chris@mimsy.umd.edu (Chris Torek) (11/29/90)
(I have no idea what this has to do with computer architecture...) In article <00940487.15804140@KING.ENG.UMD.EDU> sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: [$799 ACM CD-ROM] >So what format are the papers in? Were they scanned in or are we >talking simple ASCII text? Compilation of the works cost someone a >certain amount of time and energy, plus the initial mastering of the >disk (some number greater than $2 bucks). Musician friends tell me that you can walk into a CD house with a digital master tape, plunk down $2000, and walk out with a digital master tape and 1000 CDs. This means that the cost is $2/CD for very low volumes, hence only lower for reasonable volumes. I have no idea whether CD-ROMs use the same mastering techniques as music CDs, but a good estimate for 100% markups at two levels of delivery would put CD-ROM prices at around $5 each, *provided* that the equivalent of `making the master tape' was free. Personally, I would imagine people would happily pay the same as the total price of the year's issues for a year-end CD-ROM of a technical journal. Shipping this should cost much less than shipping a year's worth of the journal! (Libraries would pay this, and then get rid of the paper version, so as to fit more in less space. Shelf space is expensive!) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 405 2750) Domain: chris@cs.umd.edu Path: uunet!mimsy!chris
sysmgr@KING.ENG.UMD.EDU (Doug Mohney) (11/29/90)
In article <28083@mimsy.umd.edu>, chris@mimsy.umd.edu (Chris Torek) writes: >Musician friends tell me that you can walk into a CD house with a digital >master tape, plunk down $2000, and walk out with a digital master tape >and 1000 CDs. This means that the cost is $2/CD for very low volumes, >hence only lower for reasonable volumes. I have no idea whether CD-ROMs >use the same mastering techniques as music CDs, but a good estimate for >100% markups at two levels of delivery would put CD-ROM prices at around >$5 each, *provided* that the equivalent of `making the master tape' was >free. Hum...I remember reading something somewhere (yes, memory fails again. Guess I'll have to get an AI prompter like Henry has), whereby making a CD-ROM in small production was $5-10K. Again, I was wondering about the labor costs involved to put everything together...of course, it wouldn't surprise me if they were just gouge-happy mongrels wanting to make a buck :-) >Personally, I would imagine people would happily pay the same as the >total price of the year's issues for a year-end CD-ROM of a technical >journal. Shipping this should cost much less than shipping a year's >worth of the journal! (Libraries would pay this, and then get rid of >the paper version, so as to fit more in less space. Shelf space is >expensive!) I wonder if people will be giving up their stacks of National Geographic for a CD-ROM? Would kill a tradition. %%%%% Signature v1.1 %%%%% Doug Mohney, Operations Manager, CAD Lab/ME, Univ. of Maryland College Park * Why do VMS system managers get more sleep and less ulcers than their * * UNIX(TM) counterparts, despite the sophistication of UNIX? *
brett@cayman.amd.com (Brett Stewart) (11/30/90)
In article <28083@mimsy.umd.edu> chris@mimsy.umd.edu (Chris Torek) writes: >(I have no idea what this has to do with computer architecture...) > >Musician friends tell me that you can walk into a CD house with a digital >master tape, plunk down $2000, and walk out with a digital master tape >and 1000 CDs. This means that the cost is $2/CD for very low volumes, >hence only lower for reasonable volumes. I have no idea whether CD-ROMs >use the same mastering techniques as music CDs, but a good estimate for >100% markups at two levels of delivery would put CD-ROM prices at around >$5 each, *provided* that the equivalent of `making the master tape' was >free. What this has to do with computer architecture is that the economics here pretty much predict that the things on our desks will have CD slots Real Soon Now, and this stuff will become indispensable to us. Lots of our customers are talking to us about how to do this now. And what I wanted to report to the net was similar economics for CD ROM mastering: last year during COMPCON I walked down to the CD-ROM show and browsed around. A guy from Denon (yeah, like in audio CD) told me that if you show up with your master tape and $1800, 48 hours later you could start getting your CD ROMS for under 2 bucks in Jewel Cases, you supply your own art. Other interesting news was that Full Motion Video compressed via DCT was getting 70 minutes per CD-ROM of normal US TV resolution video, and people expect this to quadruple within a few years. Best Regards; Brett Stewart Advanced Micro Devices, Inc. 1-512-462-5321 FAX 5900 E. Ben White Blvd MS561 1-512-462-4336 Telephone Austin, Texas 78741 USA brett@cayman.amd.com
sysmgr@KING.ENG.UMD.EDU (Doug Mohney) (11/30/90)
In article <1990Nov29.162726.11411@mozart.amd.com>, brett@cayman.amd.com (Brett Stewart) writes: >What this has to do with computer architecture is that the economics >here pretty much predict that the things on our desks will have CD >slots Real Soon Now, and this stuff will become indispensable to us. I confess, I was pretty tempted by the offer in DAK Electronics for the mondo 5 CD-ROM set which had a full library of reference materials (All yours if you had a PC-class machine with VGA...) and I usually wouldn't give very much in DAK a second thought. Sun wants to issue all their software on CDs by, what, '91? Digital has a similiar program, but they put most of their layered products on a 3 CD set. All you need is the proper License Paks and you can run the stuff. I think Digital will be multi-media (ie: reel, cartridge, TK50/70, CD-ROM) for quite a while, tho, due to their installed base of equipment. With Sun, you just throw it on the SCSI and you're good to go. >A guy from Denon (yeah, like in audio CD) >told me that if you show up with your master tape and $1800, 48 >hours later you could start getting your CD ROMS for under 2 bucks >in Jewel Cases, you supply your own art. VMS backup format? UN*X tar? And what gets written out? Don't Sun's CD's differ from VMS's format? Or do they both adhere to the "High Sierra" format? I never got that straight. After all, there is only one format for audio CDs. >Other interesting news was >that Full Motion Video compressed via DCT was getting 70 minutes per >CD-ROM of normal US TV resolution video, and people expect this to >quadruple within a few years. %%%%% Signature v1.1 %%%%% Doug Mohney, Operations Manager, CAD Lab/ME, Univ. of Maryland College Park * Why do VMS system managers get more sleep and less ulcers than their * * UNIX(TM) counterparts, despite the sophistication of UNIX? *
albaugh@dms.UUCP (Mike Albaugh) (11/30/90)
From article <0094070F.B370E000@KING.ENG.UMD.EDU>, by sysmgr@KING.ENG.UMD.EDU (Doug Mohney): > In article <1990Nov29.162726.11411@mozart.amd.com>, brett@cayman.amd.com (Brett Stewart) writes: > >>What this has to do with computer architecture is that the economics >>here pretty much predict that the things on our desks will have CD >>slots Real Soon Now, and this stuff will become indispensable to us. Which might, of course, revive the raging flame-war from sci.electronics about the increase in error rate with time on this supposedly "read only" media. I don't mind the occasional drop out in ZZ top, and probably wouldn't even see the "interpolation" that audio CDs do when they are a little more confident that the errors were "small", but I don't really want to deal with "just a few little errors" in, say, "as" or "ld". :-) BTW: I know about error correcting codes, etc, but the story from the audiophiles is that the little "correction light" (_If_ you have one) tends to light more and more often with age (of the disk), which perhaps explains why newer decks tend not to have them. "Don't lets alarm the users, now" :-). And a n-bit correction scheme with n-1 _known_ bad bits gets us a bit close to the edge... Mike | Mike Albaugh (albaugh@dms.UUCP || {...decwrl!pyramid!}weitek!dms!albaugh) | Atari Games Corp (Arcade Games, no relation to the makers of the ST) | 675 Sycamore Dr. Milpitas, CA 95035 voice: (408)434-1709 | The opinions expressed are my own (Boy, are they ever)
