FIB@PSUVM.BITNET (12/15/86)
I've been accumulating questions to post regarding a system I'd like to put
together. Some may be easy, others may not. This is cross-posted to a number
of categories, but please respond to either comp.sys.ibm.pc , or directly to
me ... FIB at PSUVM.
1) Is there any way to lock out users of an AT system from portions of a hard
disk? I'd like two different groups of people to use it, but one group
has data that they would like to remain confidential. Can it be done in
software or hardware?
2) Has anybody had any experience (good or bad) with the Bernoulli Box? My
understanding is that it acts like a hard disk, but is removable. Is this
true? If so, it may be the answer to the above data security problem.
Also, any details on its operation would be appreciated. Since I may be
working with some good size files, how is access speed as compared to a
standard hard disk?
3) I would like to begin Desktop Publishing. I currently own a Xerox 4045
laser printer, but very few software packages support it. Has anyone used
or read any reviews on Xerox's Ventura Publisher software? Has anybody
used or seen any other packages that supports the Xerox 4045. I was at an
Aldus Pagemaker demo last week and was very impressed...only thing is that
Aldus is in competition with Xerox in this field, and probably won't
support the Xerox 4045 for "QUITE SOME TIME." Ventura looks like my best
bet right now, but I'd like some more input.
4) Is there any problem installing multiple hard disks on an AT system?
5) What is the general consensus on mice? What are the good and bad points
between BUS mice systems and Serial Port mice? How about Mechanical v/s
optical?
6) I'm also taking recomendations on low/medium price Letter quality (18-24
pin) Dot Matrix Printers. The only printers I've ever used have been
Epson... none of which have the quality I'm looking for. Suggestions?
7) Does anybody have any experience with using the AT as a remote S/36
workstation or up/downloading capability via communications software?
I'd like to hear anyone's experiences (harware/software) with this type
of application.
8) Finally, what are your feelings about the difference of buying a REAL
IBM-PC/AT v/s clones. I've been looking at the PC-Limited AT's with
a lot of interest. Does anybody have one that can tell me how they
perform? If you're really against or for buying clones, tell me why.
Whewww! What a load off my chest. Now all I have to do is sit back and read
the reviews. At any rate, I'm very serious about all of these questions, and
would really appreciate the input from the net.
Thanks in advance,
Glenn S Piper
Bitnet:FIB@PSUVM
Berks Campus,
Penn State Universityron@brl-sem.ARPA (Ron Natalie <ron>) (12/17/86)
In article <9073FIB@PSUVM>, FIB@PSUVM.BITNET writes: > 2) Has anybody had any experience (good or bad) with the Bernoulli Box? My > understanding is that it acts like a hard disk, but is removable. Is this > true? If so, it may be the answer to the above data security problem. > Also, any details on its operation would be appreciated. Since I may be > working with some good size files, how is access speed as compared to a > standard hard disk? Yeah, I wrote the driver for the IOMEGA 10.5 for XENIX. True it acts like any other SCSI hard disk except that it is SLOW! Despite what the advertisments claim, the data rates are no where near a decent hard disk. The cartridges are removable and damn near indestructible. The performance is at best that of a floppy disk. Note that in the dual drive model, you can't do overlapped seeks (move the heads on one drive while doing a transfer on the other) since the controller will refuse additional commands during a seek. In addition, it doesn't obey the SCSI protocol which required some additional hacks to the driver. > 6) I'm also taking recomendations on low/medium price Letter quality (18-24 > pin) Dot Matrix Printers. The only printers I've ever used have been > Epson... none of which have the quality I'm looking for. Suggestions? See last months PC TECH JOURNAL. They had an extensive review of printers with print samples.
