edu Date: Sun, 22 Aug 93 00:13:04 EDT Subject: Linux-Activists Digest #154 Linux-Activists Digest #154, Volume #6 Sun, 22 Aug 93 00:13:04 EDT Contents: Re: Linux Distribution? (Kevin Fluet) UnixChat (JL Gomez) Almost-diskless linux (Liem Bahneman) Unix for sale (Scott Barker) Re: SCSI Performance (Keith Smith) DOSEMU:0.49 doesn't work, won't compile. (Allen M. Ashley) Re: Almost-diskless linux (Byron A Jeff) Re: Almost-diskless linux (Jon Tombs) SLS boot disk and ne2000 type network cards (Brian Quandt) Re: DOSEMU:0.49 doesn't work, won't compile. (Howlin' Bob) Dat driver (FMGTNG Student9) Re: Ctrl-Alt-Del with SLS 1.03 (Frank Lofaro) Re: Linux CD-ROMs in general (Charlie Lear) LILO and disk geometry: major puzzlement! (CHRISTIAN D. ARMOUR) Unix for sale (David W. Summers) Re: /dev/tty (Whatis it for?) (Oliver Paulzen) Re: SCSI Performance (Yet Again) (Remco Treffkorn) Setting up bi-dir uucp (Gary Kline) Re: What's the FASTEST FS? (Bruce Evans) Re: Ethernet Cards informations needed (Donald J. Becker) Re: Serial: My IRQ is 5 NOT 4! Change? (Craig T Manske) ---------------------------------------------------------------------------- From: user1@valis.ampr.ab.ca (Kevin Fluet) Subject: Re: Linux Distribution? Date: Fri, 20 Aug 1993 19:59:51 GMT In <1993Aug18.003949.3568@bcars6a8.bnr.ca> markem@bcarh10d.bnr.ca (Dave Mielke) writes: >Anybody know if their is an FTP site with one big tar-file containing the >full Linux Distribution? I believe that tsx-11 allows you to do a 'tar SLS' (in the /pub/linux/packages dir) and get the whole 36 Megs as one big tar file. Maybe it was sunsite. I did it once about 8 months ago, before bode.ee.ualberta.ca started mirroring SLS. -- Kevin /--------------------------------------------------------------------\ | Kevin Fluet Call V.A.L.I.S. Public Access Linux | | user1 or kevin@valis.ampr.ab.ca (Down for the Summer) | | fluet@ee.ualberta.ca Ask me about Linux, the FREE Unix clone! | \--------------------------------------------------------------------/ ------------------------------ Crossposted-To: comp.os.linux.help From: gomez@enuxsa.eas.asu.edu (JL Gomez) Subject: UnixChat Date: Sat, 21 Aug 1993 18:45:22 GMT Has anyone got UnixChat as posted in alt.sources to work under Linux? Thanks for your patches! ------------------------------ From: roland@cac.washington.edu (Liem Bahneman) Subject: Almost-diskless linux Date: 21 Aug 1993 20:28:37 GMT I've been trying to work on a rahter interesting problem: Say I've got 10 386's with only a single high-desnsity floppy drive, and I wanted to nfs mount an entire operational filesytem from a single linux server, would it be possible with a single bootdisk? (Basically I'd like to run everything on the pcs via nfs.) Is this possible? Each machine has it's own ip address si I'd like to have all the abilities, like telnet, ftp, etc. I guess it'd be like these computers were dumb terminals, but not so dumb, as they would do all the work instead of the single server doing everything. Has anyone had any success with such a situation? Someone posted long ago about diskless nfs booting, if they could respond would be much appreciated as I've lost their post. - liem -- =======[roland@cac.washington.edu]=====[The Last Gunslinger]================== Outside of a dog, computers are a man's best | UCS Consulting friend, inside a dog it's too dark to type. | UW Ice Hockey Co-captain =============================================================================== ------------------------------ Crossposted-To: comp.unix.misc,comp.unix.sys5.r3,comp.unix.sys5.r4,comp.unix.sys5.misc,comp.unix.pc-clone.16bit,comp.unix.pc-clone.32bit,comp.os.linux.misc From: barker@enel.ucalgary.ca (Scott Barker) Subject: Unix for sale Date: Sat, 21 Aug 1993 20:29:05 GMT Reply-To: barker@enel.ucalgary.ca I'm selling my DELL Unix SVR4, Issue 2.2. Included is: Installation disks (3.5 and 5.25) Installation tape (250M QIC and 2G DAT) Unlimited user upgrade "Working with Dell Unix" Release notes, etc. This all costs about $1700 (Cdn) from Dell ($1000 (Cdn) if you have them install it on a Dell purchased machine). I'm asking $900 (Cdn) OBO. I'm selling because I need the cash (DELL Unix works wonderfully, it hasn't crashed on me yet - I've been up since December 1992). Interested