rcd@ico.isc.com (Dick Dunn) (11/02/90)
(Mutter...why don't people change Subject: lines when they change topics? I passed by once because this was still carrying the Tektronix/88k subject.) richard@aiai.ed.ac.uk (Richard Tobin) writes: my@frapper.UUCP (Michael Yip) writes: > >After all this price arguments, did anyone include the cost of the > >software (eg the UNIX OS and util) into the comparsion? That, BTW, is a very good question, since that factor makes some of the low-end RISC machines--which include OS--very competitive. > With luck, within a year or so, they may be two industrial-strength > free Unixes available - GNU, from the Free Software Foundation, and > 4.4-detox (ie BSD "detoxified" - with the AT&T code removed). Fine. Now show me how many people will buy a machine assuming they can get the OS free "with luck" and "in a year"! Yeah, you bet, I'm going out to buy a few $K of hardware hoping I'll get lucky in the next year or so to have free software for it. COME ON! You can wish for free UNIX all you want (I do!:-), but it hasn't happened, it's not close, and it's not happening very fast. People are paying for OSes because they want to use machines NOW, not next year. Yes, GNU wants a freed UNIX-like system, but is there any reason to think they'll actually have one within the next year? I don't think so. Last public statement I saw, they hadn't even settled on a base for it. (And no, Mach is not going to be free in the near future. A freed micro-kernel is not a free kernel; any usable Mach-based system is going to be heavily tied up in licenses for quite some time yet.) And yes, Berkeley is freeing code...slowly. Has anyone seen the BSD boys put a firm date on when 4.4 will be out, let alone when there would be a de-tox'ed version? If you've watched Berkeley in the past, you know that if they don't set a date, they're not ready to set a date and it's point- less to second-guess when they'll have something out. (Before you flame: Consider that I'm stating what I think is an accurate assessment, *NOT* what I would like to see happen! Hey, I'd like to see a free BSD for my 386 next month.) -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...but Meatball doesn't work that way!
richard@aiai.ed.ac.uk (Richard Tobin) (11/06/90)
In article <1990Nov1.234831.2066@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes: >Fine. Now show me how many people will buy a machine assuming they can get >the OS free "with luck" and "in a year"! Yeah, you bet, I'm going out to >buy a few $K of hardware hoping I'll get lucky in the next year or so to >have free software for it. Fine. Now show me where I suggested anyone should. I merely pointed it out because when it happens, which might or might not be within a year, it's likely to make quite a difference to the workstation market. In particular, I think it will extend that market downwards to cheap 386 machines. -- Richard -- Richard Tobin, JANET: R.Tobin@uk.ac.ed AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nsfnet-relay.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin
rcd@ico.isc.com (Dick Dunn) (11/06/90)
richard@aiai.ed.ac.uk (Richard Tobin) writes: rcd@ico.isc.com (Dick Dunn) writes: > >Fine. Now show me how many people will buy a machine assuming they can get > >the OS free "with luck" and "in a year"!...[snide remarks deleted]... ... > Fine. Now show me where I suggested anyone should. That's the first, and lesser, part of my point, simply that few people will buy based on speculated availability of software. The more important part is that there's little chance of a cheap/free UNIX within a year or so... > I merely pointed it out because when it happens, which might or might > not be within a year, it's likely to make quite a difference to the > workstation market. In particular, I think it will extend that market > downwards to cheap 386 machines. My quarrel is with the idea that there's any significant likelihood of it happening. Sure, if it *did* happen in a year, it would make waves. But there's no chance of it. And if it takes two years, the workstation market coming down from above (the RISC folks) will already be there. They're already close enough in price for packaged systems (workstation + OS) to start lopping off the top of the 386/486 PC + UNIX market. In other words, by the time it happened, it wouldn't be extending the market "downwards" but only "sideways". -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...but Meatball doesn't work that way!