muir@cae780.UUCP (David Muir Sharnoff) (11/30/90)
I looked into distributing my company's software on CD ROM. I didn't save my findings, but I remember most of them... Young Minds Inc. sells software to make the master tapes. Using their software, you can build a CD ROM image on your hard disk. You can mount the image for testing purposes. Once tested, you can transfer it to tape and have it pressed. Young Minds' software costs about $7k. CD ROMs can hold up to 600MB. Pressing a CD requires that you have a master. There are two kinds of masters: metal & glass. A metal master is good for about 50,000 CD's. It costs between $800 and $2400 depending on how fast you want it, who you go to, etc... Each CD costs about $1.20 (again, milege may vary). A glass master costs about $250 and is only good for one CD. There are several formats for CD ROM out there. The format Sun (some standard, what was it???) uses seems to the winner. Digital uses a different format. Rumors (sure, you can trust 'em): Digital will be switching over. Apollo will be soon support it. NeXT either already supports it or will be shortly. You can buy device drivers for most OSs. Sorry to be so vague, -Dave -- David Muir Sharnoff. "RISC is about one year ahead" muir@csi.com (415) 358-3664 (415) 644-0441 Comdisco Systems Inc. 919 East Hillsdale Blvd, Foster City, CA 94404
bill@bilver.uucp (Bill Vermillion) (11/30/90)
In article <009406EF.82F93E60@KING.ENG.UMD.EDU> sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: >In article <28083@mimsy.umd.edu>, chris@mimsy.umd.edu (Chris Torek) writes: > >>Musician friends tell me that you can walk into a CD house with a digital >>master tape, plunk down $2000, and walk out with a digital master tape >>and 1000 CDs. This means that the cost is $2/CD for very low volumes, >>hence only lower for reasonable volumes. I have no idea whether CD-ROMs >>use the same mastering techniques as music CDs, but a good estimate for >>100% markups at two levels of delivery would put CD-ROM prices at around >>$5 each, *provided* that the equivalent of `making the master tape' was >>free. Well making a music master tape isn't free. When I worked in the studio the cost to do a major label production was from $100,000 to $250,000 dollars, but that's another story. >Hum...I remember reading something somewhere (yes, memory fails again. Guess >I'll have to get an AI prompter like Henry has), whereby making a CD-ROM in >small production was $5-10K. Again, I was wondering about the labor costs >involved to put everything together...of course, it wouldn't surprise me if >they were just gouge-happy mongrels wanting to make a buck :-) > I have a question, how about a wild guess as to how long it would take for you to assemble, on 1/2 tape or equivalent, 500 MEGABYTES of program to put on that CD, and to put it in the format needed. I suspect that even if the programs were FREE, the cost of assembly would raise the cost quite a bit. It's one thing to make a disk, but to make the tape equivalent of 500 before you master it is a bit intimidating. So labor costs of assembling such a beast are not a small item if it is a limited run CD-ROM. -- Bill Vermillion - UUCP: uunet!tarpit!bilver!bill : bill@bilver.UUCP
herrickd@iccgcc.decnet.ab.com (11/30/90)
In article <009406EF.82F93E60@KING.ENG.UMD.EDU>, sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: > In article <28083@mimsy.umd.edu>, chris@mimsy.umd.edu (Chris Torek) writes: > >>Musician friends tell me that you can walk into a CD house with a digital >>master tape, plunk down $2000, and walk out with a digital master tape >>and 1000 CDs. This means that the cost is $2/CD for very low volumes, >>hence only lower for reasonable volumes. I have no idea whether CD-ROMs >>use the same mastering techniques as music CDs, but a good estimate for >>100% markups at two levels of delivery would put CD-ROM prices at around >>$5 each, *provided* that the equivalent of `making the master tape' was >>free. > > Hum...I remember reading something somewhere (yes, memory fails again. Guess > I'll have to get an AI prompter like Henry has), whereby making a CD-ROM in > small production was $5-10K. I read the same thing. In something like The New Papyrus, the first Microsoft CDROM Symposium. 4 or 5 years ago. Another price learning curve. Slower than silicon. >Again, I was wondering about the labor costs > involved to put everything together...of course, it wouldn't surprise me if > they were just gouge-happy mongrels wanting to make a buck :-) > >>Personally, I would imagine people would happily pay the same as the >>total price of the year's issues for a year-end CD-ROM of a technical >>journal. Shipping this should cost much less than shipping a year's >>worth of the journal! (Libraries would pay this, and then get rid of >>the paper version, so as to fit more in less space. Shelf space is >>expensive!) > > I wonder if people will be giving up their stacks of National Geographic for a > CD-ROM? Would kill a tradition. I would pay $50 or $100 for a CDROM holding all of the IEEE periodicals for a year. Willing to wait a year for a clearance sale. They may not fit on one CDROM. The ACM collection might. It would make sense for the initial price to be significantly less than the subscription price of the paper or microform copies. Take the current cost figures, back out all printing and mailing costs, put in digitizing, mastering, CDROM production, and mailing costs. The cost should come out about two thirds of current costs. Lower if manuscripts were submitted in electronic form, but that makes a different production path for the two versions, meaning they would diverge. The National Geographic is another game entirely. They could begin a new publishing venture. Each monthly issue is one CDROM. Subscription price does not change. The photographs are GIF or something with higher fidelity, some are animated, the maps are digitized data that can be displayed at various levels of detail, the CDROM holds access software for MSDOS, MAC, and Amiga (or some other list - maybe there is one CDROM with software for fifty architectures distributed annually with occasional updates stuck on the monthly). The articles are there in their present form, but there is one or two levels of primary sources behind them (a keypress away). Interesting supplement (or an article) - 100 years of National Geographic maps of Europe, viewable as an animation through time. Lots of libraries have CDROM drives on computers, but not many homes. (Interesting use of relative quantifiers - "lots" means a few thousand, "not many" means a few thousand, probably several thousand more than "lots".) It might be that a way for Sony or Philips to increase the penetration of CDROM drives into the market is to subsidize National Geographic in the startup costs of such a venture. Why are CDROM drives moving down the price learning curve so much slower than CD players? > %%%%% Signature v1.1 %%%%% > Doug Mohney, Operations Manager, CAD Lab/ME, Univ. of Maryland College Park > * Why do VMS system managers get more sleep and less ulcers than their * > * UNIX(TM) counterparts, despite the sophistication of UNIX? * dan herrick dlh Performance Marketing POBox 1419 Mentor, Ohio 44061 (216)974-9637 herrickd@astro.pc.ab.com
msp33327@uxa.cso.uiuc.edu (Michael S. Pereckas) (11/30/90)
In <1181@dms.UUCP> albaugh@dms.UUCP (Mike Albaugh) writes: > Which might, of course, revive the raging flame-war from >sci.electronics about the increase in error rate with time on this >supposedly "read only" media. I don't mind the occasional drop out >in ZZ top, and probably wouldn't even see the "interpolation" that >audio CDs do when they are a little more confident that the errors >were "small", but I don't really want to deal with "just a few little >errors" in, say, "as" or "ld". :-) >BTW: I know about error correcting codes, etc, but the story from >the audiophiles is that the little "correction light" (_If_ you have >one) tends to light more and more often with age (of the disk), which >perhaps explains why newer decks tend not to have them. "Don't lets alarm >the users, now" :-). And a n-bit correction scheme with n-1 _known_ bad >bits gets us a bit close to the edge... Note that CD ROM puts an additional layer of error correcting code on top of the usual ECC on CDs. Does anyone know of any good research on bit decay on CDs? We keep hearing about people who have friends with neat CD players who say that they think that... but who knows how true it is. And of course, CDs haven't been around all that long yet. I heard a while ago about CDs made of glass instead of plastic, the idea being that while it is hoped that CDs will last 20 years, the glass ones should last 100 years, of so it was claimed. Does anyone know what happend to that? -- Michael Pereckas * InterNet: m-pereckas@uiuc.edu * just another student... (CI$: 72311,3246) Jargon Dept.: Decoupled Architecture---sounds like the aftermath of a tornado