parties can e-mail me at barker@enel.ucalgary.ca -- Scott Barker barker@enel.ucalgary.ca ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: Re: SCSI Performance Date: Sat, 21 Aug 93 19:26:47 GMT In article <1993Aug19.031941.5586@jonh.wimsey.bc.ca> jhenders@jonh.wimsey.bc.ca (John Henders) writes: [SCSI vs. IDE Performance] > A me too here as well. My Conner ide actually gets slightly better >times than my Maxtor scsi drive. This is also tru under DOS. The numbers >are 650 for scsi, 725 for ide. Neither even comes close to my old Atari >ST figure of 1150 for a Quantum lps52. OK, Now, Hook up 2 FAST SCSI disks and 2 IDE Disks and run the same test simultaneously on both disks. Next, Run a background job that EATS the CPU at the same time with only 1 of each and post your results. -- 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: ashley@cco.caltech.edu (Allen M. Ashley) Subject: DOSEMU:0.49 doesn't work, won't compile. Date: 21 Aug 1993 20:28:10 GMT I tried posting to comp.os.linux.help but got no response. Sorry to keep this channel open. It would seem that the dosemu0.49 should run out of the box since the /etc/dosemu/config file is interpreted at run time. Not so. After loading up diskimage I get a never ending sequence of: SYSCALL ERROR:38 function not implemented in file mutex.c expr=semop(sem,semoa,1) line 52 or 79. I tried several tweaks of the config file and I do get the time/date prompt but it ends there. When I do compile dosemu0.49 by: make config, make dep, make install I get an ld error since the lib ipc can't be found( this is SLS 1.03, the latest on tsx-11). I read an earlier post that I should just drop the -lipc from the ld statement, and the make install goes through. However the emulator then just hangs, as if it were waiting for input and never sees the keyboard. I got the earlier version of dosemu to run, but this one seems to be a giant step backward. Any clues for the clueless? ------------------------------ From: byron@cc.gatech.edu (Byron A Jeff) Subject: Re: Almost-diskless linux Date: Sat, 21 Aug 1993 22:11:33 GMT In article <2560hl$b5h@news.u.washington.edu>, Liem Bahneman wrote: > >I've been trying to work on a rahter interesting problem: > >Say I've got 10 386's with only a single high-desnsity floppy drive, and I >wanted to nfs mount an entire operational filesytem from a single linux >server, would it be possible with a single bootdisk? (Basically I'd like >to run everything on the pcs via nfs.) Is this possible? >Each machine has it's own ip address si I'd like to have all the abilities, >like telnet, ftp, etc. I guess it'd be like these computers were dumb >terminals, but not so dumb, as they would do all the work instead of the >single server doing everything. Has anyone had any success with such a >situation? Someone posted long ago about diskless nfs booting, if they >could respond would be much appreciated as I've lost their post. It's definitely possible. I mean the distribution floppies can mount and run programs via NFS. I also read that NFS has been fixed so that shared libs can be done. So build a boot/root floppy (or use a distribution one) and mount /usr and /etc from the server on boot. You may want to to set up a link from /etc and /usr to somewhere local on the floppy. A couple of issues: - Swap. You think HD is slow wait til you try over the network. - Root on floppy. Very slow. You could put it an a RAMDISK. How much memory do these 386's have? You'd probably not want to have to swap and you would like to have a RAMDISK root. 8 is good, 12 is better, 16 or better is best. But for the cost of 16 Meg of memory you can just buy a 340 (or much smaller) Meg HD. It may be usable with 8 Meg of memory and a RAMdisk root. But I'd test it first. Hope this helps, BAJ --- Another random extraction from the mental bit stream of... Byron A. Jeff - PhD student operating in parallel! Georgia Tech, Atlanta GA 30332 Internet: byron@cc.gatech.edu ------------------------------ From: jon@robots.ox.ac.uk (Jon Tombs) Subject: Re: Almost-diskless linux Date: Sat, 21 Aug 1993 21:50:26 GMT In article <2560hl$b5h@news.u.washington.edu> roland@cac.washington.edu (Liem Bahneman) writes: > >I've been trying to work on a rahter interesting problem: > >Say I've got 10 386's