madd@bucsb.bu.edu.UUCP (Jim "Jack" Frost) (12/18/86)
In article <9073FIB@PSUVM> FIB@PSUVM.BITNET writes: >1) Is there any way to lock out users of an AT system from portions of a hard > disk? I'd like two different groups of people to use it, but one group > has data that they would like to remain confidential. Can it be done in > software or hardware? You can do it, and you can do it in software or hardware. Hardware is the easiest -- buy a bernoulli box and lock up the disks (if you get a 20+20, you have 40 meg of online data and you can lock it up when you leave). Through software, there are lots of programs out there. I recommend VFeature by Bow Systems (I think). We have been using this for almost a year now with no bugs found (however we don't use the security mechanism that you want). It gives you the option of protecting DOS volumes with a password. It does this in such a way as to leave the drive inaccessible even if some smart guy uses his own DOS to boot. As an extra feature, it allows DOS volumes of >32mb, which is why we got it. >2) Has anybody had any experience (good or bad) with the Bernoulli Box? My > understanding is that it acts like a hard disk, but is removable. Is this > true? If so, it may be the answer to the above data security problem. > Also, any details on its operation would be appreciated. Since I may be > working with some good size files, how is access speed as compared to a > standard hard disk? The Bernoulli Box (I mentioned it above) is fantastic. It is not hyper- fast, but it has speed very comparable to your standard XT drive (65ms, I think, but you can check). I used one for awhile with absolutely no problems. Good points: * removable (security, multiple volumes cheaply and portable) * relatively fast * cannot have a head crash (because of bernoulli effect) * Easy backups Bad points: * Rather expensive to start up compared to a hard drive Check out the reviews in several magazines (PC-Week has one, I believe). I think they've all been pretty favorable. You can also boot off one of the Bernoulli disks if you buy a special controller (marginally more expensive, I think). >4) Is there any problem installing multiple hard disks on an AT system? Nope. We ran 1 85mb and a 20mb on one for awhile. Also installed: 1 1.2mb floppy and a 8" floppy controller. Everything ran together with zero problems. Even the power supply was adequate. >6) I'm also taking recomendations on low/medium price Letter quality (18-24 > pin) Dot Matrix Printers. The only printers I've ever used have been > Epson... none of which have the quality I'm looking for. Suggestions? We have a Nissho printer that runs great. Lots of paper usage with no problems (except one paper jam where the paper got caught coming out of the box). It emulates several different printers and has several faceplate/software switchable fonts and qualities. I'd recommend it. >8) Finally, what are your feelings about the difference of buying a REAL > IBM-PC/AT v/s clones. I've been looking at the PC-Limited AT's with > a lot of interest. Does anybody have one that can tell me how they > perform? If you're really against or for buying clones, tell me why. For clones: Most of them work fine in nearly all applications. No opinions on which are better or worse. 8mhz clones tend to be good, but stay away from clones w/ higher speeds, more below. Definitely cheaper. Against clones: Some have poor quality and substandard power supplies. Any clone that runs faster than 8mhz will have bus speed problems with many add-on cards. Some faster then 6mhz will have problems with some memory cards, but this is rapidly disappearing because of 8mhz pc-clones. For what it looks like you're doing, you can probably get away with a clone. Try to check with your periferal manufacturer to see if their card or whatever will work in your machine. If they don't know, ask the machine's manufacturer or try to borrow the same periferal from someone to see if it will work in your machine. We use coprocessor cards that are really finiky -- they REALLY want IBM equipment. Notes: We run an IBM PC AT w/ 85mb+20mb hard drives (under VFeature), an 8" floppy drive (using Flagstaff's driver and software), and Alloy's PC-SLAVE coprocessor cards (4 of 'em) using ATNX. Disclaimer: I am unrelated to: Alloy, Bow Systems, Flagstaff Engineering, IBM, IOmega, Nissho, and any other company that may have been mentioned. Good luck in your venture. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - Jim Frost * The Madd Hacker - UUCP: ..!harvard!bu-cs!bucsb!madd | ARPANET: madd@bucsb.bu.edu CSNET: madd%bucsb@bu-cs | BITNET: cscc71c@bostonu -------------------------------+---+------------------------------------ "Oh beer, oh beer." -- Me | [=(BEER) <- Bud the Beer (cheers!)