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (11/06/90)
In article <3699@skye.ed.ac.uk> richard@aiai.UUCP (Richard Tobin) writes: | I merely pointed it out because when it happens, which might or might | not be within a year, it's likely to make quite a difference to the | workstation market. In particular, I think it will extend that market | downwards to cheap 386 machines. If by "cheap" you mean "low part of the price range," as in "cheap car," at the moment you just can't buy a decent unix machine which fits the "cheap 386" class. I have been trying for over a year to identify a configuration which will run unix cheaply, as part of my "cheap-ix" project. I've evaluated about 15 machines (with the help of friends), and concluded that to run a useful unix environment (as opposed to a toy), you want 8MB RAM and 200MB disk. Otherwise you give up compilers, or X, or news, or whatever, and you certainly give up performance. An SX, even at 16MHz, seems to be a reasonable chip to do the job, but you can't get acceptable characteristics without RAM and disk. I would therefore dispute that you will ever be able to use the "cheap 386," which will be forever a capon intended for MS-DOS. Oh, and if you get your o/s from FSF, remember the size of their grep, bison, flex, emacs, etc. Even on V.4, with everything in the world in my kernel, and drivers for every device ever invented, emacs is bigger than the kernel. I have no information on what the size of their system will be, but past software shows that in the past they have favored functionality and performance over size reduction. However, the prices are coming down to the point where even the midrange system is getting affordable. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) The Twin Peaks Halloween costume: stark naked in a body bag
rcd@ico.isc.com (Dick Dunn) (11/07/90)
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: > If by "cheap" you mean "low part of the price range," as in "cheap > car," at the moment you just can't buy a decent unix machine which fits > the "cheap 386" class... I think you can; in fact, I have. But that's because I have exactly one different assumption from Davidsen... > I've evaluated about 15 machines (with the help of friends), > and concluded that to run a useful unix environment (as opposed to a > toy), you want 8MB RAM and 200MB disk. Otherwise you give up compilers, > or X, or news, or whatever, and you certainly give up performance. The problem is X. You can have a good, *real* UNIX machine for well under $2000 for hardware as long as you don't want X. A 386SX has enough power, 4 Mb is plenty of memory, and an 80-100 Mb disk will do. Now add X, and some layers above to let you do something with it (like OpenLook or Motif, and a window manager). Oh, sure, it's neat; you've got bas-relief window borders (that you actually use at least once every few thousand keystrokes:-) and pleasant colors and cute little icons. And it increases the cost of the machine by more than 50%. It's a bit like going out to buy a car, costing $8000 for the basic vehicle plus $5000 for a trim-and-instrumentation package. Figure it out: You need a monitor capable of 1024x768, probably color. If your brain operates at anything close to normal frequency (some suits can apparently tolerate a 56 Hz monitor, but few other folks can), you need a pretty good display card. That's about $700 over the cost of a good char- acter display and card. Davidsen's right about memory--add 4 Mb @ $50-55. Add a mouse (we'll waste a precious serial port) for $75 for a decent one. Oh, and you'll need a bunch more disk. Better add at least 20 Mb; incre- mental cost is $3.50-4.00/Mb. You've added over $1000 to the hardware cost of a machine that was under $2000. It ain't the fault of the hardware folks...and it ain't the fault of UNIX. The hardware exists right now (in fact, has existed for some time) to make a cheap, good UNIX system. It's just that we allow software bloat to move so fast it gives us zero or negative net technical progress (*if* we buy into it). See the discussion in comp.misc, Subject: A tirade about in- efficient software & systems. Duly noted, Davidsen's (valid) statements about how big emacs is. But I see that the X server here on my machine is also larger than the (absurdly large) kernel. -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 ...but Meatball doesn't work that way!
mslater@cup.portal.com (Michael Z Slater) (11/07/90)
With respect to the cost of the OS for RISC workstations: If you buy a SPARC chip set from LSI Logic, the OS license is built in. You pay $25K for a design kit, which includes a complete hardware documentation package and a working system board -- and a tape with SunOS, C compiler, NFS, etc. You can then distribute binary copies without any additional fees -- as long as each one goes with an LSI chip set. The OS port was done by Opus Systems, so I suppose they'll be providing support. Michael Slater, Microprocessor Report mslater@cup.portal.com 707/823-4004
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (11/07/90)
