In fact, some observers worried about such issues refer to an 
approaching "Digital Dark Ages", in which the late 20th/early 21st 
centuries will come to be looked upon as having as much of a dearth 
of documentary evidence as the "Dark Ages" of the fall of the Roman 
Empire until the 15th century (mistakenly, or at least, exaggerated) 
were.

There's a number of problems, both physical/hardware and interpretive/software.

No current media has a lifetime measured in more than decades.  It's 
not just that the hard disk may mechanically fail, but that, 
effectively, the "bits" will decay.  Similarly with CDs and DVDs (let 
alone diskettes), which have even shorter lifetimes.   So, not only 
do there have to be backups made against mechanical failure, but also 
backups over a longer period to just insure that the disks are 
earlier in their "lifetime", migrating to "newer" technology.

And then there's the problem that the software needed to read the 
media is no longer available - obsolete operating systems for 
no-longer manufactured entire computers or even just peripheral 
storage devices (think of the spinning tape drives of science fiction 
movie computers).  This is the intersection of the above physical and 
interpretive.

But then there's the format of the data/documents themselves.  For 
example, some of the early adopters of computer word processing may 
have used one of the most famous of the early ones, "Electric 
Pencil".  I doubt any current word processor still reads that 
document file format.

This problem has already had impacts.  Some of NASA's data from 
1970's Mars landers was stored on equipment that is no longer 
working, no longer available, and the software to read the tapes no 
longer available.  A government social service organization (I 
believe in the UK) found that all their data was on diskettes, which 
were either a) bad or b) did not have software to read the outmoded 
diskette formats.

A lot of people are noticing this coming problem, and starting to 
think hard about it (one of whom is even Stewart Brand, of Whole 
Earth and the Clock of the Long Now), as well as institutions, such 
as the Internet Archive (aka Wayback Machine).  Some of this is doing 
the data migration mentioned above; some of it is ensuring that there 
exists available translation software to read the old file formats.

But this sort of approach only works for large institutions.  How 
many of us who are "serious" - but individual - computer users can 
say we have a solid backup strategy against even mechanical failure? 
Let alone, make a point to migrate old media to new media, convert 
document files to newer file formats?

By the way, Nancy, I'm glad to see your mention of the importance of 
living in Serendip below.  This is the most acute argument against 
ordering books (or music) on-line, instead of browsing in stores, 
libraries, shops.  You can at most find what you're looking for; you 
don't find the serendipitous items that usually end up having an even 
greater impact.

At 11:06 AM -0500 1/6/10, Nancy Gish wrote:
>
>books be always available and more and more be made so.  But I 
>personally think paper books were an incredible technology likely to 
>long outlast electronic material: computers keep being updated and 
>crashing and older files cannot even be accessed on new 
>equipment.  Unless some uniform and sustained structure develops as 
>a standard, I fear all these digitized books will just disappear for 
>use even if they exist eternally in cyberspace.  (Any computer savvy 
>answers on the list?) As someone who loves libraries with books in 
>paper, I am dubious in any case; I find things that just would never 
>turn up on the internet (like the letters Eliot wrote to Maurice 
>Lindsay).  The problem with working on databases is that you only 
>find what you look for--I love serendipity, and much of my most 
>interesting discoveries come from that.  Who would ever find those 
>letters by deciding to put Lindsay and Eliot together into Google? I 
>doubt that would turn it up anyway.  (OK--I just tried it, and a lot 
>came up linking the names because Lindsay died last April and 
>obituaries mention the connection.  He died while I was in Scotland 
>and had already read the letters. But what is on Google might send 
>you to some of the stuff, not all.  And who would think to do that 
>anyway unless they read the Scottish and English obits?)

--