with only a single high-desnsity floppy drive, and I >wanted to nfs mount an entire operational filesytem from a single linux >server, would it be possible with a single bootdisk? (Basically I'd like >to run everything on the pcs via nfs.) Is this possible? >Each machine has it's own ip address si I'd like to have all the abilities, >like telnet, ftp, etc. I guess it'd be like these computers were dumb >terminals, but not so dumb, as they would do all the work instead of the >single server doing everything. Has anyone had any success with such a >situation? Someone posted long ago about diskless nfs booting, if they >could respond would be much appreciated as I've lost their post. Get my networking root/bott disks from ftp.robots.ox.ac.uk in /pub/linux/ox-usr/rootdisk. This is exactly what I have used it for. Currently I've given up with putting the kernel on the root disk and use a separate boot and root disks. If you remove scsi, math emu, support for all but one ethernet card etc, then you can still fit a kernel and all the networking binaries you need on a single floppy. Previous to linux-0.99.pl12 the lack of nfs mmap made this pretty pointless, but with my nfs mmap code in pl12, you can indeed run X on a networked PC without touching the hard disk. With the kenel removed there is even enought /tmp space to use gcc (but with $TMPDIR on nfs it is faster). Jon. ------------------------------ From: quandt@cs.umr.edu (Brian Quandt) Subject: SLS boot disk and ne2000 type network cards Date: Sat, 21 Aug 1993 22:04:21 GMT The SLS boot disk (a1.3) locks up with my system when I have my network cards installed (a ne2000 compatible). The version of the boot disk is the one current as of this message. The point at which this locks up is the 8390.c test. I remember reading a while back about this problem, and the fix was to remove the card, recompile linux with only the ne2000 support. However, I don't like this, I was hoping not to have to rip my machines apart when installing linux-SLS for the first time. Any better ways, like is there a real fix out there yet? (Oh to clarify, yet removing the card recompiling and does solve the problem). Brian Quandt ------------------------------ From: gt8134b@prism.gatech.EDU (Howlin' Bob) Subject: Re: DOSEMU:0.49 doesn't work, won't compile. Date: 21 Aug 93 23:33:03 GMT ashley@cco.caltech.edu (Allen M. Ashley) writes: >run out of the box since the /etc/dosemu/config file is interpreted >at run time. Not so. After loading up diskimage I get a never ending >sequence of: >SYSCALL ERROR:38 function not implemented in file mutex.c >expr=semop(sem,semoa,1) >line 52 or 79. This seems to say that your kernel doesn't have SYSV IPC support compiled in. What version of the kernel do you run? I recommend 0.99pl12 *with* SYSV IPC enabled. >I got the earlier version of dosemu to run, but this one seems to be >a giant step backward. Any clues for the clueless? Uh, thank you for the confidence. Good luck. -- Robert Sanders Georgia Institute of Technology, Atlanta Georgia, 30332 uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt8134b Internet: gt8134b@prism.gatech.edu ------------------------------ From: student9@strider.sps.mot.com (FMGTNG Student9) Subject: Dat driver Date: Sat, 21 Aug 1993 22:36:34 GMT ------------------------------ From: ftlofaro@unlv.edu (Frank Lofaro) Subject: Re: Ctrl-Alt-Del with SLS 1.03 Date: Sun, 22 Aug 93 00:17:52 GMT As for things not getting reset when one would expect... If I set my USRobotics Sportster 9600 to not hang-up on DTR loss (usually I set it so DTR drop brings me to on-line command mode*), I can press the reset button, and stay logged in!. Weird, I know. I guess the reset line doesn't reset everything..... * P.S. AT&D1 is the command for this. It is sort of undocumemted. ------------------------------ From: clear@actrix.gen.nz (Charlie Lear) Subject: Re: Linux CD-ROMs in general Date: Sun, 22 Aug 1993 00:34:54 GMT In article <1993Aug19.172528.15787@valis.ampr.ab.ca>, Kevin Fluet wrote: > That Yggdrasil group seems to have disappeared, too (at least I haven't > heard anything about their final release CD). The beta CD I have doesn't > have an image of tsx-11 or sunsite, so now that the version of linux on it is > old, the disk is kind of