In article <1990Nov6.222057.17797@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes: | The problem is X. You can have a good, *real* UNIX machine for well under | $2000 for hardware as long as you don't want X. A 386SX has enough power, | 4 Mb is plenty of memory, and an 80-100 Mb disk will do. Remember, the idea of project cheap-ix is to have a practical system which will cost about the same as a color X terminal (and could therefore be bought by may with X-terminal budget). You can certainly scale back the level of performance and capability to reduce the price, but there are a lot of cheap systems which will function well in "glass tty" mode now. I really believe that by March I will be able to publish a list of parts and vendors which will allow the average user to assemble a machine which runs color X at useful speed, includes the fun parts of UNIX, and will cost less than $3000 in at most quantity three. Hopefully Q1 if I can convince a few vendors to run promotional deals for "cheap-ix" configurations. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) The Twin Peaks Halloween costume: stark naked in a body bag
rpeglar@csinc.UUCP (Rob Peglar) (11/07/90)
In article <1990Nov6.222057.17797@ico.isc.com>, rcd@ico.isc.com (Dick Dunn) writes: (stuff deleted) > > Figure it out: You need a monitor capable of 1024x768, probably color. If > your brain operates at anything close to normal frequency (some suits can > apparently tolerate a 56 Hz monitor, but few other folks can), you need a > pretty good display card. That's about $700 over the cost of a good char- > acter display and card. Davidsen's right about memory--add 4 Mb @ $50-55. > Add a mouse (we'll waste a precious serial port) for $75 for a decent one. > Oh, and you'll need a bunch more disk. Better add at least 20 Mb; incre- > mental cost is $3.50-4.00/Mb. You've added over $1000 to the hardware cost > of a machine that was under $2000. 1kx768? It is a myth that 1kx768 resolution is necessary for X. A prime example is television; very low resolution, but the images are spectacular. Why? Analog color. Learning from this example, the key is not resolution, but color ability. The brain is capable, relatively speaking, of seeing many more colors (distinguishing) than very small dots. (One of the reasons "wide lines" present problems currently). Monitor refresh, on the other hand, is very important. 70 Hz at least for flicker-free (zero eye strain). One wonders how many eyes have lost retinal capability by looking into bad monitors. 640x480x256 is marginal for adequate imaging. 640x480x1k is better, 640x480x16k is virtually indistinguishable (oxymoron, i know) from television. Sound odd? Perhaps. Resolution is key if one is drawing huge, 2D (e.g. AEC-type) drawings, and you want to be able to zoom into the drawing at factors like 1x10^15 and such. For X, color ability and monitor capability is key, not resolution. Rob -- Rob Peglar Comtrol Corp. 2675 Patton Rd., St. Paul MN 55113 A Control Systems Company (800) 926-6876 ...uunet!csinc!rpeglar
rcpieter@svin02.info.win.tue.nl (Tiggr) (11/08/90)
rpeglar@csinc.UUCP (Rob Peglar) writes: >1kx768? It is a myth that 1kx768 resolution is necessary for X. A prime >example is television; very low resolution, but the images are spectacular. Indeed. And as a result nobody uses television when displaying two moderately sized (say, 90x40 characters) windows. >640x480x256 is marginal for adequate imaging. 640x480x1k is better, >640x480x16k is virtually indistinguishable (oxymoron, i know) from >television. Sound odd? Perhaps. 640x480 is barely suitable for displaying even a single window. There are people who actually do other things than image viewing/processing. Tiggr
mh2f+@andrew.cmu.edu (Mark Hahn) (11/08/90)
there is hope in the future for using commodity parts for cheapix, but it's not here now. pervious posts have identified a major weak point in the plan: monitor and graphics adapter. there are some bits of news in the press about a new IBM standard called XGA, which is, as I recall, 1024x768 in color. it seems reasonable to expect the same painful bank-selecting interface as VGA, but hopefully not the same slow initial implementation. if you buy the argument that volume makes for cheap prices, you'll want to wait for XGA to become a commodity. on the other hand, I'm beginning to think an x terminal and a headless PC is the way to go. does anyone have reviews of to offer? regards, Mark
peter@ficc.ferranti.com (Peter da Silva) (11/08/90)
In article <2840@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: > and concluded that to run a useful unix environment (as opposed to a > toy), you want 8MB RAM and 200MB disk. Otherwise you give up compilers, > or X, or news, or whatever, and you certainly give up performance. Sigh. 4 Meg and 72 Meg disk is planty for a useful UNIX development system. I've done real work with a 40 Meg disk, but it's not very happy. X is a frill, not a necessity. News is a frill, not a necessity. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
spot@WOOZLE.GRAPHICS.CS.CMU.EDU (Scott Draves) (11/08/90)