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/01/90)
In article <009406EF.82F93E60@KING.ENG.UMD.EDU> sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: | I wonder if people will be giving up their stacks of National Geographic for a | CD-ROM? Would kill a tradition. Since I have gotten to the point where I have to throw away lots of (still possibly useful) magizines, buy a bigger house, or get a divorce, I certainly would be interested in buying "year end" CD-ROMs of most, although not all, or the magazines I get. I currently scan a *very* small number of articles in for online use, and cut a few out before the bulk of the magazine departs. Note that there is already one bimonthly magazine published in image format, the "GIF news." While it's hardly a best seller (it's a giveaway) it shows that some people are thinking about this. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/01/90)
In article <1181@dms.UUCP> albaugh@dms.UUCP (Mike Albaugh) writes: | BTW: I know about error correcting codes, etc, but the story from | the audiophiles is that the little "correction light" (_If_ you have | one) tends to light more and more often with age (of the disk), which | perhaps explains why newer decks tend not to have them. "Don't lets alarm | the users, now" :-). And a n-bit correction scheme with n-1 _known_ bad | bits gets us a bit close to the edge... Or, if you have the software, look at the soft error count on a copy of your latest backup tape to /dev/null. Just read through it and see what the error count is. Having an error correcting device and driver is helpful, but I would rather have better reliability at the front end. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/01/90)
In article <11212@charm.UUCP> muir@cae780.csi.com (David Muir Sharnoff) writes: | Rumors (sure, you can trust 'em): | Digital will be switching over. Haven't heard that one. I thoungth they'd change the day after Sun switches from Open Look to Motif. | Apollo will be soon support it. Heard that one. | NeXT either already supports it or will be shortly. And that one. | You can buy device drivers for most OSs. Yes, I heard that from a guy at UNIX Expo who sells the drives and drivers. We are thinking of getting one or two as a network resource, or even some kind of jukebox arrangement if it's debugged yet. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
sysmgr@KING.ENG.UMD.EDU (Doug Mohney) (12/01/90)
In article <2209.27561d1c@iccgcc.decnet.ab.com>, herrickd@iccgcc.decnet.ab.com writes: >Geographic in the startup costs of such a venture. Why are CDROM >drives moving down the price learning curve so much slower than CD players? >dan herrick >dlh Performance Marketing I'm still waiting for a release of a multi-disk CD-ROM player. Not like the mechanics are all that tough, since you can purchase up to 10-stack CD players at a time. Response time might not be blazing (oooh, gotta wait the 2-5 seconds to switch disks out), but it would be MUCH cheaper than racking 4-5 CD players together off a SCSI. %%%%% Signature v1.1 %%%%% Doug Mohney, Operations Manager, CAD Lab/ME, Univ. of Maryland College Park * Why do VMS system managers get more sleep and less ulcers than their * * UNIX(TM) counterparts, despite the sophistication of UNIX? *
esmith@goofy.apple.com (Eric Smith) (12/01/90)
In article <1181@dms.UUCP> albaugh@dms.UUCP (Mike Albaugh) writes: > Which might, of course, revive the raging flame-war from > sci.electronics about the increase in error rate with time on this > supposedly "read only" media. I don't mind the occasional drop out > in ZZ top, and probably wouldn't even see the "interpolation" that > audio CDs do when they are a little more confident that the errors > were "small", but I don't really want to deal with "just a few little > errors" in, say, "as" or "ld". :-) > BTW: I know about error correcting codes, etc, but the story from > the audiophiles is that the little "correction light" (_If_ you have > one) tends to light more and more often with age (of the disk), which > perhaps explains why newer decks tend not to have them. "Don't lets alarm > the users, now" :-). And a n-bit correction scheme with n-1 _known_ bad > bits gets us a bit close to the edge... I agree that one wrong bit is too many, but the error correction in CD-ROM is up to the challenge. Keep in mind that CD-Audio players have two levels of error correction (real honest-to-God correction, as in get the original data back). Error concealment (interpolation) is only used when the error correction fails. Many CD-Audio players don't implement the maximum possible error correction, because the manufacturers deem it acceptable to have error concealment used with moderate frequency. CD-ROM incorporates one additional layer of error correction above and beyond the error correction of CD-Audio. CD-ROM drives don't use concealment (except when playing audio CDs), but they still use the error correction of CD-Audio, and usually implement it fully. CD-ROM drives also typically have better tracking than CD-Audio players, which reduces the error rate considerable from its already low level. Most uncorrectable errors on CDs are the result of pressing defects. CD-ROMs are less succeptible to this as described above. I expect that more care is taken when CD-ROMs are pressed, although I have no evidence to support this. Every case of bit-rot on an Audio CD that I have seen has been due to improper handling of the CD. Most people don't realize that the label side of the CD is actually much more succeptible to damage than the "data" side. There is only an extremely thin layer between the aluminum and the label of the disk. *Don't* write on or scratch the top of a disk within the data area! (I write my name immediately around the center hole where no data is stored). The error correction for CD-Audio is interleaved. On a hard disk, a burst error of sufficient length can wipe out a block. On a CD, the bytes that make up a block are not consecutive on the disk, so a burst will really just cause a single correctable errors in each of many blocks. The error correction is capable of dealing with multiple errors per block. When you add the third layer of correction, the error rate for CD-ROM becomes lower than that of any other current data storage medium (that I've heard of). Would you rather trust your archives to floppies? Hard disk? Quarter or half inch tape? I'll take CD-ROM any day. -- Eric L. Smith Opinions expressed herein do not necessarily reflect those esmith@apple.com of my employer, friends, family, computer, or even me! :-)
dwells@fits.cx.nrao.edu (Don Wells) (12/01/90)