useless. I've had no replies to my email, in fact the last communication I had from them was a couple of months ago asking for contact details so that I could be listed in the 1.0.x documentation as a distributor. I have not sold any Yggdrasil CDs since running out of stock around six weeks ago, and the people who have ordered since then have been advised that I'm not supplying until a new version arrives. According to Walnut Creek, the new Yggdrasil CD is expected at the end of this month, but "be prepared to be flexible about it." -- Charlie "The Bear" Lear clear@actrix.gen.nz Ph/Fax +64 4 564-5307 Shareware Distribution NZ PO Box 2009, Wellington, New Zealand NZ Distributors of Walnut Creek CD-ROM Products DoD #221 ------------------------------ From: cs911461@ariel.yorku.ca (CHRISTIAN D. ARMOUR) Subject: LILO and disk geometry: major puzzlement! Date: Sun, 22 Aug 1993 01:16:47 GMT Ok, here's the long and the short of it: I have a crappy old 40Mb MFM drive onto which I have already installed a previous version of linux. LILO was able to make the drive bootable at one point but recently I tried again with the pl 12 kernel and LILO was no longer able to boot everything up. At first, LILO stalled after printing out "LIL-", then after some hacking around, it printed out just "LI". Yuck. So this README tells me that in both cases, the disk geometry is wrong. It also says something about "moving" boot.b and map, which I don't understand. (BTW, where can I find the technical overview as mentioned in the LILO README?) Anyway, the important thing here is that on line 643, there is talk of a DOS program called DPARAM.COM. I tried looking for this program but was unsuccessful. If anyone could provide me with an ftp site and directory, I'd be grateful. Could it also be possible that LILO is not getting the correct disk geometry from the kernel? Considering this is an MFM drive and all... cak Geekium Rex ------------------------------ Crossposted-To: comp.unix.misc,comp.unix.sys5.r3,comp.unix.sys5.r4,comp.unix.sys5.misc,comp.unix.pc-clone.16bit,comp.unix.pc-clone.32bit,comp.os.linux.misc From: dws@cseg03.engr.uark.edu (David W. Summers) Subject: Unix for sale Date: Sun, 22 Aug 1993 02:01:32 GMT Xenix for sale. Nothing wrong with it. I used it for years! I upgraded to Linux. I have two Xenix 286 for sale and two Xenix 386 for sale. Please e-mail me for details. I am offering these at rock-bottom prices! I can't find anyone to buy them! - David Summers -- "Never under-estimate the bandwidth of a station-wagon David Summers full of tapes, hurtling down the highway." dws@engr.uark.edu - Tanenbaum, "Computer Networks" ------------------------------ From: paulzen@salyko.cube.net (Oliver Paulzen) Crossposted-To: alt.os.linux Subject: Re: /dev/tty (Whatis it for?) Date: 20 Aug 1993 20:41:43 GMT C.W. Southern (cws9669@ultb.isc.rit.edu) schrieb: > I am having a program with serial logins on my machine. I am running SLS > 1.03. I have a good connection between the two, and when I try to connect > the to together it will get as far as the Password: prompt then stop. So I > ran strace on getty and found for some reason that the login program switches > ttys from /dev/ttys2 (COM3) to /dev/tty. When this file is removed the login > process continues as normal. But this file (/dev/tty) is required when loading > xterm up under X-windows. I have a few questions. What is the reason for this > file (/dev/tty)? The only way that I can find around it is to patch login > program, is there another? I have the same probleme on my machine, which is running with SLS-1.03 too. Quite the same probleme I had with an old version of SLS wich used the pl6 of 0.99. In the first two weeks running the system, connection was quite o.k. After this periode, I don't know what happend, there were the same problems as above. If anybody is having an answer for this problem, please send an eMail. TNX Oliver Paulzen (paulzen@cubenet.sub.org) ------------------------------ From: root@hip-hop.suvl.ca.us (Remco Treffkorn) Subject: Re: SCSI Performance (Yet Again) Date: Sat, 21 Aug 1993 16:47:58 GMT soup@SonOSam.execnet.com (John R. Campbell) writes: : I keep seeing bitching and moaning about "poor SCSI performance" during : reads with all kinds of hand waving and inconsistent numbers. Yes, the numbers are not the same as mine. I think theirs must be inconsistent ;-) : : Please- Remember something called a Buffer Cache? If you've got a 16MB : box, up to 6MB can be (and, if no programs take up the space, will be) : the Disk's Buffer cache. Isn't that the thing that makes writes lightning fast as compared to raw writes ?