madd@bucsb.bu.edu.UUCP (Jim "Jack" Frost) (12/18/86)
In article <524@brl-sem.ARPA> ron@brl-sem.ARPA (Ron Natalie <ron>) writes: >In article <9073FIB@PSUVM>, FIB@PSUVM.BITNET writes: >> 2) Has anybody had any experience (good or bad) with the Bernoulli Box? My >> understanding is that it acts like a hard disk, but is removable. Is this >> true? If so, it may be the answer to the above data security problem. >> Also, any details on its operation would be appreciated. Since I may be >> working with some good size files, how is access speed as compared to a >> standard hard disk? >Yeah, I wrote the driver for the IOMEGA 10.5 for XENIX. True it acts >like any other SCSI hard disk except that it is SLOW! Despite what the >advertisments claim, the data rates are no where near a decent hard disk. >[...] The performance is at best that of a floppy disk. NO! When I used the Bernoulli Box 10+10, it ran plenty faster than a floppy, and with about the same apparent speed as the IBM PC XT using the supplied 10mb drive. I don't think the drive quite held up to the speed IOmega claimed, but it wasn't poor. The only thing that slows it down is when it stops spinning after 5min (I think) of inactivity. Then it has to speed up again, which takes some time. Note that I am telling this from the user standpoint, not programmer. To FIB@PSUVM.BITNET: Want to make your driver much, much faster? Have it seek for one sector on the current track if the drive isn't used in awhile. This way you will stop the BBox from stopping and will improve overall performance. Cost: Wear and tear on the disk, but minimal, and your driver will be a little bigger because you need junk area for the sector you read. I suppose you could reread the last sector and stash the data in the same buffer area that you had previously used, too. Note that the wear on the drive is negligible: they've run for weeks at a stretch in reviews with absolutely no problems. If there are technical problems to this, don't blame me. I used this technique on a driver for a floppy and it worked great. The controller would stop the floppy after 2 or 3 seconds of dead time, so requesting a sector ever second (by timer interrupt) kept it running. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - Jim Frost * The Madd Hacker - UUCP: ..!harvard!bu-cs!bucsb!madd | ARPANET: madd@bucsb.bu.edu CSNET: madd%bucsb@bu-cs | BITNET: cscc71c@bostonu -------------------------------+---+------------------------------------ "Oh beer, oh beer." -- Me | [=(BEER) <- Bud the Beer (cheers!)
ron@brl-sem.ARPA (Ron Natalie <ron>) (12/19/86)
In article <675@bucsb.bu.edu.UUCP>, madd@bucsb.bu.edu.UUCP (Jim "Jack" Frost) writes: > In article <524@brl-sem.ARPA> ron@brl-sem.ARPA (Ron Natalie <ron>) writes: > NO! When I used the Bernoulli Box 10+10, it ran plenty faster than > a floppy, and with about the same apparent speed as the IBM PC XT > using the supplied 10mb drive. Which is pretty god awful slow as well. > performance. Cost: Wear and tear on the disk, but minimal, and your > driver will be a little bigger because you need junk area for the > sector you read. Nah, the IOMEGA supports SEEKING without READING, it's just that the controller locks up while the seek is occuring so you can't do any other I/O.
mdf@osu-eddie.UUCP (Mark D. Freeman) (12/19/86)
In article <9073FIB@PSUVM> you write: >1) Is there any way to lock out users of an AT system from portions of a hard > disk? There are software packages to do this. I say buy a Bernoulli Box though. >2) Has anybody had any experience (good or bad) with the Bernoulli Box? BBs are as fast as an AT hard disk when used on an AT. They do not get head crashes, and the cartridges are portable and very rugged. Buy one! >4) Is there any problem installing multiple hard disks on an AT system? You can use two hard disks on an AT, each up to 32M. Of course, there are ways around these limits, but this is the stock answer. >5) What is the general consensus on mice? What are the good and bad points > between BUS mice systems and Serial Port mice? How about Mechanical v/s > optical? They are better than rats :-). Bus mice take up interrupts you may want for other things. On the other hand, you may want two serial devices AND a mouse. Mechanical require cleaning and can wear out, optical require a rigid pad to work on. >6) I'm also taking recomendations on low/medium price Letter quality (18-24 > pin) Dot Matrix Printers. Buy a NEC!!! I have a P5XL which I worship. I understand the P6 and P7 are good also and less expensive, although not as fast. >7) Does anybody have any experience with using the AT as a remote S/36 > workstation or up/downloading capability via communications software? > I'd like to hear anyone's experiences (harware/software) with this type > of application. CXI has a 5251 emulation board with file transfer software. I know very little about it, although I have installed one. It works. >8) Finally, what are your feelings about the difference of buying a REAL > IBM-PC/AT v/s clones. I've been looking at the PC-Limited AT's with > a lot of interest. Does anybody have one that can tell me how they > perform? If you're really against or for buying clones, tell me why. Do not buy an IBM, unless you routinely use three stamps to mail a postcard. I have heard wonderful things about the PCs Limited AT clones, particularly the 12 and 16 MHz models. -- < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > > Mark D. Freeman mdf@osu-eddie.uucp InfoSolv Corporation mdf@Ohio-State.arpa 6099 Riverside Drive ...!cbosgd!osu-eddie!mdf Dublin, OH 43017 Guest account at The Ohio State University (614) 761-8444 < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > >
news@cit-vax.Caltech.Edu (Usenet netnews) (12/22/86)
Organization : California Institute of Technology Keywords: mouse From: tim@tomcat.Caltech.Edu (Tim Kay) Path: tomcat!tim In article <2750@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: > > I say buy a Bernoulli Box though. > >>2) Has anybody had any experience (good or bad) with the Bernoulli Box? > > Buy one! Does Mark own stock in Bernoulli? >>5) What is the general consensus on mice? What are the good and bad points >> between BUS mice systems and Serial Port mice? How about Mechanical v/s >> optical? I have a suspicion that a bus mouse from Microsoft is just a serial mouse and an async adapter that uses non-standard ports. Can anyone confirm this? Timothy L. Kay tim@csvax.caltech.edu Department of Computer Science Caltech, 256-80 Pasadena, CA 91125
mdf@osu-eddie.UUCP (12/23/86)
Summary: In <1396@cit-vax.Caltech.Edu> tim@tomcat.UUCP (Tim Kay) writes: >In article <2750@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: >> >> I say buy a Bernoulli Box though. >> >>>2) Has anybody had any experience (good or bad) with the Bernoulli Box? >> >> Buy one! > >Does Mark own stock in Bernoulli? No. I'm just a satisfied user of both a 10+10 (the original model) and a new 20+20. On an XT, it is about as fast as an XT hard disk. On an AT it is about as fast as an AT hard disk. -- < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > > Mark D. Freeman mdf@osu-eddie.uucp InfoSolv Corporation mdf@Ohio-State.arpa 6099 Riverside Drive ...!cbosgd!osu-eddie!mdf Dublin, OH 43017 Guest account at The Ohio State University (614) 761-8444 < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > >
george@rebel.UUCP (George M. Sipe) (12/23/86)
In article <2750@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: >In article <9073FIB@PSUVM> you write: > >There are software packages to do this. I say buy a Bernoulli Box though. > >>2) Has anybody had any experience (good or bad) with the Bernoulli Box? > >BBs are as fast as an AT hard disk when used on an AT. They do not get head >crashes, and the cartridges are portable and very rugged. Buy one! I have used the Bernoulli Box for about 3 years now. I find them fast, reliable, and rugged. My only complaint is that the cost of the media has increased from the initial $30 to several times that. Low media cost was a big factor for me in buying them. Beyond that, no problems... never broken, never lost data, always interchangable, fast access, fast duplication. Highly recommended. I should note that I wrote my own drivers, my own copy program, and my own 'formatter' (don't ask, all in Z80 assembly). Many people don't realize that each cartridge has a (redundant) info area which the drive reads to learn how to handle it. The info includes bad track information, dwell timer info (the drive will idle the motor after 5, 7.5, 10, etc. minutes on a per cartridge - not drive - basis), two different ECC options, and sector interleave. If you don't 'format' (i.e. set these things specifically) you will be very lucky if the settings are suitable for your specific hardware and application, particularly for ECC and skew. Of course, that will make a substantial difference in reliability and performance.
ben@catnip.UUCP (Bennett Broder) (12/24/86)
In article <2750@osu-eddie.UUCP> mdf@osu-eddie.UUCP (Mark D. Freeman) writes: >>2) Has anybody had any experience (good or bad) with the Bernoulli Box? > >BBs are as fast as an AT hard disk when used on an AT. They do not get head >crashes, and the cartridges are portable and very rugged. Buy one! No way. The Bernoulli box is slower by several orders of magnitude than the voice coil technology drives available for the AT. The reason that Bernoulli boxes can't suffer from head crashes is because the r/w head always contacts the media, just like a floppy disk. No doubt the Bernoulli box is a good product, but its strengths lie in its removable media and ruggedness, not in its speed and performance. -- Ben Broder {ihnp4,decvax} !hjuxa!catnip!ben {houxm,clyde}/
ron@brl-sem.ARPA (Ron Natalie <ron>) (01/02/87)
In article <434@catnip.UUCP>, ben@catnip.UUCP (Bennett Broder) writes: > The reason that Bernoulli boxes can't suffer from head crashes is because > the r/w head always contacts the media, just like a floppy disk. Wrong, the bernoulli box is immune from head crashes because it the heads are designed to be able to hit the media, but they are not always in contact with the media. They achieve their faster turning rates over floppies by forcing air under the heads. > No doubt the Bernoulli box is a good product, but its strengths lie > in its removable media and ruggedness, not in its speed and performance. Well, it is better than floppies.