In article <239@csinc.UUCP>, rpeglar@csinc.UUCP (Rob Peglar) writes: |> |> 640x480x256 is marginal for adequate imaging. 640x480x1k is better, |> 640x480x16k is virtually indistinguishable (oxymoron, i know) from |> television. Sound odd? Perhaps. |> |> Resolution is key if one is drawing huge, 2D (e.g. AEC-type) drawings, |> and you want to be able to zoom into the drawing at factors like 1x10^15 |> and such. For X, color ability and monitor capability is key, not |> resolution. |> |> Rob |> who uses X for imaging? Most people use it to run a window system, and that means there is a lot of text on the screen. For text you need high resolution. 640x480 is puny. That's only 80-100 characters across!! My current screen is 191 across, and I wish it were more. |> very low resolution, but the images are spectacular. |> Why? Analog color. yea. right. Consume Scott Draves Be Silent spot@cs.cmu.edu Die
jmaynard@thesis1.hsch.utexas.edu (Jay Maynard) (11/08/90)
In article <T_Y6OYC@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: >In article <2840@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: >> and concluded that to run a useful unix environment (as opposed to a >> toy), you want 8MB RAM and 200MB disk. Otherwise you give up compilers, >> or X, or news, or whatever, and you certainly give up performance. >Sigh. 4 Meg and 72 Meg disk is planty for a useful UNIX development >system. I've done real work with a 40 Meg disk, but it's not very >happy. ...and 4 meg and 160 MB of disk is enough for development and 2 days' worth of a full news feed...on a 286, yet. 2.6 meg was enough smaller to induce significant thrashing. >X is a frill, not a necessity. News is a frill, not a necessity. What? I thought you were a newsaholic...as for X, I got along fine without it, though Microbug's multiple virtual console driver helped quite a bit. -- Jay Maynard, EMT-P, K5ZC, PP-ASEL | Never ascribe to malice that which can jmaynard@thesis1.hsch.utexas.edu | adequately be explained by stupidity. "With design like this, who needs bugs?" - Boyd Roberts
peter@ficc.ferranti.com (Peter da Silva) (11/08/90)
In article <1553@svin02.info.win.tue.nl> rcpieter@svin02.info.win.tue.nl (Tiggr) writes: > 640x480 is barely suitable for displaying even a single window. I'm sitting in front of a (click, click, click) 684 by *217* display that holds an 80 by 25 window, including several dozen pixels of borders and available backdrop. A multisync monitor would be nice, but in real life you don't really need all that display. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (11/09/90)
In article <239@csinc.UUCP> rpeglar@csinc.UUCP (Rob Peglar) writes: | 1kx768? It is a myth that 1kx768 resolution is necessary for X. A prime | example is television; very low resolution, but the images are spectacular. Yes, and anyone who has every tried to put even 80x25 on a TV knows you can't read it. | Why? Analog color. Learning from this example, the key is not resolution, | but color ability. The brain is capable, relatively speaking, of seeing | many more colors (distinguishing) than very small dots. Unfortunately drawings and text are composed of very small dots. | 640x480x256 is marginal for adequate imaging. 640x480x1k is better, | 640x480x16k is virtually indistinguishable (oxymoron, i know) from | television. Sound odd? Perhaps. And for text and line drawings you need a total of two colors, preferably with high contrast to one another. | Resolution is key if one is drawing huge, 2D (e.g. AEC-type) drawings, | and you want to be able to zoom into the drawing at factors like 1x10^15 | and such. For X, color ability and monitor capability is key, not | resolution. You need 1024x768 just to get two regular "glass tty" screens of 80x25 up. Many people like 132 columns (you can tell by their C code and postings). Then you need room for icons, so that takes some more. The low resolution solution is to first drop from 9x16 to 8x13, losing serifs in the process, then to 8x8, losing decenders, and looking like a 1978 dot matrix printer. Most people use X for mostly text, and anything less than 9x16 is going to lead to eyestrain (maybe a well chosen 9x14). And this on a big screen. After looking at what people do with their X color capability, I think that most people don't make use of it more than 10% of the time (having pastel borders is not making use of it in any productive sense, unless the border color means something). -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) The Twin Peaks Halloween costume: stark naked in a body bag
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (11/09/90)
