Even digital memories can fade

The problem of preserving digital photos and other electronic records for future decades confounds even the experts.
The New York Times

The story "Even digital memories can fade" published November 10, 2004 at 6:08 AM is no longer available on CNET News.

Content from The New York Times expires after 7 days.

[fmcgowan]

There are several problems here:

a hardware problem (Where can I find an Apple Lisa 1MB 5.25" floppy drive so I can read my old diskette?) and at least 2 software problems: device drivers (How can my current PC communicate with the drive?) and file formats (How can I read the proprietary Lisa "Write" document files?)<br /><br />The hardware problem may well be beyond practical solution if the "museum" approach fails. Indeed, NASA has millions of feet of magnetic tape containing data that cannot be read because the 7 track low density open reel tape drives that could read it are no longer available. Neither is the specific design information that would allow NASA to build one or two such drives to permit migration of the data to modern storage.<br /><br />The device driver problem is also probably already beyond hope of solution. Even if NASA could buy or build some of the drives it needs, the information required to build drivers for them may no longer be available.<br /><br />The file format problem for old PC files is probably nearing the point of impossibility as well. <br /><br />Our hope for the future lies in the fact that digital content is still in its infancy. What we will lose *now* is a time fraction of what we stand to lose in the future.<br /><br />Open file formats based on open standards are the only hope for the long term. Such formats would permit the migration from one archival medium to the next over an extended time period. <br /><br />Standardized hardware *may* make migration unnecessary, but a periodic "refresh" cycle will still be a requirement. In any case, *real* standards (as opposed to the "de facto" variety the IT industry has found so attractive in the past) are a *requirement* for the future. <br /><br />The British "Domesday Book Project" has already pointed to *all* of these issues (<a class="jive-link-external" href="http://www.theregister.co.uk/2003/07/11/bbc_domesday_project_saved/" target="_newWindow">http://www.theregister.co.uk/2003/07/11/bbc_domesday_project_saved/</a>) but the industry seems to have regarded this episode more as a curiosity than a harbinger.<br /><br />If you think this problem is not important to *you*, I hope you file your taxes on paper rather than electronically...

[wolf1420]

Worry about Video Tape Too

Digital Silo Inc., anonline digital media company, just today <br />announced the availability of the first affordable consumer <br />online video storage and sharing service, giving people an easy <br />way to permanently store and share their favorite home videos. <br />igital Silo Inc. Digital Silo Inc. (www.digitalsilo.com), is the first <br />online consumer digital video format conversion, storage, <br />archiving and streaming media service. The Digital Silo site <br />provides consumers with a secure and easy way to archive, store <br />and share their home videos with friends and family. The site <br />also provides additional services and creative tools to enhance <br />the viewing experience.

[Razzl]

don't believe the stuff about libraries

Firstly, don't buy this comment: "Professional archivists and librarians have the resources to duplicate materials in other formats and the expertise to retrieve materials trapped in obsolete computers". The author must have gotten that impression from talking to the people at the National Archives and Library of Congress but I guarantee you the number of Academic libraries throughout the world that have such capabilities could be counted on fingers and toes. Only law enforcement has such softwares widely available to them and preservation is not their mission for it. <br /><br />Other than that there are a lot of facets to this problem. One is a matter of perspective--how well did most of the information in earlier formats survive? How well are we doing in relation to that? 99.999% of the information that went through telephones in the past 130 years is gone, but now much of the same information is surviving today in email, so there's a net gain at this point. We agonize over the preservation of family photos locked in computers but forget that, all in all, little of that material really survived from the era of photography (if Kodak could tell us how many frames of 35mm film it sold over the 20th century I believe the survival rate would be as bad as the 99.999% of lost phone calls mentioned above). <br /><br />Writable discs are chemically unstable in their current form but the commercially produced versions are chemically inert and may be able to survive millennia when stored properly. The Compact Disc and CD-rom (the Sony/Philips proprietary format) are the most stable data storage platforms ever invented outside of rock carvings.<br /><br />Having a critical mass of material in existence improves the chances of a given technology remaining in play for the next generation to remaster it. The Compact Disc is a good example again; the number of playing devices and discs in use are so great that devices using later technologies are constructed to be able to read the old discs. Phonograph records should be as dead as a doornail but the vast number of discs and players floating around keeps the format alive. Standardized digital file types such as MPEG and JPEG ensure a long life to old digital files during a transitional period.

[hadaso]

Vynil is almost dead but revivable

It is true that right now CDs are a good way to keep things. But what will happen in a hundred years? If the CDs as media will not be used, then having no hardware available will make the CDs that can survive millenia quite useless. Hand building a CD player at a time they are not mass produced will be much harder than hand building a phonograph to play vynil records (and perhaps not possible at all). The phonograph records have the advantage that any researcher will be able to have a phonograph made available at reasonable cost (i.e, cost that a research grant can cover, and that can be justified). I am not sure the same would apply to a CD player. You'd need much more expensive equipment to manufacture those, even if you only need one.

[hadaso]

Anyone has a punched cards reader?

I have a box full of punched cards of programs I wrote around 1980. No card punch though. Their advantage over magnetic media is that a human can recover the data manually.<br /><br />On my PC I have a file called "Files from old Desktop" with files I copied from the old desktop I bought in 1998 and replaced in 2002. I meant to copy the files I needed to the new folder structure of the new PC, but ended up using subfolders of that folder. No time for rearranging things. Inside that folder I have a file named "Floppies" containing copies of all the floppies I had lying around in 1999. I copied all of them to the HD so I can store them away. Inside that folder I have another folder named 5.25" that contains copies of the 3.5" floppies onto which I copied my older 5.25" floppies sometime in the mid-90's. Some of the 5.25" floppies were already unreadable (CRC errors) when I tried to transfer the data. Some of them are from the early 80's. I still have all those floppies, but no 5.25" drive. I also don't have a punched card reader, of course, but I can read them manually. Anyway, I would not have much use now for the cool JCL procedures on some of those cards...<br /><br />I also have folders on my PCs of my home directories from several UNIX accounts I had in various universities during the 90's. Some of them contain copies of files from other universities. Among them is my email store from the 90's that I actully copied from one UNIX to the next, so it is one email storage system for my online life in the 1990's scattered in many files in mbox format around the copy of my UNIX folder structure (or perhaps directory structure?) I plan to copy all those emails sometimes and integrate them to my current mail store using IMAP (I downloded IMAPSize from broobles.com for this project).

[hadaso]

Copyright might prevent preservation

One problem libraries face that sometimes prevent archiving is copyright, and DRM. Libraries cannot keep copies of what was traditionally kept in them, such as scientific periodicals, that nowadays are moving to online subscription services. This means that copied exist only on publishers' servers, and by the time they lose their copyright protection they may be long gone, since the publisher might not find keeping them financially possible.

[LANjackal]

File conversion will save the day

What's all the panic about? Simply write code to convert files in older formats to newer ones, and have it run when a generational leap is made. Duh. Stop worrying.