mdf@osu-eddie.UUCP (Mark D. Freeman) (01/04/87)
Summary: In <551@brl-sem.ARPA> ron@brl-sem.ARPA (Ron Natalie <ron>) writes: >In article <434@catnip.UUCP>, ben@catnip.UUCP (Bennett Broder) writes: >> The reason that Bernoulli boxes can't suffer from head crashes is because >> the r/w head always contacts the media, just like a floppy disk. >Wrong, the bernoulli box is immune from head crashes because it the heads >are designed to be able to hit the media, but they are not always in contact >with the media. They achieve their faster turning rates over floppies by >forcing air under the heads. Wrong, the Bernoulli Box is immune from head crashes because the media is designed to be in contact with the head (like a floppy). Also, when the drive sustains a shock, the media falls away from the head, because the air flow that pushes the media against the head gets interrupted. -- < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > > Mark D. Freeman mdf@osu-eddie.uucp InfoSolv Corporation mdf@Ohio-State.arpa 6099 Riverside Drive ...!cbosgd!osu-eddie!mdf Dublin, OH 43017 Guest account at The Ohio State University (614) 761-8444 < < < < < < < < < < < < < < < < < < < <> > > > > > > > > > > > > > > > > > > >
ralf@cad.cs.cmu.edu (Ralf Brown) (01/04/87)
In article <551@brl-sem.ARPA> ron@brl-sem.ARPA (Ron Natalie <ron>) writes: >In article <434@catnip.UUCP>, ben@catnip.UUCP (Bennett Broder) writes: >> The reason that Bernoulli boxes can't suffer from head crashes is because >> the r/w head always contacts the media, just like a floppy disk. >Wrong, the bernoulli box is immune from head crashes because it the heads >are designed to be able to hit the media, but they are not always in contact >with the media. They achieve their faster turning rates over floppies by >forcing air under the heads. Wrong again: The Bernoulli Box is so named because it uses the Bernoulli effect, in which a rapidly moving stream of fluid (either liquid or gas) experiences a pressure decrease. The BB can't have a head crash, because in normal operation, the flexible media is pulled toward the head by the Bernoulli effect on the air moving between the head and the disk surface (this is the opposite of hard disks, in which the heads are shaped so that are flowing under them lifts the heads off the disk). If anything interrupts that airflow, such as a piece of dirt, the Bernoulli effect is lost, the media is no longer pulled toward the heads, and the media is now safely out of harm's way. THE HEAD NEVER TOUCHES THE DISK SURFACE--if it did, the Bernoulli effect would immediately be lost, again moving the disk surface away from the head. -- +=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=+ | ARPA: RALF@B.GP.CS.CMU.EDU "Teaching COBOL ought to be | | AT&T: (412) 268-3053 (school) regarded as a criminal act" | | Snail: Ralf Brown --- Edsger Dijkstra | | Computer Science Department | | Carnegie-Mellon University DISCLAIMER? Who ever said I claimed | | Pittsburgh, PA 15213 anything? | +=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=+
drew@packard.UUCP (01/05/87)
In article <1029@cad.cs.cmu.edu> ralf@cad.cs.cmu.edu.UUCP writes: >In article <551@brl-sem.ARPA> ron@brl-sem.ARPA (Ron Natalie <ron>) writes: >>In article <434@catnip.UUCP>, ben@catnip.UUCP (Bennett Broder) writes: >>> The reason that Bernoulli boxes can't suffer from head crashes is because >>> the r/w head always contacts the media, just like a floppy disk. >>Wrong, the bernoulli box is immune from head crashes because it the heads >>are designed to be able to hit the media, but they are not always in contact >>with the media. They achieve their faster turning rates over floppies by >>forcing air under the heads. > >Wrong again: The Bernoulli Box is so named because it uses the Bernoulli >effect, in which a rapidly moving stream of fluid (either liquid or gas) >experiences a pressure decrease. > >The BB can't have a head crash, because in normal operation, the flexible >media is pulled toward the head by the Bernoulli effect on the air moving >between the head and the disk surface (this is the opposite of hard disks, >in which the heads are shaped so that are flowing under them lifts the >heads off the disk). If anything interrupts that airflow, such as a piece of >dirt, the