-) But without slowing down reads? : : Sure, it can take < 30 seconds to write a big 8MB file and 45-50 seconds : to read it, since the write() system call merely delivers it to the buffer : cache; the read() call may have to wait for the disk to come around before : it returns to the user. If you watch the little disk intensity light and : time it, the numbers should be about even. If a memory move from user space to kernel space and queing it in the buffer cache takes 30 seconds without the actual write taking place then I would consider that a major slowdown. I might agree with your reasoning regarding reads taking longer then writes *but* only if the program doing the write does not do a (f)sync. There is a compiletime option in iozone to do that. If the (f)sync is properly done reads should *not* take longer than writes. : : Linux's performance is EXTREMELY respectable against $CO. I'm not even : counting Bang for the Buck, either. I could not agree more! It is that way because people were not happy with the prformance of the early drivers and bitched about it. Thanks to all the developers who would listen and finding ways to tweak performance! : : Remember that the buffer cache can be VERY BIG on a Linux system and that : this will screw up any attempt to measure performance on a disk drive. Agreed, but normally the existence of a cache gives you numbers that are an order of magnitude too high. It makes things *look* better than they actually are. That is their purpose. I would find cache code that makes things look worse than the raw device pretty useless, unless it does other things that are worth the penalty. : : Of course, if we had raw drives, we'd be able to get better numbers... Maybe so, but that is not how it should be. : : -- : John R. Campbell soup@sonosam.execnet.com : Listen, the point here is simply that some people are wondering if the SCSI performance is all the way up there *or* if it could be improved. Natually they are comparing notes. The problem is only that nobody can agree on the correct way to measure performance. Iozone is certainly not perfect, but gives you a relative measure if compiled correctly. Again, nobody (I hope!-) is whining. I am subjectively happy with the disk performance, but after having seen some people with IDE presenting numbers that beat the sh*t out of my setup with an Adaptec SCSI controller I am certainly motivated to look into the matter a bit further. I think your understanding of how the buffer cache affects (should affect) performance is just wrong. But then, maybe things have changed since I left the university ;-) Cheers Remco remco@hip-hop.suvl.ca.us <<<== use this one for replies ------------------------------ Subject: Setting up bi-dir uucp From: kline@tig (Gary Kline) Date: 21 Aug 93 22:06:35 -0600 I've used all shades of Unix for 15 years--from v6 on to SVR4--and Linux is the newest. I could use some help from any Linux gurus out there. I've installed Linux of my 386SX/25 Acros laptop. With 126MB of diskspace, everything fits with a little extra space. So far, Linux looks like a superior WIN, esp'ly for hardcore hackers. But the docs are lean....or else I haven't found the right ones. So far, kermit work with whatever generic hayes clone is in the slot. But cu has troubles. Which probably means that I'm missing the proper config files. Can anyone tell me what files I need to make ``cu -s2400 -l cua1 -c 555 5555'' or ``cu '' work?? And do these files belong in /etc/default or /usr/lib/uucp? In /etc/inittab I'm using uugetty. Out of habit as much as anything. So far, the modem (ATS0=1&W) answers when I dial my data-line, but that's all I know. Can anywiz clue me in on what I should put in my /etc/default/uugetty file? Any assist much appreciated!! gary kline ------------------------------ From: bde@kralizec.zeta.org.au (Bruce Evans) Subject: Re: What's the FASTEST FS? Date: 17 