In article <1990Nov29.162726.11411@mozart.amd.com> brett@cayman.amd.com (Brett Stewart) writes: In article <28083@mimsy.umd.edu> chris@mimsy.umd.edu (Chris Torek) writes: >Musician friends tell me that you can walk into a CD house with a digital >master tape, plunk down $2000, and walk out with a digital master tape >and 1000 CDs. This means that the cost is $2/CD for very low volumes, >hence only lower for reasonable volumes. I have no idea whether CD-ROMs >use the same mastering techniques as music CDs, but a good estimate for >100% markups at two levels of delivery would put CD-ROM prices at around >$5 each, *provided* that the equivalent of `making the master tape' was >free. ... And what I wanted to report to the net was similar economics for CD ROM mastering: .... A guy from Denon (yeah, like in audio CD) told me that if you show up with your master tape and $1800, 48 hours later you could start getting your CD ROMS for under 2 bucks in Jewel Cases, you supply your own art. Other interesting news was Several months ago I visited the Nimbus CD production facility about 15_miles north of Charlottesville. I learned that CD-ROMs are produced on exactly the same production line as music CDs. The master molds are produced on the same mastering engines. The test procedures are identical. Nimbus has a subsidiary called Nimbus Information Systems which produces CD-ROMs; their offices are 100_meters from the CD factory. They write the digital master tapes that drive the mastering engines at the factory. The prices mentioned above are in the right range. A rough cost estimate is $1500 mastering charge plus $2 per CD-ROM. The price includes jewel cases and multicolor artwork printed onto the CD-ROMs; you supply the master for the artwork, and you can get booklets printed separately at a printing shop. The masters are retained indefinitely in a vault and can be used for additional production runs. I was told that, contrary to my a_priori assumption, most production runs -- even for music CDs -- are modest in size, a few hundred to a few thousand. For further information about producing your own CD-ROMs, contact: John L. Sands Support Services Manager, CD-ROM Division Nimbus Information Systems, Inc. P.O. Box 7305 Charlottesville, VA 22906 804-985-1100 804-985-4625(Fax) -- Donald C. Wells, Assoc. Scientist | dwells@nrao.edu Nat. Radio Astronomy Observatory | 6654::DWELLS Edgemont Road | +1-804-296-0277 38:02.2N Charlottesville, VA 22903-2475 USA | +1-804-296-0278(Fax) 78:31.1W
przemek@liszt.helios.nd.edu (Przemek Klosowski) (12/02/90)
Re: magazines distributed by electronic means There is a host of purely electronic 'newspapers' around: periodical and aperiodical. I even know of one that appears daily: several people at Warsaw University in Poland about a year ago started a daily news bulletin covering the current events there. The distribution is exclusively by e-mail. Initially their recipients were Polish researchers temporarily abroad, but lately more and more people with interest in Eastern European affairs started reading it, especially since the English translation became available. This initiative is a non-profit thing, with no fee charged. However, with proliferation of commercial networks, I imagine there will be more and more specialized news services like that, with the result that everyone will be able to tailor their news selection to their taste, in effect being able to customize the news they are getting. I tend to rip the ads (and articles about Intel 80x86 :^) from the magazines I get right after I get them, to make it easier to leaf thru them. Why not do that electronically? przemek -- przemek klosowski (przemek@ndcva.cc.nd.edu) Physics Dept University of Notre Dame IN 46556
tbray@watsol.waterloo.edu (Tim Bray) (12/04/90)
There's been discussion of CD-ROM here recently. Should just point out one important architectural issue: CDROMs have a seek time of on the order of 500 ms - that's right, half a second. This means that the selection of algorithms and data structures available for use on CD-ROM is very highly constrained. Also, it means that a CD-ROM is unlikely to be a satisfactory database access medium for more than one user at a time. It is unlikely that the access time issue will improve. The reason CD-ROM's are used at all is their extreme cheapness, which is achieved by leveraging off the high-volume manufacturing technology due to the music industry. Make any significant changes to the medium or the players to speed it up, and you lose that leverage, and the cost advantage compared to magnetic or more conventional optical media. On the other hand, CD's just CAN'T BE BEAT as an electronic *dissemination* medium. I wouldn't be surprised, in a few years, to see all the major database and software vendors shipping stuff out on CD's; the recipients, of course, copy the stuff onto a *real* disk to use it... Cheers, Tim Bray, Open Text Systems
minich@d.cs.okstate.edu (Robert Minich) (12/04/90)
tbray@watsol.waterloo.edu (Tim Bray): | There's been discussion of CD-ROM here recently. Should just point out one | important architectural issue: CDROMs have a seek time of on the order of | 500 ms - that's right, half a second. This means that the selection of | algorithms and data structures available for use on CD-ROM is very highly | constrained. Also, it means that a CD-ROM is unlikely to be a satisfactory | database access medium for more than one user at a time. | | It is unlikely that the access time issue will improve. The reason CD-ROM's | are used at all is their extreme cheapness, which is achieved by leveraging | off the high-volume manufacturing technology due to the music industry. Make | any significant changes to the medium or the players to speed it up, and | you lose that leverage, and the cost advantage compared to magnetic or | more conventional optical media. There was rumor printed recently (can't find the citation) that Apple is working on a CD-ROM player with access times around 35ms. Needless to say this would be avery useful product if priced reasonably. Apple has put a lot of resources into using CD-ROM for distribution. All registered developers get CDs at regular intervals that are jam packed with goodies. They also distribute training and info CDs to dealers. One service provides updates on all of Apple's MPW development tools all at once for a yearly fee. I believe the reason is not ideological but rather economical. Why distribute a couple disks and some paper when you could, for the same price, have hundreds of megabytes of useful info? I have no affiliation with Apple other than owning a Mac. Mistakes are clearly unmarked. -- |_ /| | Robert Minich | |\'o.O' | Oklahoma State University| "Get bent." |=(___)= | minich@d.cs.okstate.edu | -- Bart Simpson | U | - Ackphtth |
esmith@goofy.apple.com (Eric Smith) (12/04/90)
In article <1990Dec3.220850.18352@watdragon.waterloo.edu> tbray@watsol.waterloo.edu (Tim Bray) writes: > ... Should just point out one important architectural issue: CDROMs have > a seek time of on the order of 500 ms - that's right, half a second. The seek time depends mostly on the drive, not the media. There are already drives much faster than that. Heck, my CD-Audio player seeks faster than that. > It is unlikely that the access time issue will improve. This is definitely not true. There is no inherent reason why CD-ROM drives can't be made to seek as rapidly as any other current optical disks. -- Eric L. Smith Opinions expressed herein do not necessarily reflect those esmith@apple.com of my employer, friends, family, computer, or even me! :-)
gsarff@meph.UUCP (Gary Sarff) (12/04/90)
In article <009406EF.82F93E60@KING.ENG.UMD.EDU>, sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: >In article <28083@mimsy.umd.edu>, chris@mimsy.umd.edu (Chris Torek) writes: > >>Musician friends tell me that you can walk into a CD house with a digital >>master tape, plunk down $2000, and walk out with a digital master tape >>and 1000 CDs. This means that the cost is $2/CD for very low volumes, >>hence only lower for reasonable volumes. I have no idea whether CD-ROMs > .... [excised] >Hum...I remember reading something somewhere (yes, memory fails again. Guess >I'll have to get an AI prompter like Henry has), whereby making a CD-ROM in >small production was $5-10K. Again, I was wondering about the labor costs >involved to put everything together...of course, it wouldn't surprise me if > .... [excised] >I wonder if people will be giving up their stacks of National Geographic for a >CD-ROM? Would kill a tradition. We are in the education and training business, and have video disks and CD's mastered. I was just at comdex las vegas and talked to sony, phillips and 3-M and all quoted about $1000 or less to master and _less than_ $2.00 per copy after that, sony was I think something like $1.44 for quantities in only the hundreds. Interesting isn't it? There are several things I would like to see on CD-ROM as well. --------------------------------------------------------------------------- I _don't_ live for the Leap! ..uplherc!wicat!sarek!gsarff
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/04/90)
