----- Original Message ----- From: "T Rittenhouse" <[EMAIL PROTECTED]>
> Now here is a related question, what percentage of ancient records remain > available today? That's a valid question. And it should probably be balanced against the percentage of analphabets in the populations in ancient times. > Another, what is the difference between a written record in a language no > one knows, and a computer record in an unknown format? You need different kind of geeks to decode them. :-) > And another, where is all that information that was lost in the year 2000 > change over (remember that non-problem)? Some people did a good job to make that a non-poblem, Tom. > And yet another, what is the difference between a destroyed piece of paper > and a destroyed CD. The former is envionment friendly? Oh, well you would have to destroy a lot of paper per CD, I guess... > In fact, commercial, and governmental hard drives are regularly backed up. > When new larger ones replace older smaller ones most often they just copy > the data onto the new one. When old file cabinets are replaced, quite often > the old files are just dumped into boxes and put into damp basements to be > forgotten. Which files are more likely to be preserved? Well not the ones on CDroms in the damp basement, for sure. I think your are twisting things to an angle to make your point here, but I shall not claim to see things better. I thinks this is an imoprtant issue, so I would like to give an example from an area of business I know something about. And where the issue of long-term storage is important. Pharmaceutical companies are obliged to keep all records regarding a drug from the days of the first experiments with the active ingredient, and until 30-100 years after the medicine have been removed from the market (depending on national laws). In Sweden-based Astra, they did a project on the logistics of long-term storage a couple of years ago. They found that to maintain a digital archive in a readable state for a 100 years would be extremely expensive, cosidering the amount of info that will accumulate between each format or media transferral. OTOH, a digital archive has very good availability. Information is searchable, so retrieval time is short. For long-term storage, however, they had too look at other options. A paper-based archive could be one solution, but it takes up a huge volume, and laser prints and photocopies don't last more than 20-30 years before the toner comes off. The best suited medium for long-term storage, according to Astra, was microfilm. Small volume, cheap and long expected life time. The downside is, of course, that info is hard to find. However, there is also a nice trade-off to be made against how often information is requested. One can accept that older and less sought-for information takes longer to find. So Astra concluded that the best way was to find some optimal transition age for information to be moved from one media to the next. IIRC, they allowed 3 years on network drives, then another 5-7 years on CDROMs as XML files, before everything was printed out and kept in paper format for 15 years. Then, everything should be microfilmed, and the paper destroyed. It is interesting that the total cost for Astra was lower for transfer to chemical media than to keep it digital. They even applied Moore's law to the size and price of new hardware in their estimates. Unfortunately, Astra never got around to trying out their logistics, as the whole project was stopped in the merge with Zeneca half a year later, but it goes to show that the issue of long-term storage is an issue that one shouldn't take too lightly. Especially with regards to the actual volume of digital info we now have. Jostein