Aug 1993 06:46:28 +1000 In <108355@hydra.gatech.EDU> gt8134b@prism.gatech.EDU (Howlin' Bob) writes: >In dfox@hip-hop.suvl.ca.us (David Fox) writes: >>Ed Carp (erc@apple.com) wrote: >>: What's the FASTEST FS? I know that ext is much slower than ext2, but is minix, >>: xiafs, or ext2 the fastest? >>Why, 386BSD's. (Berkeley FFS.) On my system (486/33, WD1007V ESDI controller, one 10 Mb/sec (35 sectors/ track) drive formatted with 1:1 interleave, another 15 Mb/sec (54 sectors/ track) drive formatted with 2:1 interleave), the standard Minix fs is much faster under Linux than a highly tuned ufs under 386BSD (tuning actually didn't help because of deficiencies in the driver and slowness of the controller). The only real advantage that ufs might have is less fragmentation over time. However, the standard 386BSD wd driver can't keep up for writes of contiguous sectors, so the default values of 4 msec for 'rotdelay' and 1 for 'maxconfig' have to be used to get decent performance for writes, so all files are laid out with discontiguous blocks. In effect they suffer from 2:1 fragmentation immediately. ufs enforces an implicit fsync on directory operations. This makes creation and deletion of files 10-20 times slower than under Linux and negates many of the advantages of having a disk cache. Standard 386BSD suffers from a tiny disk cache. It can easily be 50 times smaller than the Linux cache on Linux systems with plenty of memory. >Well, this isn't really a useful answer (not that there is one). He >wanted to know about the filesystems available under Linux. If you >don't look at *that* particular, then you're probably wrong, too. >I'm sure the 4.4BSD FFS is better that 386BSD's older implementation, >as is the log-stuctured file system in many cases. I've read that the log-structured file system turned out to be not so fast compared with ffs when one old bug and some implementation details in ffs were fixed. Both the bug and the implementation details have to do with contiguous blocks not being handled right and/or used enough. The reason that the symple Linux fs's work so well is that they only attempt a simple contiguous layout, and some Linux drivers can handle contiguous blocks at full speed. The reason that some people say that the simple Linux fs's are slow may be that some Linux drivers cannot handle contiguous blocks at full speed :-). >And do you have numbers comparing xiafs/ext2/FFS on the same hardware >with comparable aging/fragmentation? I would indeed like to see them. I have numbers for iozone on unfragmented file systems. Linux is 1.5 to 2 times faster because there is no 2:1 hit from the layout policy. This is not very interesting. I've been thinking about writing some simple real-world tests. Perhaps something involving unzipping untarring and deleting multiple copies of the gcc sources many times. -- Bruce Evans bde@kralizec.zeta.org.au ------------------------------ From: becker@super.org (Donald J. Becker) Subject: Re: Ethernet Cards informations needed Date: Sun, 22 Aug 1993 01:59:44 GMT [[ My newsfeed is falling further behind each day, so this may not be timely.]] In article <24oglf$7hf@news.univ-rennes1.fr> lebailli@univ-rennes1.fr (LEBAILLIF Olivier) writes: > What are the ethernet cards from 3COM which run well with linux. > > It seems that the 3C509 doesn't (what about the 3C501 ???) The 3c509 driver is still in the alpha stage. The 3c503 driver has been working for about nine months. It's not the fastest ethercard around, but it's from 3Com. The 3c501 driver once worked (for about three minutes), but was never released. The hardware is too brain-damaged to use on a modern network. Analog: you wouldn't expect Linux to run on an 8088...why expect networking to work with an 8088-era ethercard? -- Donald Becker becker@super.org Supercomputing Research Center 17100 Science Drive, Bowie MD 20715 301-805-7482 ------------------------------ From: albion@csd4.csd.uwm.edu (Craig T Manske) Subject: Re: Serial: My IRQ is 5 NOT 4! Change? Date: 22 Aug 1993 04:06:33 GMT Reply-To: albion@csd4.csd.uwm.edu From article <24pugk$abo@hdxu03.telecom.ptt.nl>, by arthur@ptt-iat.uucp (Arthur Donkers): > In article <1993Aug16.183236.11735@mnemosyne.cs.du.edu> gcortevi@nyx.cs.du.edu (Greg Corteville) writes: <