In article <1990Dec3.220850.18352@watdragon.waterloo.edu> tbray@watsol.waterloo.edu (Tim Bray) writes: | It is unlikely that the access time issue will improve. The reason CD-ROM's | are used at all is their extreme cheapness, which is achieved by leveraging | off the high-volume manufacturing technology due to the music industry. Since it is improving I can't agree, but even the 500ms time is largely irrelevant, since you pay that once in a well designed system, then the data is transferred to scratch fast media. A number of vendors sell NFS servers, based on a few cheap GB of disk, and a lot of optical in jukeboxes. Given some common sense in managing this, it works really well. I've seen the Epoch presentations, and the few people I've met who have it say the access is not a problem. I assume that other vendors have similar systems. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/04/90)
In article <00045@meph.UUCP> gsarff@meph.UUCP writes: | We are in the education and training business, and have video disks and CD's | mastered. I was just at comdex las vegas and talked to sony, phillips and | 3-M and all quoted about $1000 or less to master and _less than_ $2.00 per | copy after that, sony was I think something like $1.44 for quantities in only | the hundreds. Interesting isn't it? There are several things I would like | to see on CD-ROM as well. If FSF were a for-profit organization they would be taking advantage of this, and selling absolutely everything they ever wrote, in source and compiled for a number of popular machines, and including other non-FSF software covered by the copyleft or a similar agreement. And for, maybe $200/year, you could get a disk every quarter with the latest and greatest and hopefully the previous few versions of things like gcc and emacs which have been known to stop working on some machines when upgraded. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
brett@cayman.amd.com (Brett Stewart) (12/04/90)
In article <1990Dec3.220850.18352@watdragon.waterloo.edu> tbray@watsol.waterloo.edu (Tim Bray) writes: >There's been discussion of CD-ROM here recently. Should just point out one >important architectural issue: CDROMs have a seek time of on the order of >500 ms - that's right, half a second. This means that the selection of >algorithms and data structures available for use on CD-ROM is very highly >constrained. Also, it means that a CD-ROM is unlikely to be a satisfactory >database access medium for more than one user at a time. Of course, there are things, like full motion video in hypermedia, where this becomes less important. Accessing such an object with a 1-second penalty followed by transfer of vast quantities of digital; picture data works. Adding full-motion video to a user interface may be from a technical point of view a 'constraint' but many users will likely not see it that way. I believe there are CD jukebox servers now. The idea is anybody can query them, and they hold all the stuff. I think that the satisfactory or unsatisfactory characteristic of these as database servers depends on the type of data, its organization, whether the servers do anything smart, like keep all the indexes from every CD loaded on a hard disk, to cut down on multiple CD seeks, etc. One point made in this string has been that authoring of .6-2.4 Gigabytes of anything is difficult. Video can burn up that storage capacity really fast. (But authoring a hypermedia CD is still hard.) Best Regards; Brett Stewart Advanced Micro Devices, Inc. 1-512-462-5321 FAX 5900 E. Ben White Blvd MS561 1-512-462-4336 Telephone Austin, Texas 78741 USA brett@cayman.amd.com
lyndon@cs.athabascau.ca (Lyndon Nerenberg) (12/05/90)
bill@bilver.uucp (Bill Vermillion) writes: >In article <009406EF.82F93E60@KING.ENG.UMD.EDU> sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: >Well making a music master tape isn't free. When I worked in the studio >the cost to do a major label production was from $100,000 to $250,000 >dollars, but that's another story. Just to inject some reality here, the numbers quoted above include the studio time necessary to get all the audio down to mix onto that master. The actual cost of creating the master, if you were to walk in with all the material pre-recorded, would be *significantly* lower. Just to add a data point, it's possible to record, mix, master, and press 200 45 RPM singles for under $500 (Canadian dollars at that). The input format that the CD mastering machines want to see can be obtained by calling the CD mastering shops. Once you have that, you can create your own "master tapes" at the cost of renting time on a system with the ability to write media that the CD mastering machine can read (assuming you don't already have the necessary equipment in house already). -- Lyndon Nerenberg VE6BBM / Computing Services / Athabasca University {alberta,cbmvax,mips}!atha!lyndon || lyndon@cs.athabascau.ca Packet: ve6bbm@ve6mc [.ab.can.na] The only thing open about OSF is their mouth. --Chuck Musciano
norman@d.cs.okstate.edu (Norman Graham) (12/05/90)
From article <ESMITH.90Dec3183800@goofy.apple.com>, by esmith@goofy.apple.com (Eric Smith): > In article <1990Dec3.220850.18352@watdragon.waterloo.edu> tbray@watsol.waterloo.edu (Tim Bray) writes: > >> It is unlikely that the access time issue will improve. > > This is definitely not true. There is no inherent reason why CD-ROM drives > can't be made to seek as rapidly as any other current optical disks. Well, actually there is a reason: Other optical disks are laid out in a fashion similar to magnetic disks (tracks laid out in circles on the disk surface); CD-ROMs are laid out with a single track spiraling out from the center of the disk. Data density is constant along this spiraling track (that's the secret of their data capacity). Because of this scheme, the disk must spin with a constant tangential velocity at the point of the read head. What all this means is that when seeking the drive must move the read head to the proper location _and_ change the rotation speed. Thus CD-ROM drives have more overhead than other random access devices. Couple this with the fact that the proper seek position is not a precise read head position and you get slow seek times. [The older CD-ROM drives used to just guess where the read head should go and then start reading until they found the proper position. The better drives merely guessed better than the others.] Note: I'm not saying that these problems can't be overcome; I'm just pointing out that there's more to the problem than most people think. -- Norman Graham <norman@a.cs.okstate.edu> {cbosgd,rutgers}!okstate!norman The opinions expressed herein do not necessarily reflect the views of the state of Oklahoma, Oklahoma State University, OSU's Department of Computer Science, or of the writer himself.
aglew@crhc.uiuc.edu (Andy Glew) (12/05/90)
>There's been discussion of CD-ROM here recently. Should just point out one >important architectural issue: CDROMs have a seek time of on the order of >500 ms - that's right, half a second. This means that the selection of >algorithms and data structures available for use on CD-ROM is very highly >constrained. Also, it means that a CD-ROM is unlikely to be a satisfactory >database access medium for more than one user at a time. Why are CD-ROM seeks so slow? Is it related to the spiral tracking? Other optical disks (WORM and RW) have seeks on the order of 70ms. Proximal seeks are faster. Do CD-ROMs have proximal seeks? -- Andy Glew, a-glew@uiuc.edu [get ph nameserver from uxc.cso.uiuc.edu:net/qi]
carlton@aldebaran (Mike Carlton) (12/05/90)
In article <00045@meph.UUCP> gsarff@meph.UUCP writes:
+We are in the education and training business, and have video disks and CD's
+mastered. I was just at comdex las vegas and talked to sony, phillips and
+3-M and all quoted about $1000 or less to master and _less than_ $2.00 per
+copy after that, sony was I think something like $1.44 for quantities in only
+the hundreds. Interesting isn't it? There are several things I would like
+to see on CD-ROM as well.
You know that CD-ROMs are getting cheap when you get junk mail with one!
That's right -- Discovery Systems sent me a demo CD of their new service,
completely unasked for (I don't even own a CD-ROM drive).
They want something like $10/month for a monthly CD-ROM with the latest
Mac shareware/freeware etc. I haven't looked at the disk to see if it is
useful info, but it certainly is cheap.
