Slashdot Mirror


Hard Drive Memory Lane

Chabil Ha' writes "CNET has gathered together some good old nostalgia from the photo vault. What high-tech product advances the fastest? It's probably the hard drive. The capacity doubles easily every two years and sometimes every year, faster even than the chip progress described by Moore's Law. The first drives took up storage closets. Now, a 5GB drive can fit in a phone."

17 of 156 comments (clear)

  1. Why oh why??? by countach · · Score: 1, Interesting

    I've always wondered why these advances in hardware ALWAYS follow such a relatively linear trajectory. I mean, instead of releasing a hard drive 2x the size of last year, why can't they skip a generation and release one 10x the last year? What was stopping them 5 years ago releasing a drive of the size now on offer?

    Seems almost like a conspiracy to have a continual flow of incrementally better product without going too far at once and leaving nowhere to go for upgrades. Because once they make the ultimate density hard drive there'll be only people replacing dead drives, and nobody upgrading.

    1. Re:Why oh why??? by NixieBunny · · Score: 2, Interesting

      The reason is that it's easier to redesign a gizmo to be just a bit smaller or faster than the current version. It's much harder to design something that's 10x better. The incremental improvements in technology are usually about 20-30%, which is a small enough improvement that the manufacturing people don't have to change their process too much. Think of it as an iterative process of making it just a little better; after 30 iterations you have an astounding increase in performance.

      --
      The determined Real Programmer can write Fortran programs in any language.
  2. Scoffing Posts Are From Those With Sort/No Memory by Kozar_The_Malignant · · Score: 3, Interesting

    First hard drive my emplopyer paid for was 5MB. First one I paid for with muy owm money was 40MB, and that was a trade-off for a whopping 4MB of RAM. If I'd gone with 1MB of RAM, I could have had a 110MB drive at the same price. At that time, RAM cost way more than drivespace, and that RAM let me multitask Quattro Pro and Paradox under DR DOS (I think you could actually do it with 2MB). Life was good!

    --
    Some mornings it's hardly worth chewing through the restraints to get out of bed.
  3. Re:Scoffing Posts Are From Those With Sort/No Memo by aussersterne · · Score: 5, Interesting

    In the '80s I paid thousands for a 5MB hard drive that sounded like an airplane engine and required three controllers: the servo/logic board, an MFM-to-SASI adapter board (yes, these really existed, for RLL and ESDI and to and from SCSI too), and a SASI/SCSI-to-host-bus board.

    I remember benchmarking the thing in excitement and getting a speed of 1 megabyte read in 96 seconds. A-W-E-S-O-M-E!

    Later I replaced it with a 5MB SyQuest removable drive (yes, there was a time when SyQuest made 5.0MB removable disks that were 5.25" to a side by about 1" high) that had a window on the front and weatherstripping on the door to keep the dust out. Unfortunately, all of those disks eventually developed bad sectors (despite the weatherstripping!) and by the mid-'80s I was running my BBS on an ST-213 10MB half-height (what we'd now call "huge") MFM hard drive in a PC, having become fully commodified in my computing self. ;-)

    --
    STOP . AMERICA . NOW
  4. Maximum speed by earnest+murderer · · Score: 3, Interesting

    Not wholly on topic, but this BBC article discusses the theoretical maximum speed of (modern) magnetic media.

    http://news.bbc.co.uk/1/hi/technology/3647055.stm

    2.3 picoseconds is pretty quick, at least until someone makes a faster material.

    --
    Platform advocacy is like choosing a favorite severely developmentally disabled child.
  5. Holographic pr0n? by quokkapox · · Score: 2, Interesting
    It still won't be enough to store all your holographic porn.

    Even today, I can store essentially all the music I've collected over the years that I really care about having at my fingertips all on my laptop, backed up on a handful of DVDs and my other machines if I'm paranoid about losing it.

    I can store everything I've ever written and all the digital images I've ever taken or scanned on another handful of DVDs. A few dozen hours of my childhood were filmed in Super-8, and have been converted to a bookshelf's worth of VHS tape in the late 1980s (which are now collecting dust) and digitized again to a handful of DVDs which are geographically backed up among several family members. Recordings of my grandparents singing folk songs, maybe that's another DVD worth of MP3 data. My dad and mom have separately scanned virtually all of the family photos back to the late 1800s. I could easily back up all of this content on my laptop and still have room to store the TV shows that are downloaded for me automatically and then normally deleted after watching.