In article <T_Y6OYC@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: | Sigh. 4 Meg and 72 Meg disk is planty for a useful UNIX development | system. I've done real work with a 40 Meg disk, but it's not very | happy. I've done real work on 64k CP/M systems, but that doesn't mean I want to go back there again. A good CP/M system with two drive and a full 64k was about $3000 in 1978, and my wife sold hundreds of them. Given inflation I think that's a reasonable hardware cost for a UNIX system *with* frills, thanks. | X is a frill, not a necessity. News is a frill, not a necessity. Depends on what you're developing. You can say the same thing about screen editors. You can edit nicely with ed, and we did on V6. And as soon as vi came along everybody dropped ed. X is a useful, portable way to do certain graphical things which are not adapted to text. For one screen text work I usually don't use X, or miss it. UNIX is a frill. You can always use MS-DOS and get your cost *real* low. That doesn't make it comfortable. Henry rewrote nroff in awk, but that doesn't mean I intend to scrap to original. It's a great tool for those who don't have the "real thing," and I believe that's how it was intended. I want to see a full featured UNIX at prices anyone can afford, and to say that you can live without the luxuries is true, but a bit frugal for my taste. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) The Twin Peaks Halloween costume: stark naked in a body bag
steven@uicadd.csl.uiuc.edu (Steven Parkes) (11/09/90)
In article <O8Z6:J3@xds13.ferranti.com>, peter@ficc.ferranti.com (Peter da Silva) writes: |> In article <1553@svin02.info.win.tue.nl> rcpieter@svin02.info.win.tue.nl (Tiggr) writes: |> > 640x480 is barely suitable for displaying even a single window. |> I'm sitting in front of a (click, click, click) 684 by *217* display that |> holds an 80 by 25 window, including several dozen pixels of borders and |> available backdrop. A multisync monitor would be nice, but in real life you |> don't really need all that display. Maybe people are thinking about the days when televisions were used as monitors ... although TV's have about 480 active lines, 1) they are interlaced, which limits vertical resolution and 2) the signal is band-limited, which limits horizontal resolution. A 640/480 progressive scan monitor is going to be a lot better than the old days of using a TV with an RF modulator. [I'm not saying 640/480 is great (I use 1024/864), but its a lot better than the 24x40 windows that old systems using TV's got.] steven parkes --------------------------------------- University of Illinois Coordinated Science Laboratory steven@pacific.csl.uiuc.edu -------------------------
tif@doorstop.austin.ibm.com (Paul Chamberlain) (11/09/90)
In article <239@csinc.UUCP> rpeglar@csinc.UUCP (Rob Peglar) writes: >1kx768? It is a myth that 1kx768 resolution is necessary for X. A prime >example is television; very low resolution, but the images are spectacular. Spectacular? 640x480x16k may be great to look at the stuff from alt.sex.pictures but running multiple ~80x~24 windows just isn't useful with dinky resolution. X's usefulness is supposed to come from multiple windows. For real development work, I'd rather have 2048x2048x2 (_LESS_ bits). I agree with 1kx768 minimum. If it'd make a difference economically, I'd settle for 16 colors (and sacrifice the quality of my gif's :-) ). And, BTW, I'm interested in cheap-ix but don't know that I can help. Paul Chamberlain | I do NOT represent IBM. tif@doorstop, sc30661 at ausvm6 512/838-7008 | ...!cs.utexas.edu!ibmchs!auschs!doorstop.austin.ibm.com!tif
rcd@ico.isc.com (Dick Dunn) (11/09/90)
mh2f+@andrew.cmu.edu (Mark Hahn) writes: > ...there are some bits of news in the press > about a new IBM standard called XGA, which is, > as I recall, 1024x768 in color... OK, but 1024x768 is old news. Even 1024x768 with 256 colors has been around for a while. XGA is supposedly for MicroChannel architectures, which makes it uninteresting for most of the 86ish world. One trade- rag report put it at "up to 10% faster" than IBM's 8514, which would be respectable speed, but nothing remarkable there either. I don't see what XGA is going to provide that's new or different. -- Dick Dunn rcd@ico.isc.com -or- ico!rcd Boulder, CO (303)449-2870 Cellular phones: more deadly than marijuana.
bzs@world.std.com (Barry Shein) (11/09/90)
You can buy a Sun3/50 these days for about $1200 max (under $1000 if
you work at it.) That's a 4MB 68020 with ethernet, SCSI, and two
serial ports, 1150x900 monochrome, keyboard, mouse. An OS license is
bundled (I believe it's transferable, if not I think single user is
something like $495 with docs.)
Now to that you can add a 327MB Wren-IV for around $1500, that's $2700
or about $3200 with the OS. If the OS is bundled then you can throw in
a tape device (e.g. a used 60MB QIC-24 can probably be had for $500, I
know new ones go for around $750.)
So how far off are we? Or does it have to be x86 for some reason I'm
missing?
That's a very solid Unix workstation environment. For an additional
$1200 you can probably find a color Sun3/60 which is expandable to
24MB (about $2500 for the base system), that might be a used price (is
that cheating if there were sources for a lot of them?)
An on-site service contract for my 3/60HM with 327MB disk, 60MB tape
is about $135/mo. I suspect you could find depot service for under
$100 from one of the third-party service orgs (you probably don't want
service people coming into your bedroom anyhow.) Actually,my contract
is written for a color monitor as that's what they would probably give
me if my hi-res (1280x1600) ever dies, which was ok by me (they didn't
have a price for an HM, and doubted they'd have stock on the day it
ever died..)
???
