Alex Barker <[log in to unmask]> writes:
 
> I know this may open a can of worms, but have studies been done
> establishing the archival permanence (or lack thereof) of CD-ROMs?
 
There have been several good pointers, including the current article on
Digital media in the Jan85 "Scientific American."  What has been lacking
is a practical observation - the 15 year quoted average life for CD
media is for _manufactured_ CD media.  If an institution is planning on
producing its own CDROMs using CD-R (recordable) media, the life expectancy
is MUCH lower.  Simple explanation - manufactured CDs use cuts/"pits" in
a foil to represent the bits, while CD-R media uses dyes - the recorder
heats up and then burns a layer of dye in the CD-R media, to fake the
"pits."  This dye-based media is more heat and environment sensitive than
foil-based manufactured CDs.
 
        And there haven't been any good studies on the life of these disks - until
recently, CD-R was used for mastering CDROMs that were then manufactured.
Recently, prices on recorders have made it possible for smaller institutions
and individuals to buy the recorders, and it is now being sold as an archival
media.  And of course, a museum probably would rather pay to make a single
disk, instead of mastering and producing a minimum order from a CD
manuf., so CD-R is bound to be attractive.
 
        My personal experience has been with some CD-R media failing after about
three years.
 
        What I would suggest (and I think has been suggested already) is if an
institution wants to archive data using CD-R technology, they should
make at LEAST three copies of each disk, and refresh those disks every
three years to similar or newer media.  Furthermore, a tape backup using
analog recording would be a good failsafe backup - what affects the CDROM
would hopefully not affect the tape.  That tape should be periodically
refreshed as well.
 
        With a "multiple-source" plan such as that, it would take a major disaster
to  lose your data.  (Need I mention that these backups should be kept in
different sites, apart from each other?)
 
Greg Trimper
Technology Consultant