    I'm not sure I will ever need more than a few terabytes. I'm not into holographic pr0n and I don't want a TV-quality recording of my life archived for posterity. Nobody is ever going to watch it. I can barely keep up with current reality as it is.

    --
    it's a blue bright blue Saturday hey hey
    1. Re:Holographic pr0n? by evilviper · · Score: 4, Interesting
      I'm not sure I will ever need more than a few terabytes. I'm not into holographic pr0n and I don't want a TV-quality recording of my life archived for posterity.

      Music is still only stereo, and most people are only storing lossy copies of it. When you have lossless 48 channel music at 384KHz, then we'll talk.

      How about video? Even with lossy MPEG-2, you can still only store a few dozen hours of HDTV on the largest hard drives. Switch to lossless video, or perhaps holographic, and you'll need a hell of a lot more space.

      We don't know what will develop. In a few years, will we all have full-fledged Earth Simulators running on our desktops, deciding when the next rainstorm will be?

      How about wearing a device that monitors EVERY neuron, every muscle fiber, etc., to be analyzed to determine if we are beginning to develop any health problems?

      Maybe a full copy of your own genome, which can be analysed in-detail by software.

      Perhaps with the development of software radio, we'll just set our computers to record ALL of the electromagnetic spectrum, and pick out anything we might want to watch/hear later.

      Maybe computer control of cars and servant robots will be possible, not because of wonderful A.I., but because every single possible senario being mapped to an appropriate response, and stored on a gigantic hard drive.

      Maybe we'll have our own personal "Hitchiker's Guide to the Galaxy", that detects that you're looking at a specific car, and automatically tells you everything there is to know about it, the company that made it, the driving record of the person associated with the license plate number, etc. Personal histories of every person you look at. Reviews of the movie poster you glanced at. etc.

      Or maybe a Matrix-like senario... You'd want to have a lot more movies if you could watch each of them each in a fraction of a second.

      Well, now I'm drawing a blank, but that's not bad for what I could come up with in a few minutes. I'm sure in a few years time I could have an incredibly long list.
      --
      Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
    2. Re:Holographic pr0n? by swilver · · Score: 2, Interesting
      Although storing everything lossless will definitely fuel the need for greater capacity storage for decades to come, I think that at some point people will be very happy with their lossy MP3's, AVI's and so on.

      It all depends on how good the quality is of the lossy storage:

      1) Lossless

      2) Lossy, but in such a way that you can only distinguish it from the original by looking directly at a waveform / video-still and being able to tell "yes, they're different", but without knowledge of which is the original you wouldn't be able to pick which one of the two is the original.

      3) Lossy, but in such a way you would be in doubt whether you are listening to the original or not (no direct comparison allowed, just tell me if this song I'm playing is lossless or lossy).

      4) Lossy, but with artifacts that could be distinguished by experts if they knew what to look for.

      For most people, this last level is more than good enough. At some point, you're just trying to figure out new ways of wasting storage -- I think that space could be put to better use in the form of redundancy :)

  6. Re:depends on how you measure improvements by cowbutt · · Score: 2, Interesting
    For the majority of use the speed of even slow drives is plenty fast. The main problem is that most computers have way to little RAM which means the system has to swap which slows the system in general.

    Actually, I'd dispute that. Even el-cheapo Dell models now come with 512MB of RAM. I have 512MB on my FC3 workstation and that still leaves me with ~200MB for buffers/cache in normal usage (Windows XP is a different matter, however).

    I'd argue that these days, architecturally speaking, the front-side bus is probably the greatest bottleneck, relative to the speed of CPUs. You can work around it by making caches larger and larger, but those caches need to be filled from somewhere, and meanwhile, it makes caching the wrong thing through mis-prediction even more expensive. This is even more pertinent to SMP systems that don't have dedicated memwory buses for each CPU.

  7. Dual Head Hard Drive? by nbritton · · Score: 2, Interesting

    Why not make hard drives with two heads per platter, It seems trivial to make it work? Just place the heads at opposite sides of the drive, shrink the platter a bit to accommodate the two heads, and implement a abstracted queueing algorithm so the two heads can work together.