--
-Barry Shein
Software Tool & Die | {xylogics,uunet}!world!bzs | bzs@world.std.com
Purveyors to the Trade | Voice: 617-739-0202 | Login: 617-739-WRLDpositron@cosmic.berkeley.edu (Shigeki Misawa) (11/09/90)
In Message-ID: <8bC=xse00VpgQb9lNc@andrew.cmu.edu> Mark Hahn says : >there is hope in the future for using commodity parts >for cheapix, but it's not here now. pervious posts have >identified a major weak point in the plan: monitor and >graphics adapter. there are some bits of news in the press >about a new IBM standard called XGA, which is, >as I recall, 1024x768 in color. it seems reasonable to expect >the same painful bank-selecting interface as VGA, >but hopefully not the same slow initial implementation. >if you buy the argument that volume makes for cheap prices, >you'll want to wait for XGA to become a commodity. Which gets to an interesting question, how much does it really cost to manufacture a large color monitor and graphics card? an how much more is this over the price of your run of the mill color monitor. In one of the recent computer rags (I forget which one, PC Week, Computer World, Digital Review?), there was a quote from a spokesman from one of the Taiwanese SPARC clone makers (I think it was Tatung) saying that large color monitors are not significantly more expensive to make than your run of the mill monitor and that the reason why the price is high is that the workstation market is willing to pay the extra buck for it. ______________ Shigeki Misawa UCB Physics Department
sysmgr@KING.ENG.UMD.EDU (Doug Mohney) (11/09/90)
In article <BZS.90Nov8173010@world.std.com>, bzs@world.std.com (Barry Shein) writes: > >You can buy a Sun3/50 these days for about $1200 max (under $1000 if [cut comparison shopping guide] >or about $3200 with the OS. If the OS is bundled then you can throw in >a tape device (e.g. a used 60MB QIC-24 can probably be had for $500, I >know new ones go for around $750.) >An on-site service contract for my 3/60HM with 327MB disk, 60MB tape >is about $135/mo. I suspect you could find depot service for under >$100 from one of the third-party service orgs (you probably don't want >service people coming into your bedroom anyhow.) Hum. $135/mon. * 12 months = $1620/year? Or a "mere" $1200/month for carry-in. Effectively, you're paying the initial cost about $3200 + $1200 = $4400, plus $1200+ each following year, unless you are getting a warranty for the used thing (maybe 3 months?) Now, take the cheap-ix. Component parts = ~$3000. Should any single part break after the first year (say, do the used things have warranties or are they sold on the 'lots of luck' policy?), you pay some amount of money which is less than $1200 (say the motherboard blows out and takes something with it, call the price, oh around $1200) and replace it yourself. More than likely, it gives one an excuse to upgrade because equilivant parts will either be A) Much cheaper (cost of plunging PC boards and hard disk) and B) More capability for same price (shift upward from 20Mhz to 25Mhz to ? clock speeds on chips). Meanwhile, Mr. Sun, looking long in the tooth since the early 80s, has probably blown a gizmo or two since you've had it, and your service contract increase anywhere from 10-20% a year because it is A) Getting older, so your odds of system failure for a major part are greater and B) Those Sun parts are getting scarcer, because other people's workstations are starting to bust a gut. Bill's box is more sustainable over the long run, and will be easier & cheaper to keep running.
herrickd@iccgcc.decnet.ab.com (11/09/90)
In article <2859@crdos1.crd.ge.COM>, davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: > I want to see a full featured UNIX at prices anyone can afford, and to > say that you can live without the luxuries is true, but a bit frugal for > my taste. > -- > bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) > The Twin Peaks Halloween costume: stark naked in a body bag Bill, you said that you are trying to spec a system that will let us have a reasonable cost Unix system. I have already spent too much on a computer and am living with msdos. Could you post the pieces as you find them? Especially the software pieces? I want to convert it to unix with minimum money, and reasonable effort. I have to at least be able to move files from DOS systems (to preserve what I have and to continue to communicate with other systems). Being able to execute DOS programs in the unix environment would be nice, but not crucial. Can you reveal to us the parts of your goal that you have found? dan herrick dlh Performance Marketing POBox 1419 Mentor Ohio 44061 (216)974-9637 herrickd@astro.pc.ab.com
richard@aiai.ed.ac.uk (Richard Tobin) (11/10/90)
In article <239@csinc.UUCP> rpeglar@csinc.UUCP (Rob Peglar) writes: >1kx768? It is a myth that 1kx768 resolution is necessary for X. That depends what you want X for. You could run X on a 100x100 monchrome screen, but it wouldn't be much use. >For X, color ability and monitor capability is key, not resolution. No. For displaying images, that may be the key. *I* want X so that I can have multiple text windows, with two readable 80x66 windows side-by-side. And for that, 1024x768 is barely adequate. -- Richard -- Richard Tobin, JANET: R.Tobin@uk.ac.ed AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nsfnet-relay.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin
gsh7w@astsun.astro.Virginia.EDU (Greg Hennessy) (11/10/90)
Peter da Silva writes:
#X is a frill, not a necessity. News is a frill, not a necessity.
