HELLER@cs.umass.edu.UUCP (02/17/87)
It seems that my recent message stirred up some dust. First of all let me clarify a few things. First of I don't object to having a MMU. It is just that I don't think a MMU is *neccessary* for *most* applications. In a development/hacker environment and with real milti-user systems it is needed. It is doubtfull that MMUs will ever be common on *little* home/office systems (i.e. ST's, Amiga's, Mac's etc.) - most of these system will be running well behaved software (i.e. canned business and home software), so these system can easily get by with an unprotected multi-tasking O/S (like OS9 or AMIGADOS). I really doubt that UNIX/MINIX will ever be the O/S of choice amongst home and office users: UNIX is just too "weird" (why is the directory command "ls"? why is the type command "cat"? what is an inode? - only UNIX hackers really understand UNIX). Secondly, I have found that I have had little or no trouble with runaway pointers totally trashing the system, at worst it just trashes the current "process". Re-booting just that process is suffientent. I guess I tend to write good code. Also, I have little need to run jobs in the background much. I guess I am just a synchronious user. I suspect that if I had UNIX on my system it would just cause an *increase* in the system overhead, because I would not use enough of it to justify most of what it would supply me with. I just don't have much need of forks and pipes (or even virtual memory - I will be going to a 4-meg system soon, which will provide me with more than enough memory for anything I'm likely to run on my system. UNIX would just eat up more memory for page tables and system code than I would actually be using most of the time!). I suspect that that is true of most users of little systems. I have used UNIX a little and have *not* liked it at all. The documentation tends to be poor and the error messages not very informative. Unless MINIX imporoves on this (along with some other changes) I probably won't bother with it. If I go for a more advanced O/S than CP/M-68K, I'll probably go to OS9/68K or CDOS-68K (neither of which need a MMU). I just as soon not spend money for a MMU when I don't really need one. Robert Heller ARPANet: Heller@UMass-CS.CSNET BITNET: Heller@UMass.BITNET BIX: Heller GEnie: RHeller FidoNet: 101/27 (Dave's Fido, Gardner, MA) or 101/147 (Orange Fido, Orange, MA) CompuServe 71450,3432 Local PV VAXen: COINS::HELLER UCC Cyber/DG: Heller@CS
jon@blit.UUCP.UUCP (02/18/87)
Robert Heller writes: >First of I don't object to having a MMU. It is just that I don't think a >MMU is *neccessary* for *most* applications. In a development/hacker >environment and with real milti-user systems it is needed. It is doubtfull >that MMUs will ever be common on *little* home/office systems (i.e. ST's, >Amiga's, Mac's etc.) The i286 (IBM-AT) and i386 (Compaq 386) already include on chip MMU's. So does the Mot 68030. Other manufacturers will soon follow. "Anything that can be put on silicon will eventually cost a dollar." - Noyce. Many home systems already include an MMU, and the rest will soon follow. Jonathan Ryshpan MS D3645 USENET: ...hplabs!nsc!blit!jon National Semiconductor ARPA: decwrl!nsc!blit!jon@ucbvax.arpa 2900 Semiconductor Dr. DOMAIN: jon@blit.sc.nsc.com P.O. Box 58090 Sunnyvale, CA 95052-8090 PHONE: 408-721-5481
hsu@ENEEVAX.UMD.EDU.UUCP (02/19/87)
In article <8702181533.AA27804@ucbvax.Berkeley.EDU> Robert Heller writes: >It seems that my recent message stirred up some dust. First of all let me >clarify a few things. > >First of I don't object to having a MMU. It is just that I don't think a >MMU is *neccessary* for *most* applications. In a development/hacker >environment and with real milti-user systems it is needed. Today, these points appear to be true. > It is doubtfull >that MMUs will ever be common on *little* home/office systems (i.e. ST's, >Amiga's, Mac's etc.) - most of these system will be running well behaved >software (i.e. canned business and home software), so these system can >easily get by with an unprotected multi-tasking O/S (like OS9 or AMIGADOS). This, however, is a truly remarkable if not wholly ridiculous claim. MMU's are coming, whether you like them or not. Both Motorola's and Intel's latest offerings have them built in, and in a sense, the iAPX86 family has always had them. Also remarkable is this concept of "well- behavedness"; having spent some years in a software store, I can assure you that few programs are "well-behaved" out of the box, at least for the first year. Often longer. An unprotected multitasking system is only as safe as its most dangerous process, and I for one don't intend to restrict myself to one-vendor canned packages. This is the sort of mentality that breeds dead-end products. "Simpler systems tend to run simpler software." >I really doubt that UNIX/MINIX will ever be the O/S of choice amongst home >and office users: UNIX is just too "weird" (why is the directory command >"ls"? why is the type command "cat"? what is an inode? - only UNIX hackers >really understand UNIX). Honestly, you should have attended the Winter USEnix conference. "ls" is no weirder than "dir" or "catalog". Hell, you should have seen the fuss I made when I discovered that nobody's BASIC implementation used "scr" anymore; they all had this thing called "new". "cat", like most other Unix commands, is either an abbreviation or an acronym. Yes, it's terse, but you at least have the ability to redefine everything to your heart's content with the shell's alias mechanism. And nobody needs to know about inodes anymore than you need to know about VTOCs or FATs or whatnot. >Secondly, I have found that I have had little or no trouble with runaway >pointers totally trashing the system, at worst it just trashes the current >"process". Re-booting just that process is suffientent. I guess I tend to >write good code. Also, I have little need to run jobs in the background >much. I guess I am just a synchronious user. Either that, or you tend to write uncomplicated code. That you can detect the trashing is quite remarkable in itself; how do you know it's not subtly wrecking something else very quietly? >I have used UNIX a little and have *not* liked it at all. The documentation >tends to be poor and the error messages not very informative. >... > Robert Heller While the Unix documentation is cryptic to most (if not all) beginning users, the strength of Unix lies in its flexible and relatively consistent command interpreter. Most of the mysteries go away once you get the hang of the piping output through filters. And the quality and breadth of the tool assortment is unrivaled. The tutorials probably could stand to be simplified. -dave -- David "bd" Hsu hsu@eneevax.umd.edu <or> seismo!mimsy!eneevax!hsu EE Computer Facility, Maryversity of Uniland, College Park, MD 20742