--mike
Mike Carlton carlton@cs.berkeley.eduherrickd@iccgcc.decnet.ab.com (12/05/90)
In article <2989@crdos1.crd.ge.COM>, davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: > > If FSF were a for-profit organization they would be taking advantage > of this, [cheap CD-ROM publication costs] >and selling absolutely everything they ever wrote, in source > and compiled for a number of popular machines, and including other > non-FSF software covered by the copyleft or a similar agreement. And > for, maybe $200/year, you could get a disk every quarter with the > latest and greatest and hopefully the previous few versions of things > like gcc and emacs which have been known to stop working on some > machines when upgraded. > -- > bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) > VMS is a text-only adventure game. If you win you can use unix. I've only read the Dr. Dobbs version of the Manifesto and the copyleft distributed with an obsolete copy VMS EMACS, but it seems to me that they explicitly encourage people to provide the service of wider distribution of FSF distributions. Their condition is that nothing be left out. Sounds to me like you have identified an entrepreneurial opportunity. dan herrick herrickd@astro.pc.ab.com
bruce@zuhause.MN.ORG (Bruce Albrecht) (12/05/90)
>In article <009407D6.4101A640@KING.ENG.UMD.EDU> sysmgr@KING.ENG.UMD.EDU (Doug Mohney) writes: >I'm still waiting for a release of a multi-disk CD-ROM player. Not like the >mechanics are all that tough, since you can purchase up to 10-stack CD >players at a time. There's apparently a 6 disk changer available from Pioneer. It's advertised in MacWeek 12/4/90 by Relax Technology for $1100 in quantities of one, $999 for ten or more. I don't know if it's actually shipping, or just being advertised in anticipation of availability. -- bruce@zuhause.mn.org
ccplumb@spurge.uwaterloo.ca (Colin Plumb) (12/05/90)
At the World of Commodore show last Friday (Nov 30), Commodore was showing off their CDTV thingie... a CD player (built by Sony, looks like Sony) that also happens to be an Amiga inside. Anyway, a guy in Ottawa announced he'd turn your tape (or whatever) into 5,000 CD's for $7500 Canadian. That's about $1.25 U.S. apiece. (For my computer, I'd rather have a DAT tape, but I realize tapes are more expensive than CD's to produce in bulk.) -- -Colin
Bruce.Hoult@bbs.actrix.gen.nz (12/05/90)
In article <1990Dec4.224107.6589@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: > this scheme, the disk must spin with a constant tangential velocity at > the point of the read head. > > What all this means is that when seeking the drive must move the read > head to the proper location _and_ change the rotation speed. Not necessarily so. There's no reason that a drive with a fast seek movement but slow rotational speed adjustment (or constant speed) couldn't just read at odd speeds and buffer the data. Anyone remember those paper-tape readers on some early 8 bit micros where the user pulled the tape through by hand? -- Bruce.Hoult@bbs.actrix.gen.nz Twisted pair: +64 4 772 116 BIX: brucehoult Last Resort: PO Box 4145 Wellington, NZ
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/05/90)
In article <AGLEW.90Dec4170052@treflan.crhc.uiuc.edu> aglew@crhc.uiuc.edu (Andy Glew) writes: | Why are CD-ROM seeks so slow? Is it related to the spiral tracking? That seems to be it. Also, the drives were designed to deliver continuous data, and seek was not a time critical issue for the musical usage. Finally, I believe these drives use constant bpi, so the rotational speed in radians per sec changes as the head moves. | Other optical disks (WORM and RW) have seeks on the order of 70ms. | Proximal seeks are faster. Do CD-ROMs have proximal seeks? I'm going to let someone else handle that, I believe the answer is yes, but don't speak with certainty. The question is really one of what the buyers want... would a 4x improvement help anything? I think not, it would still be a very slow drive for random access. Perhaps it's better to keep the drive cheap and use caching to get the effective speed up. If you know the start of sections you can have a "branch target cache" for CD-ROM, too. That still won't help with truely random access to any byte on the media, or with very small records. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
meissner@osf.org (Michael Meissner) (12/05/90)
In article <2279.275bf1a4@iccgcc.decnet.ab.com> herrickd@iccgcc.decnet.ab.com writes: | I've only read the Dr. Dobbs version of the Manifesto and the | copyleft distributed with an obsolete copy VMS EMACS, but it | seems to me that they explicitly encourage people to provide | the service of wider distribution of FSF distributions. Their | condition is that nothing be left out. | | Sounds to me like you have identified an entrepreneurial opportunity. Doesn't Young Minds already ship GNU-ware on CD's. I seem to remember them at the Baltimore USENIX. -- Michael Meissner email: meissner@osf.org phone: 617-621-8861 Open Software Foundation, 11 Cambridge Center, Cambridge, MA, 02142 Considering the flames and intolerance, shouldn't USENET be spelled ABUSENET?
dmocsny@minerva.che.uc.edu (Daniel Mocsny) (12/06/90)
In article <2279.275bf1a4@iccgcc.decnet.ab.com> herrickd@iccgcc.decnet.ab.com writes: >In article <2989@crdos1.crd.ge.COM>, davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) writes: [ FSF could make $$$ selling CD-ROMs with all their stuff ] >Sounds to me like you have identified an entrepreneurial opportunity. Wasn't there an outfit called "Young Minds" or some such that did/is doing exactly this? I saw their ad in one of the UNIX trade rags. However, they were asking ca. $500 for a CD of PD UNIX software. I think that's a little steep, unless they were somehow adding value (?) How about a CD-ROM collection of the SIMTEL20 archives? That wouldn't even be very hard to do, since SIMTEL puts out listing files that let you automate the task of mirroring their archives (as wuarchive.wustl.edu does). I bet a LOT of people without ftp access to the Internet would like to get a copy of the SIMTEL20 archives...especially if the price was only a small multiple of the cost of a CD-ROM. -- 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
herrickd@iccgcc.decnet.ab.com (12/06/90)
c.uiuc.edu> Followup-To: .uiuc.edu> Lines: 22 In article <AGLEW.90Dec4170052@treflan.crhc.uiuc.edu>, aglew@crhc.uiuc.edu (Andy Glew) writes: >>important architectural issue: CDROMs have a seek time of on the order of >>500 ms - that's right, half a second. This means that the selection of > > Why are CD-ROM seeks so slow? Is it related to the spiral tracking? The audio Compact Disk spec includes constant linear speed along the spiral track that runs the entire length of the disk. The average seek includes moving a significant distance along the radius of the disk, requiring a noticable change in the rotational speed of the disk. That average change in rotational speed is the critical path factor in the half second seek time. Seeks that skip only a few revolutions of the spiral are typically done by rotating a mirror and happen quite fast (they don't require much change in the rotation speed of the disk, either). A half second between selections in an audio program is not at all unpleasant, the slow seek time was a good design decision, darn it. dan herrick herrickd@astro.pc.ab.com