Not everyone shares in these two opinions.
--
-Greg Hennessy, University of Virginia
USPS Mail: Astronomy Department, Charlottesville, VA 22903-2475 USA
Internet: gsh7w@virginia.edu
UUCP: ...!uunet!virginia!gsh7wbzs@world.std.com (Barry Shein) (11/10/90)
Hasn't anyone seen a TELETEXT system like they use on Cable (e.g.
those local for-sale channels), or airports, or a computer that hooks
up to a TV monitor?
The resolution for text is pretty bad, 12x40 (chars) is about all you
can get without your eyes falling out of your head, and even those
characters are badly formed. And you suddenly notice a lot of jitter
even on a good TV that you just don't notice when there's a picture
there.
I don't understand why anyone is even taking the claim that normal TV
is adequate seriously, we've all had any number of experiences with
that, it isn't.
Animated images (normal TV programming) is just a whole other thing,
not comparable really.
--
-Barry Shein
Software Tool & Die | {xylogics,uunet}!world!bzs | bzs@world.std.com
Purveyors to the Trade | Voice: 617-739-0202 | Login: 617-739-WRLDpeter@ficc.ferranti.com (Peter da Silva) (11/10/90)
Just as an aside, I've been playing around with Superhires on my Amiga. It is actually possible to get ~1500 by ~500 pixels from a tube with about the resolution of a good TV set if you don't go through the ghastly NTSC conversion (Amiga 3000 with 1080 monitor (analog RGB) in SuperHires Interlace). -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
staff@cadlab.sublink.ORG (Alex Martelli) (11/10/90)
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: ... > And for text and line drawings you need a total of two colors, >preferably with high contrast to one another. A few intermediate shades would be better - four instead of two are plenty - to make use of anti-aliased rendering for lines, curves, and fonts. Although not very popular (who knows why!), anti-aliasing is a good way to get more "apparent resolution/definition" out of a given number of pixels, for line and text work. Apart from that, I agree with your main point - 640 x 480 is not adequate for text and line work; still, I believe I'd rather have, say, 1024 x 768 with 4 grey levels used for antialiasing, than, oh, 1152 x 928, black-white monochrome. -- Alex Martelli - CAD.LAB s.p.a., v. Stalingrado 45, Bologna, Italia Email: (work:) staff@cadlab.sublink.org, (home:) alex@am.sublink.org Phone: (work:) ++39 (51) 371099, (home:) ++39 (51) 250434; Fax: ++39 (51) 366964 (work only), Fidonet: 332/401.3 (home only).
gdtltr@brahms.udel.edu (Gary D Duzan) (11/11/90)
In article <1990Nov9.200125.18287@murdoch.acc.Virginia.EDU> gsh7w@astsun.astro.Virginia.EDU (Greg Hennessy) writes: =>Peter da Silva writes: =>#X is a frill, not a necessity. News is a frill, not a necessity. => =>Not everyone shares in these two opinions. => X is something you need for your terminal; News is something you need from your server. I admit a bias, however, since I am getting seriously into Distributed Systems & Operating Systems. Plan 9 from Bell Labs has the right idea, but lacks transparency. Gary Duzan Time Lord Third Regeneration -- gdtltr@brahms.udel.edu _o_ ---------------------- _o_ [|o o|] An isolated computer is a terribly lonely thing. [|o o|] |_O_| "Don't listen to me; I never do." -- Doctor Who |_O_|
dmocsny@minerva.che.uc.edu (Daniel Mocsny) (11/14/90)
In article <1990Nov9.200125.18287@murdoch.acc.Virginia.EDU> gsh7w@astsun.astro.Virginia.EDU (Greg Hennessy) writes: >Peter da Silva writes: >#X is a frill, not a necessity. News is a frill, not a necessity. > >Not everyone shares in these two opinions. Computers are a frill. -- Dan Mocsny Snail: Internet: dmocsny@minerva.che.uc.edu Dept. of Chemical Engng. M.L. 171 dmocsny@uceng.uc.edu University of Cincinnati 513/751-6824 (home) 513/556-2007 (lab) Cincinnati, Ohio 45221-0171
gsh7w@astsun.astro.Virginia.EDU (Greg Hennessy) (11/14/90)
Daniel Mocsny writes:
#Computers are a frill.
It is kind of hard to reduct CCD images without them.