    With SCSI's command queueing a dual head drive would at the very least double random read/write performance and access times. This would also make the drive sorta more fault tolerant because you only need one working head to read data off the drive.

    I want a 15,000-RPM dual head Ultra320 hard drive!

  8. Drum memory by ribuck · · Score: 3, Interesting
    I was disappointed that the article didn't mention drum memory, which was popular in the 1950s. The magnetic surface was on the outside of a cylindrical drum.

    Sometimes there was even one head per track (fixed in position) which improved performance by eliminating seek times.

    There's a photo of drum storage about halfway down the following article (which I found more interesting and more informative than TFA): http://www.moah.org/exhibits/archives/brains/compu terage.html

    1. Re:Drum memory by matt_wilts · · Score: 2, Interesting
      Check out this story of how to use drum memory:

      "Mel never wrote time-delay loops, either,
      even when the balky Flexowriter
      required a delay between output characters to work right.
      He just located instructions on the drum
      so each successive one was just past the read head
      when it was needed;
      the drum had to execute another complete revolution
      to find the next instruction."

  9. 1990 Computer shopper magazine by suso · · Score: 2, Interesting

    I remember seeing the first consumer 4GB drive by IBM listed in Computer Shopper for a whopping $20,000. Looking back, its amazing they didn't list the price as "Call".

  10. Re:Scoffing Posts Are From Those With Sort/No Memo by Ratbert42 · · Score: 2, Interesting

    My first personal hard drive was a 40 meg full-height in my IBM PC.

    Every once in a while I catch myself throwing 50 or 500 meg files around like they're nothing and think back to how many hard drives that is.

    My favorite "making me feel old" machine is my Palm Zire 71. It's got more RAM (16m), more "disk" (1g SD card), a faster clock (144MHz), higher resolution (320x320x16-bit) than my first 5 computers combined. If I had a decent PC emulator for it, it'd emulate all of my first 5 computers too.

  11. Re:depends on how you measure improvements by mrchaotica · · Score: 2, Interesting

    It's not nitpicking. Floppy discs are close to "paper thin;" that's why they're called "floppy." Hard drive platters are solid metal, at least 1mm thick, and quite rigid. They're probably either the heaviest or second-heaviest part of the drive (the cast metal casing might be heavier). And as far as resonance goes, hard disks can actually be used as speakers.

    --

    "[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz

  12. Re:1000x every 10 years by WuphonsReach · · Score: 2, Interesting

    you'll be able to get a 500 terabyte drive in 2016

    Not unless something bigger then perpendicular recording comes along. Today's GMR drives are roughly 60 gigabits / square inch, PR drives are shipping at 100-130 gigabits / square inch and are expected to top out at around 230-245 gigabits / square inch.

    Which puts the upper limit at around 2TB in a 3.5" drive.

    Anyone know what the next "big" thing is in magnetic storage? Or have they driven PR past the 245 gigabit / square inch level?

    --
    Wolde you bothe eate your cake, and have your cake?
  13. Re:Scoffing Posts Are From Those With Sort/No Memo by aussersterne · · Score: 2, Interesting

    Oh yeah, it was that slow. The machine was an old SSB machine and at the time you sort of had to get your own CPU and bus speeds working. It wasn't really overclocking, it was that the devices on your bus all had different characteristics or operating requirements and memory costs varied wildly depending on the speed of the chips you bought, so you often bought slow.

    So depending on the rest of your kit, you might desolder the stock CPU clock generator/crystal and solder in a slightly slower or faster one. I think I had mine with a 2.8MHz crystal (divided into two, meaning ultimately a reasonably snappy 1.4MHz machine), and this was an 8-bit machine with an 8-bit bus that had shared address/data lines (they'd alternate across clock cycles), so the drive and controllers (in spite of having three controllers in series doing bus conversion) were really still much faster than the CPU, system bus, and memory, especially with the 48k of 400ns memory I had installed in the machine.

    Hell, I could only do 1200 bps reliably on that system, which is one of the reasons I finally upgraded to a "new" PC in the mid-80s... I was seeking after that holy grail, 2400bps!

    --
    STOP . AMERICA . NOW