norman@d.cs.okstate.edu (Norman Graham) (12/06/90)
From article <1990Dec5.105743.25693@actrix.gen.nz>, by Bruce.Hoult@bbs.actrix.gen.nz: > In article <1990Dec4.224107.6589@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: > >> this scheme, the disk must spin with a constant tangential velocity at >> the point of the read head. >> >> What all this means is that when seeking the drive must move the read >> head to the proper location _and_ change the rotation speed. > > Not necessarily so. There's no reason that a drive with a fast seek movement > but slow rotational speed adjustment (or constant speed) couldn't just read > at odd speeds and buffer the data. I've seen a couple of postings like this, so I'll try to clarify things a bit. CD-ROM drives vary the rotation speed (AV) so they can _read_ the disk. This has nothing to do with the drive trying to maintain a constant data rate. Let me explain: CD-ROMs don't actually store bits--they store transitions between bit runs. So the bit stream 00110000101110 has six transitions. Transitions are recorded so that the _distance_ between transitions indicates how many bits are in the bit run. Well, how do you measure the distance between transitions? With constant tangential velocity (or constant linear velocity (CLV)), the distance between transition A and transition B is a simple function of the time between transition A and transition B and the tangential velocity (which is a constant--remember?). Does the equation distance = rate * time sound familiar? But suppose we reverse the situation and use a constant rotation speed (i.e. constant angular velocity (CAV)). Then the distance between transition A and transition B becomes a more complex problem. Remember that because of CAV, the tangential velocity is increasing from transition A to transition B (points on the inside of the disk are spinning slower than points on the outside of the disk); thus we now have acceleration to worry about. [I've not reasoned through this (heck, it's just a USENET posting) but I believe the acceleration is non-linear.] So, what _do_ we have to work with to determine the distance between two transitions? Well, we still know the time between the transitions. And I suppose someone could come up with a way to measure the tangential velocity at transition A and transition B. If we assume a linear acceleration then things fall out at this point. If we assume the worst case [which is probable]--nonlinear acceleration that can't be predicted between disks or maybe not even on the same disk if the spiral doesn't conform to some mathematical function--then we'll have a much harder time trying to calculate the distance between transitions. All-in-all, I believe the best bet for decreasing seek times for CD-ROM drives is to keep the constant tangential velocity and to concentrate on speeding up the change of angular velocity. It's doubtful that CD-ROM drives will match the seek times of magnetic disks any time soon. But a 20ms seek time is _not_ necessary for good performance from CD-ROMs. I believe we can get reasonable performance from 100ms seeks--IF some intelligence is used when laying out the CD-ROM. Remember, CD-ROM is a _read-only_ medium. This means you should put a lot of time up front to choose a data layout that minimizes seeks. This means putting information that is used together close together; or perhaps putting the same information in several places on the disk; or perhaps interleaving certain files; etc. NOTE: Again, I believe we can overcome the problems hindering the reduction of seek times on CD-ROM. I'm just pointing out that the problem is not as easy as some think. -- Norman Graham <norman@a.cs.okstate.edu> {cbosgd,rutgers}!okstate!norman The opinions expressed herein do not necessarily reflect the views of the state of Oklahoma, Oklahoma State University, OSU's Department of Computer Science, or of the writer himself.
sysmgr@KING.ENG.UMD.EDU (Doug Mohney) (12/07/90)
In article <bruce.5240@zuhause.MN.ORG>, bruce@zuhause.MN.ORG (Bruce Albrecht) writes: > >There's apparently a 6 disk changer available from Pioneer. It's advertised in MacWeek >12/4/90 by Relax Technology for $1100 in quantities of one, $999 for ten or more. I >don't know if it's actually shipping, or just being advertised in anticipation of >availability. Yah, I stand somewhat corrected (should start reading "Byte" again, I guess); someone sent me a mail message on the sucker. They have one. Big drawback is access time when swapping CDs. Between this and cheap CD production (two sources previously listed on here, Nimbus I talked with myself, will print 100 CDs for around $1600), we're coming to a new age of information saturation :-) %%%%% Signature v2.0 %%%%% Doug Mohney, Operations Manager, CAD Lab/ME, Univ. of Maryland College Park * If Apple's pricing strategy had been as exciting as their commercials, * * Windows 3.0 would have never been written *
przemek@liszt.helios.nd.edu (Przemek Klosowski) (12/07/90)
In article <1990Dec6.154348.5206@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: > >But suppose we reverse the situation and use a constant rotation speed >(i.e. constant angular velocity (CAV)). Then the distance between >transition A and transition B becomes a more complex problem. Remember >that because of CAV, the tangential velocity is increasing from >transition A to transition B (points on the inside of the disk are >spinning slower than points on the outside of the disk); thus we now >have acceleration to worry about. [I've not reasoned through this >(heck, it's just a USENET posting) but I believe the acceleration is >non-linear.] So, what _do_ we have to work with to determine the >distance between two transitions? Well, if the disk rotates at constant angular speed omega, the linear speed of a point at radius r is v = omega * r. The spiral is so tightly woven that for all practical purposes it can be treated as concentric (ie. the tangent to its direction is perpendicular to the radius). Even if it weren't, the Archimedes spiral would have some angle close to 90 deg (say 89.99 deg), and the linear velocity would be linear function of radius, too. The real problems may pop up when you consider excentricity of the tracks, but then they plague constant linear velocity scheme also. -- przemek klosowski (przemek@ndcva.cc.nd.edu) Physics Dept University of Notre Dame IN 46556
phil@brahms.amd.com (Phil Ngai) (12/07/90)
In article <1990Dec6.154348.5206@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: |But suppose we reverse the situation and use a constant rotation speed |(i.e. constant angular velocity (CAV)). Then the distance between |transition A and transition B becomes a more complex problem. Remember Well, you ought to know that in the magnetic storage field, there are lots of products being shipped with more bits stored on the outer tracks than on the inner tracks, and they have very good seek times. I just saw a Seagate 89 megabyte IDE disk drive with Zone Bit Recording (ZBR) technology. It ranges from 44 sectors/track to 30 sectors/track. I assure you they are NOT varying the angular velocity! This kind of thing is both quite doable and is being done, in volume, at low cost. It is certainly more complex but that does not mean it can not be put into production in commodity products, it just means you need clever engineers to develop it. -- There is no right more fundamental than self-defense.