Bernoulli effect is lost, the media is no longer pulled toward the >heads, and the media is now safely out of harm's way. THE HEAD NEVER TOUCHES >THE DISK SURFACE--if it did, the Bernoulli effect would immediately be lost, >again moving the disk surface away from the head. I certainly thought that the head never touched the media and that it worked exactly as described in the preceding paragraph. But then I got a new half-height model Bernoulli Box and found it ticked when idle. The user documentation explained that this was the sound of random seeks being automatically performed to prevent wearing out of the media from leaving the head over just one track. I wrote to Iomega asking about this and also asked why they didn't just seek to a dedicated wear-out track instead of making that constant tick-tick of random seeks. Their response wasn't too enlightening for me. They said that the heads don't really contact the "media" but there is a lubricant coating the media. The heads apparently do contact the lubricant. (Seemed like word-games to me!) A dedicated wear-out zone wouldn't be as good as random seeks because the wear-out zone would run out of lubricant. I don't know how to reconcile this statement with my earlier understanding about the air flow and all that. Can someone enlighten me? I do like Bernoulli Boxes. I've had some cartridges in use for years and so far only one cartridge seems to have "worn out" from long use. (It eventually started to exhibit annoying uncorrectable I/O errors). R. Drew Davis ihnp4!packard!drew
toma@tekgvs.UUCP (Thomas Almy) (01/05/87)
Well I had to comment on Bernoulli Box performance matter. I "Beta tested" both their original 8" 10 Meg drive and 5" 5 Meg drive, and found that the drive performance depended greatly on the driver code. I wrote my own drivers for the drives, and the system was a Compupro S-100 based with a 8 Mhz 8086, running CP/M-86. The drives have the rotational latency and seek times of slower hard disks, but the (voice coil) head positioners are not energized between commands. This means that a seek must be performed before each read or write. The servo positioners seek time is not really that much better for seeking the current track than it is for seeking across the disk! Drivers that read and write a single sector at a time run REAL slow, hardly faster than a floppy, because of those seeks. When full track buffering is performed (I buffered several tracks) then the performace is that of a slow hard disk with full track buffering! Which, by the way, is much faster than a fast hard disk without track buffering. Tom Almy Tektronix, Inc.
ron@brl-sem.ARPA (Ron Natalie <ron>) (01/06/87)
Even certain hard disks will crash if the head is left in one place too long and the head wasn't touching the media. There is still friction and that causes heat and that can destroy things. Two notable drives that exhibit this problem are the old version CDC drives that came with our Gould and the AMPEX video effects machine. -Ron
news@cit-vax.Caltech.Edu (Usenet netnews) (01/06/87)
Organization : California Institute of Technology Keywords: From: tim@tomcat.Caltech.Edu (Tim Kay) Path: tomcat!tim Thomas Almy writes (about Bernoulli Boxes): > >Drivers that read and write a single sector at a time run REAL slow, hardly >faster than a floppy, because of those seeks. When full track buffering is >performed (I buffered several tracks) then the performace is that of a slow >hard disk with full track buffering! Which, by the way, is much faster than >a fast hard disk without track buffering. I can see how this would work for reading. Rather than read a single sector, you might as well read the entire track and put it in a disk cache. However, when writing, you can't wait for the entire track because you might not get it. Do you cache writes to the current track until a seek is performed to a different track? A cache that is not write-through makes me nervous. Timothy L. Kay tim@csvax.caltech.edu Department of Computer Science Caltech, 256-80 Pasadena, CA 91125
berger@clio.Uiuc.ARPA (01/12/87)
The factory explained to me that the oxide coating is a polymer, and that the head contacts the top few molecules of it. That's why there's a 5000 hour lifetime spec for the disk cartridges.