CD-rom longevity

The following discussion took place on PACS-L.

    From: "Philip H. Arny" <lrc1%umnhsnve.bitnet>
    Date: 25 Oct 89

    Also a question for archival status of indexes: does anyone have any
    idea of what the expected life of the CD-ROM media (the disc itself)
    is?  I've heard rumors that some of the early audio CD-ROMS are
    already dying...

    Philip Arny, University of Minnesota Bio-Medical Library, Minneapolis

    Date: 26 Oct 89
    From: Craig A. Summerhill <summerhi [at] wsuvm1__bitnet>
    Subject: CD deterioration

    There was some discussion of this topic on the MLA-L about the
    time I joined it in June.  Some of the vendors had stated that
    the lamination on early produced CD-Audio discs was
    deteriorating because of chemicals appearing in the dyes used to
    label the disk. The de-lamination of a disc can result in
    oxidation of the CD's metal core, which results in data error
    rates too high to be accounted for by the error correction
    devices currently used in CD-Audio playback units.

    I can't remember who issued the original report on this, but
    several companies in the industry damned the reports as being to
    harsh. However, all generally agree that some de-lamination
    could occur. Therefore, less caustic dyes are now used in
    marking the discs. Sony has also issued warnings against marking
    the laminated surface of their discs with indelible markers or
    adhesive labels that could cause de-lamination - an issue of
    particular concern to libraries where CD discs are loaned and
    need to be given identifying marks somehow.  Check an article in
    the June or July issue of Stereo Review)

    All told, the reports of deterioration are probably overblown.
    I think the interesting point made here on PACS (sorry I threw
    out the cite) was that the medium may become obsolete as a
    storage device at some point.  There is a whole new industry
    opening in the restoration and conversion of data "dinosaurs",
    which is likely to grow in the next few decades.  I find that
    interesting...

      - Craig A. Summerhill,
      Asst. Systems Librarian,
      Washington State University

    Date: 26 Oct 89
    From: pwillett%bingvaxc.bitnet

    There is an additional issue in counting on CD-ROM's for
    archival storage--that of changing hardware and software. We
    have some of the old (!-early 1986) Compact Disclosure CD's, and
    I have been trying to install them for a while now. The search
    software has changed considerably over the past three years, and
    the old disks will not run under the current software. I have
    the old software, and the problem, as near as I can figure out,
    is that it isn't compatible with the IBM PS/2 system, which
    wasn't in existence when the software was written. We no longer
    have a CD drive hooked up to an IBM PC, and it is too much
    trouble to switch machines for just one application. Thus,
    unless I can figure out a magic spell, I can't use the disk at
    all.

        **** Moderator's comments:   Please take the above with a
        grain of salt.  PS/2's are software compatible with
        IBM-PC's.  The problem here almost certainly lies elsewhere.
        The general point, of course, is well taken.

    Perry Willett
    Main Library
    SUNY-Binghamton
    PWILLETT@BINGVAXC

    Date: 27 Oct 89
    From: James Jay Morgan <izie100%indyvax.bitnet>

    I haven't heard anything recently, but a couple of years ago
    estimates were that the cd-rom would last 10 years or more.
    Since our vendors are constantly updating their medline indexes
    we havn't had any one disk for longer than a year before it's
    replaced.  I would however assume that a prudent vendor would
    maintain a master for when he or she runs out of copies.  Our
    vendors have been quite good about giving us free replacements
    for disks damaged by users.

    I also have some question about the archival status of the paper
    copies of indexes.  Aside from the problems of deteriorations
    with time and normal use, making copies of the data in machine
    readable format is a devil of a task.  OCR technology, even at
    the Kurzweil machine level, is primitive and costly.  Copying a
    cdrom onto a magnetic disk or WORM drive is easy and cheap
    compared to scanning the cumulated index medicus into machine
    readable format.  If you want to try backing up your cd, you
    might also try one of the public domain archive programs.  Since
    it's mostly text data you might be able to archive many 500 MB
    disks in about 200 mb of space, and then back the archived files
    up on tape.  I haven't tried this myself, since my own opinion
    is that cdrom is more reliable and more available than magnetic
    or tape storage.

