December 02, 2005

Trilobites & Kilobytes

I want to prevent cruelty to dead horses & avoid beating them whenever possible, and I hate scare tactics. But sitting on the floor of my apartment this week, trying to hoover the data off my old beige G3/266 and onto my PowerBook, I was reminded of why we’re bothering with this whole DNG thing. Innovation means change, leading to incompatibility, meaning that without some thought given to preservation, your work is at risk.
It was only 5 years ago that the G3 in question was my desktop workhorse, but in simply trying to recover its data I discovered:

  • AppleTalk transfer between OS 8.6 and 10.4 fails. The machines could see each other, but transfers would immediately stall. I get why Apple wouldn’t test this scenarios heavily, but still, it’s only been 5 years.
  • Connecting a current hard drive was out, given that USB and FireWire weren’t supported on this machine. And good luck finding SCSI components in a store now.
  • You can still track down a Zip drive these days, but the new ones can’t write Zip100 disks. Luckily they can still read them.
  • GoLive 8 (CS2) can’t read a site file produced by GoLive 4.
  • Self-running Director presentations no longer work (my fault, given that I don’t maintain Classic on this machine, but it’s an indicator of the transience of the work).

Sneakernet and Zip disks ended up providing the solution, and as I played three-card monte with the disks, I browsed Bill Bryson’s A Short History of Nearly Everything. Reading about the development of the fossil record, I had to smile. Picking through a circa-2000 machine felt a little like looking at the Burgess Shale, searching for signs of life.
In the end I was able to keep the bulk of my data, but the process offered some useful perspective. If talking to a machine from 5 years back was this tricky using the same platform, how would it be for one from ten years back? 15? And how will you get at today’s data in 15 years? Seems like a good case for making images & the edits done to them as open and portable as possible.

Posted by John Nack at 2:42 PM on December 02, 2005

Comments

  • Ben Reilly — 3:04 PM on December 02, 2005

    The problem is that AppleTalk is only supported for discovery and printing on 10.4- not for file transfer. You need to use AFP over TCP/IP, which I _think_ is supported on 8.6. The way to do it would be by connecting to afp://aaa.bbb.ccc.ddd

  • Cari Jansen — 4:17 PM on December 02, 2005

    Standardisation has to be amongst most viable answers to issues related to long-term digital data preservation.
    Archivists are running around bald headed, pulling out their hair in search of a solution for preservation AND access to gzillions of proprietary digital data formats that have arrived on their doorstep.
    How can we ensure that 100years from now when we, the creators, are no longer here, people or digital devices can still access, read and view such files?
    Would PDF/A, XML, DNG and other standards truely be the answer? Or has all been lost already?
    Your story indicates that even simple data-migration within such a short period of time poses issues.
    We can send a Rover to Mars, but can we preserve the imagery and data collected long-term?

  • Barry Pearson — 2:31 AM on December 03, 2005

    Cari, you ask “Would PDF/A, XML, DNG and other standards truely be the answer? Or has all been lost already?”
    I don’t believe there is “an answer”, in the sense of a single set of specifications that we should all be heading towards. Instead, I believe that archiving, to enable our work to be accessed in future, must always be seen as a process. Successful archives need active management.
    We know this applies to physical media, and John’s experiences are evidence. But it probably applies to file formats too. I believe the trick is to reduce the rate of change of file formats to be much less than that of the physical media and interfaces, etc. Then as we copy our files from those old ZIP discs or CDs or whatever, we should also check whether it is time to convert to the next standard formats.
    In 20 years time, or perhaps 10, when we have a standard for 3D moving images captured from occular implants, (“Just LOOK, and we will do the rest”), DNG will probably be inadequate for the purpose. We may need something much more suitable for quantum computing. But if we have managed to concentrate all of our raws into the DNG+XML formats, we can expect high quality converters between the new format and DNG. (Both ways – make a DNG image part of a 3D movie, or capture a still from the movie into DNG format).
    PDF/A, DNG, XMP, etc, can be seen as stepping stones, as an alternative to dead-ends. Or as the “hubs” in a “hub and spoke” model of how to get there from here.

Copyright © 2014 Adobe Systems Incorporated. All rights reserved.
Terms of Use | Privacy Policy and Cookies (Updated)