--
-Greg Hennessy, University of Virginia
USPS Mail: Astronomy Department, Charlottesville, VA 22903-2475 USA
Internet: gsh7w@virginia.edu
UUCP: ...!uunet!virginia!gsh7wlupienj@hpwadac.hp.com (John Lupien) (11/15/90)
In article <T_Y6OYC@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: >In article <2840@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: >> and concluded that to run a useful unix environment (as opposed to a >> toy), you want 8MB RAM and 200MB disk. Otherwise you give up compilers, >> or X, or news, or whatever, and you certainly give up performance. >Sigh. 4 Meg and 72 Meg disk is planty for a useful UNIX development >system. I've done real work with a 40 Meg disk, but it's not very >happy. >X is a frill, not a necessity. News is a frill, not a necessity. It wasn't that long ago that I was using VENIX on an 8086 machine, with 20MB disk and (shudder) 128KB memory. This was intended to be a serious development project, too, and yes, we had compilers. X, news, and other frills were certainly out of the question, but the product (such as it was) got built, and worked pretty well. It didn't take much to get swapping involved, so one had to be careful, but basically Peter is right: *NIX does not imply megabytes of memory, nor does it require enormous disks, but if you don't have them, you must give up the frills. --- John R. Lupien lupienj@hpwarq.hp.com
herrickd@iccgcc.decnet.ab.com (11/15/90)
Greg Hennessy writes: > Daniel Mocsny writes: > #Computers are a frill. > > It is kind of hard to reduct CCD images without them. Thus demonstrating neatly that CCD images are a frill. dan herrick herrickd@astro.pc.ab.com
richard@aiai.ed.ac.uk (Richard Tobin) (11/16/90)
In article <11619@alice.att.com> andrew@alice.att.com (Andrew Hume) writes: >megascan ... monitors ... 300 DPI, 4096x~3300 ... $3500 This sounds most impressive, >300dpi one as an alternate screen for my gnot. Which reminds me - is Plan 9 likely to be available for off-the-shelf workstations in the near future? I know that current workstations don't really fit the Plan 9 model, but there are lots of us who'd like to try it who can't afford two large SGI machines connected by a high-speed bus. -- Richard -- Richard Tobin, JANET: R.Tobin@uk.ac.ed AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nsfnet-relay.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin
henry@zoo.toronto.edu (Henry Spencer) (11/18/90)
In article <3775@skye.ed.ac.uk> richard@aiai.UUCP (Richard Tobin) writes: >Which reminds me - is Plan 9 likely to be available for off-the-shelf >workstations in the near future? ... Uh, as I understand it, the words "available" and "Plan 9" do not go together. It is a research effort, not a commercial product. -- "I don't *want* to be normal!" | Henry Spencer at U of Toronto Zoology "Not to worry." | henry@zoo.toronto.edu utzoo!henry
ronald@robobar.co.uk (Ronald S H Khoo) (11/19/90)
henry@zoo.toronto.edu (Henry Spencer) writes: > Uh, as I understand it, the words "available" and "Plan 9" do not go > together. It is a research effort, not a commercial product. We-ell... A few of the Plan 9 developers who spoke at the EUUG conference this July did say that they had some intentions to make it available (with the rider: "if it's made available in any form at all, it will be in *source*, and we'll do our best to ensure that it isn't standardised" or words to that effect) but gave no details as to timescale or on what terms or to whom. Richard obviously remembered this statement, and was probably fishing for more rumour :-) -- ronald@robobar.co.uk +44 81 991 1142 (O) +44 71 229 7741 (H)
henry@zoo.toronto.edu (Henry Spencer) (11/20/90)
In article <1990Nov18.192626.19401@robobar.co.uk> ronald@robobar.co.uk (Ronald S H Khoo) writes: >> Uh, as I understand it, the words "available" and "Plan 9" do not go >> together. It is a research effort, not a commercial product. > >We-ell... A few of the Plan 9 developers who spoke at the EUUG conference >this July did say that they had some intentions to make it available... >but gave no details as to timescale or on what terms or to whom. A further caution here: based on the early history of Unix, if Plan 9 does become available, what you're getting will probably be a slightly-flakey research prototype, not an industrial-strength system that can be relied on "straight out of the box" for heavy production use. Shaking a research system down into a robust production system is a lot of work, and I suspect it's rather peripheral to the purposes of the research people. (For all the messes the USG people made in PWB, and Berkeley made in 4.1BSD, they did put an awful lot of work into making the system run reliably and efficiently, to the point where the Murray Hill research people "bought back" a lot of that work for V7 and V8 respectively.) -- "I don't *want* to be normal!" | Henry Spencer at U of Toronto Zoology "Not to worry." | henry@zoo.toronto.edu utzoo!henry