aglew@crhc.uiuc.edu (Andy Glew) (12/07/90)
>> Why are CD-ROM seeks so slow? Is it related to the spiral tracking? > >The audio Compact Disk spec includes constant linear speed along the >spiral track that runs the entire length of the disk. The average >seek includes moving a significant distance along the radius of the >disk, requiring a noticable change in the rotational speed of the >disk. That average change in rotational speed is the critical path >factor in the half second seek time. Seems to me that the same techniques that are used to get constant linear density (as opposed to constant angualr density) on regular disks should apply. -- Andy Glew, a-glew@uiuc.edu [get ph nameserver from uxc.cso.uiuc.edu:net/qi]
aglew@crhc.uiuc.edu (Andy Glew) (12/07/90)
It's beginning to sound to me like the computer industry should piggy back on the CD-ROM media format, but specialized drives for the CD-ROMs should be designed. Ie. leverage media, not mechanism. -- Andy Glew, a-glew@uiuc.edu [get ph nameserver from uxc.cso.uiuc.edu:net/qi]
pavlov@canisius.UUCP (Greg Pavlov) (12/07/90)
In article <AGLEW.90Dec4170052@treflan.crhc.uiuc.edu>, aglew@crhc.uiuc.edu (Andy Glew) writes: > Why are CD-ROM seeks so slow? Is it related to the spiral tracking? > Other optical disks (WORM and RW) have seeks on the order of 70ms. > Proximal seeks are faster. Do CD-ROMs have proximal seeks? > Unless the technology has changed, part of the problem was (is) that the head mass is considerably greater. greg pavlov, fstrf, amherst, ny pavlov@stewart.fstrf.org
Bruce.Hoult@bbs.actrix.gen.nz (12/08/90)
In article <1990Dec6.154348.5206@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: >I've seen a couple of postings like this, so I'll try to clarify things >a bit. > >CD-ROM drives vary the rotation speed (AV) so they can _read_ the disk. >This has nothing to do with the drive trying to maintain a constant >data rate. > >Let me explain: CD-ROMs don't actually store bits--they store transitions >between bit runs. So the bit stream 00110000101110 has six transitions. >Transitions are recorded so that the _distance_ between transitions >indicates how many bits are in the bit run. > >Well, how do you measure the distance between transitions? With constant >tangential velocity (or constant linear velocity (CLV)), the distance >between transition A and transition B is a simple function of the time >between transition A and transition B and the tangential velocity (which >is a constant--remember?). Does the equation distance = rate * time >sound familiar? Indeed it does. The problem is that even in normal audio players, starting from the first "track" and following the spiral to the end of the data, the linear velocity doesn't actually stay exactly constant. This would require that the drive motor be infinitely controllable and infinitely accurate. What actually happens is that the motor speed is fairly crudely controlled and the data is read into a FIFO buffer, and clocked out at a precise rate. That is why the wow & flutter for a CD is so much less than for an LP -- there is no way that the drive motors in the cheapest CD player are just *so* much more accurate than the motor in a multi thousand dollar Linn turntable. Have *you* heard of the term "phase locked loop"? :-) -- Bruce.Hoult@bbs.actrix.gen.nz Twisted pair: +64 4 772 116 BIX: brucehoult Last Resort: PO Box 4145 Wellington, NZ
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (12/08/90)
In article <1990Dec7.020857.18469@amd.com> phil@brahms.amd.com (Phil Ngai) writes: | Well, you ought to know that in the magnetic storage field, there | are lots of products being shipped with more bits stored on the | outer tracks than on the inner tracks, and they have very good | seek times. I just saw a Seagate 89 megabyte IDE disk drive with | Zone Bit Recording (ZBR) technology. It ranges from 44 sectors/track | to 30 sectors/track. I assure you they are NOT varying the angular | velocity! Your facts are correct, but they don't apply here. There is a, as in one, as in single, track on CD-ROM, and the linear read rate is constant, because that's the standard. In ZBR, if you keep the RPM the same, and put more data on the outer tracks, your data rate changes, at least at the head. When you look for track N on ZBR it is always the same distance from the hub, because it's circular. In CD-ROM you seek to the right distance, more or less, then wait until *the* track comes squarely under the head, then slide along the track to the right data. And since the time between marks is used for encoding, and it's written with constant speed and varying RPM, it must be read that way. I believe you're impying that there's a better way, and you're right, but the fact that these things are already out, have a standard, and are produced in mass quantities to make them cheap is an argument for staying with them, or one of the other established optical standards which are closer in performance (and price) to magnetic disk. Victor used ZBR on it's floppy disks, which was inovative, but Shugart got rich with tired but cheap tech. Design can't ignore cost. A new design *now* would have to be very good to get people to pay more for it, or to make the startup (low volume) cost competitive. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) VMS is a text-only adventure game. If you win you can use unix.
bill@bilver.uucp (Bill Vermillion) (12/09/90)
In article <AGLEW.90Dec6211736@treflan.crhc.uiuc.edu> aglew@crhc.uiuc.edu (Andy Glew) writes: >>> Why are CD-ROM seeks so slow? Is it related to the spiral tracking? >> >>The audio Compact Disk spec includes constant linear speed along the >>spiral track that runs the entire length of the disk. The average >>seek includes moving a significant distance along the radius of the >>disk, requiring a noticable change in the rotational speed of the >>disk. That average change in rotational speed is the critical path >>factor in the half second seek time. > >Seems to me that the same techniques that are used to get constant >linear density (as opposed to constant angualr density) on regular >disks should apply. Key word in your description is density, while in the cd-rom it is velocity. In changing density you don't have to change the angular velocity of the disk, you change the rate of data flow. You keep the disc speed the same. In the cd-roms you keep the data flow the same, and increase the density by changing the rotational rate. To change the cd-roms to another format would break the CD specs and you would have to have specialized data drives, and not be able to take advantage of the much larger manufacturing quantity cost economies. -- Bill Vermillion - UUCP: uunet!tarpit!bilver!bill : bill@bilver.UUCP
bill@bilver.uucp (Bill Vermillion) (12/09/90)
In article <1990Dec7.020857.18469@amd.com> phil@brahms.amd.com (Phil Ngai) writes: >In article <1990Dec6.154348.5206@d.cs.okstate.edu> norman@d.cs.okstate.edu (Norman Graham) writes: >|But suppose we reverse the situation and use a constant rotation speed >|(i.e. constant angular velocity (CAV)). Then the distance between >|transition A and transition B becomes a more complex problem. Remember >Well, you ought to know that in the magnetic storage field, there >are lots of products being shipped with more bits stored on the >outer tracks than on the inner tracks, and they have very good >seek times. I just saw a Seagate 89 megabyte IDE disk drive with >Zone Bit Recording (ZBR) technology. It ranges from 44 sectors/track >to 30 sectors/track. I assure you they are NOT varying the angular >velocity! >This kind of thing is both quite doable and is being done, in volume, >at low cost. It is certainly more complex but that does not mean >it can not be put into production in commodity products, it just >means you need clever engineers to develop it. Except the big difference in the magnetic media is there are many tracks, with a high tpi density. In a CD-ROM there is essentialy ONE track, which is a spiral. In magnetic media you can say at track xx we will go to an xxxx bit rate and when we reach track yy we can go to a yyyy bit rate. The imbedded or dedicated servo controls will ALWAYS position you to the correct track, that's what they are there for. Ignoring for the moment that a CD-ROM uses one track, imagine counting the tracks out on a CD. In a magnetic drive the position is dependant on the drive itself, and to take a plater from one drive to another would yield strange results. In CDs the tracks are independant of the drives. And if you examine the specifications for CD mastering you will find that there is a range for spacing the "tracks". Therefore you can not position a CD-ROM the same way . You could seek to an approximate postion, and then read the data stream to see where you are, and re-seek according to that. In magnetic media you can seek to the exact track each time (of course you do read the id to make sure!). Basically it boils down to two facts. Magnetic drives were designed for highly dense data storage and are made with considerable precision while CD media were designed to be highly transportable over less than precision products. -- Bill Vermillion - UUCP: uunet!tarpit!bilver!bill : bill@bilver.UUCP
phil@brahms.amd.com (Phil Ngai) (12/11/90)
bill@bilver.uucp (Bill Vermillion) writes: >Except the big difference in the magnetic media is there are many tracks, >with a high tpi density. In a CD-ROM there is essentialy ONE track, which >is a spiral. Primarily, I was responding to the claim that CD-ROMs are inherently slow because the use of Constant Linear Velocity forces a change in RPM when seeking. I agree that finding the track is a little more difficult on CD-ROMs than on magnetic media but the big bottleneck for seeks right now seems to be the rotational inertia of the disk, which could be bypassed if the methods used in magnetic media for accomodating varying bit rates were adopted. >In magnetic media you can say at track xx we will go to an xxxx bit rate >and when we reach track yy we can go to a yyyy bit rate. Simple physics tells you what the bit rate at radius X will be approximately. >In CDs the tracks are independant of the drives. And if you examine the >specifications for CD mastering you will find that there is a range for >spacing the "tracks". Therefore you can not position a CD-ROM the same Then perhaps as part of the startup process the drive could determine the "track" spacing, assuming that it is relatively constant on a particular disk and varies from disk to disk only because of different data lengths. -- There is no right more fundamental than self-defense.