Message-Id: <mailto:199508171858.NAA29063@library.wustl.edu> Date: Thu, 17 Aug 1995 12:53:55 -0600 From: mailto:sas@LANL.GOV> Subject: Re: Alas, Poor CDR; I've Known It Well To: Multiple recipients of list IMAGELIB
Imagelib folke -
> Bearman may well have been referring to the newly-announced HD-ROM,
> developed at Los Alamos. I for one, having held on to my vinyl,
> have no delusions as to CDR's permanence. Technology hasn't stood
> still yet; why should it now?
>
Yes, the HD-ROM R&D effort here does promise to ultimately eclipse both the information density and the longevity of any currently known media. However, by the time it is a proven, production technology, affordable enough for conventional uses, we will all have spent many years using extensions of current technology.
Also realize that HD-ROM, as it is called is a specific e-beam "microlithography" - like process which, in 5 or 10 or 20 years will itself be eclipsed by nanotechnology where individual molecules, atoms or very small collections of the same will be the mechanisms for storage and processing rather than the relatively gigantic collections of atoms used today in integrated circuits or even in the patterns carved in steel pins ala HD-ROM.
I work in an environment where the actual physical encoding of my information is almost irrelevant to me. I have not lost a bit of data in 15 years due to mechanical malfunction... This is not because the gear or the technique is that good but because we treat information AS information, not data. We cherish it, when we must trust it to systems that might fail, we maintain multiply redundant copies as well as good indexing schemes to keep track of where it is and so forth. Many of you do the same I suspect.
We do not expect CD-R, as a singular technology to solve the problems of longevity, density, portability, et cetera all as one simple solution. We already have plenty of projects which are calling for petabytes (global climate modeling at 100 meter resolution at daily intervals for 100 years with hundreds of variables). Our photo-archives since 1943, if scanned exhaustively would yield terabytes of images. Our Library, our historical records and engineering drawings, similarly large numbers.
What we are looking at for CD-R is meaningful subsets of that data for distribution purposes where the master data is NOT on CD-R nor does it matter much if an individual CD fails through abuse, neglect or bad luck. We are looking at using them where the relevant data changes periodically, where a network solution would be ok but often requires greater data rates or geographic dislocation (away from high-speed networks).
Hybrid solutions for archive, online access and quick, relevant access are what I envision. A CD-R which holds the latest "issue" of a set of information, with the same (and more current or archived or more complete) information on the network and redundantly stored, checksummed, refreshed on more conventional (mag disk, tape) media.
I realize this is a little off the subject... sorry to have wandered.
- Steve Smith mailto:sas@lanl.gov