    Jim Morgan
    IU SCHL. OF MED. LIBRARY

            **** Moderator's comments:  A few definitions:  The Kurzweil
    machine is a (still rather expensive) device that can scan
    printed text and convert the image of the text into machine
    readable and searchable characters.  It is the state-of-the-art
    Optical Character Recognition (OCR) system.  A WORM (Write Once
    Read Mostly) (or Many) drive is a *very* high capacity storage
    similar to a cd-rom but unlike a cd-rom it is easy to write to
    it, so it can be used like a hard-disk.  When you write new
    information to it, it doesn't overwrite the existing
    information, as a harddisk does, but appends the new information
    to the end of a text, as in a manuscript scroll.  Thus it
    combines the high capacity of read-only cd's with some of the
    convenience and economy of read-write harddisks.

    Jim Morgan
    IU SCHL. OF MED. LIBRARY


    Date: 29 Oct 89
    From: Walt Crawford <BR.WCC%RLG.BITNET [at] Forsythe__Stanford__EDU>

    Regarding "laser rot" and delamination...
    And, sorry, I can't give actual citations for this information.

    1.  The reports of "laser rot" (two to four year old CDs
    becoming unplayable) apparently originated in Great Britain,
    from within the hardcore group of digiphobes (those who feel
    that CDs are a sacrilege, that only vinyl reproduces sound
    properly, that if God had meant us to use digital sampling...
    well, you get the drift). Note that the fringe audiophile Brits
    include one supposed "scientist" who asserted that playing
    digitally-RECORDED vinyl discs would cause hairline fractures in
    your turntable's metal members, and another person who is
    marketing bits of foil attached to clips that, when attached to
    various objects in the listening room, supposedly improve sound
    quality. We're dealing with serious voodoo science here...

    HOWEVER, there WAS a problem with one or two small and
    apparently stupid CD pressing plants using caustic inks, which
    eat into the label side of the CD, which is actually much closer
    to the information layer than the "rainbow" side. Other than
    this particular stupidity--which, indeed, raises a problem with
    applying labels or using permanent inks anywhere within the
    information-carrying portion of a CD--there have been NO
    reputable reports of problems. [It is possible to inscribe
    labels on CD-ROMs within the clear inner band, with some care;
    tools are now being marketed to do precisely that.] (This
    information comes, via memory, from a number of issues of
    Stereophile Magazine.)

    2.  LaserVision discs are NOT physically identical to CDs; they
    are two-sided laminates--basically like two CDs mounted
    back-to-back--and, in the early years, there were indeed
    problems with the laminating process, causing some LV discs to
    "come unglued" (effectively). Those problems have also been
    solved; in any case, CD manufacture is simpler, since there is
    only one information layer.

    3.  HOWEVER: to the best of my knowledge, NO manufacturer of
    polycarbonate/aluminum/resin digital media will assert a life
    span of more than "ten to thirty years." They just don't know
    enough yet to know whether CD is an archival medium; at the
    moment, the assumption is that it is NOT. As has already been
    pointed out, it should be possible to make perfect copies of the
    discs--assuming that some agency has a scheduled
    conservation/preservation program to do so.

    Hope this helps. The orphan-technology problem is really more of
    a problem than short-term self-destruction of the medium; I'll
    admit that I didn't realize we were ALREADY running into
    "orphaned" CD-ROM at the software level (!). Thanks; I'm doing a
    talk on how technology affects service and preservation, at the
    California Library Association, and I'll use the information.
    -walt crawford, rlg: br.wcc [at] rlg__bitnet-

    Date: 29 Oct 89
    From: Walter Giesbrecht <WALTERG%YORKVM2.BITNET [at] Forsythe__Stanford__EDU>

    The problem is not the polycarbonate (as far as I know) but the
    aluminum reflective layer, which will oxidize, given half a
    chance. I have heard of some CD-Audio manufacturers using gold
    instead of aluminum in order to ward off the bogie of oxidation.
    On the other hand, I also remember reading about one maufacturer
    who intended to use copper, on the theory that copper-plated
    discs would become useless even faster, and thereby have to be
    replaced even sooner.

    A number of companies are not putting much stock in these
    rumours. Nimbus Records (I think) warrants their classical CDs
    for 100 years. Whether Nimbus will be around in 100 years is
    another matter.

                                  ***
                   Conservation DistList Instance 3:3
                Distributed: Saturday, October 28, 1989
                        Message Id: cdl-3-3-004
                                  ***