Storing your digital self
Christian Lindholm had an accident with his digital media archives and now he's asking how other people deal with it.
My answer is simple: never rely on proprietary software. I don't use Lifeblog, or any other "digital life management" solutions. While I do use iPhoto, and iTunes, I only use them as a viewer. I don't add any metadata - if I do add something, I just usually just name the picture. Other than that, it's plain, basic JPEG or PNG, AAC or MP3 with metadata in the file or the filename itself.
The problem with proprietary software is that they always try to keep you as a user - and therefore they mostly never provide export functionality. Switching away is a problem, and it is also a problem if you ever lose that software, it is broken or you cannot use it for some other reason. I use Windows, Mac and Linux and keep switching between them all the time. Therefore there's no way you could use the same software on each. So you quickly learn not to rely on a particular piece of software, but try to keep yourself as agnostic as possible.
Same thing, obviously, goes for text documents (UTF-8 or Latin1 plain text, thankyouverymuch; or Word97 [it is really rather universal these days]). Incidentally, that's also one of the reasons JSPWiki ships with a plain file text repository as the default - it's really easy to move away from it, if you really need to. No weird-shit database schemas; just plain, unambiguously named files.
As a backup solution, I use a three-way synchronization solution with Unison - one replica on my laptop, one on my desktop and one in a remote server. Since hard drive space is so cheap, it's far easier to manage these automatically than it is to keep making physical DVD backups or drag around USB hard drives. Replicating is the best backup solution I've so far figured out. If I really needed to, I guess I could buy some HD space from a hosting company as well to store an additional replica.
I've had a couple of crashes, and so far I've been able to restore everything that I care about. My replication needs are not as difficult as Christian's (I only got about 30 GB of data I don't want to lose), but this works well for me.
|
More info...
Comments?
Back to weblog
|
| "Main_blogentry_301207_1" last changed on 30-Dec-2007 21:45:34 EET by JanneJalkanen. |
Comments
Hi Janne, Thanks, the thing with metadata is that it allows for great viewing and searching experiences. I had created lots of metadata and could find anything in seconds ;-) Now I have the file names sorted in yearly folders, so not too bad.Yesterday the farse continued, I had Bootcamp installed on my mac, removed it by formatting the partition, then I tried to reclaim the space into a single big hardisk, now I have a ghost drive. I need to go to the Geniuses at the Regent Street Mac shop. In my despair I rushed to buy a new HD my first Terabyte disk. It took 20 years to go from 20MB to 1TB of personal computing space. A bit of a milestone.
It will be interesting to see how Nokia will solve this with Nokia Photos and the on-line assets. It is a hard problem to solve and there is a bigger social responsibility involved. Lets hope hardisks is not the papyrys of our civilisation.
--Christian, 31-Dec-2007
Yupyup, we have adequate common format support (JPEG and PNG is ready by everybody, and the spec is open and public and could be implemented from scratch even years from now), but all proprietary programs insist on building their own metadata structures. Which makes metadata essentially useless, IMO.
It's a tradeoff: whether you want a better experience now, or whether you want longevity of your data. I'm opting for longevity on the stuff that is important.
I really think that the only way to deal with this is automated metadata generation. For email, I don't bother with metadata - Apple Mail or any other mail program already has pretty good search functionality, and the email storage I use is commonly understood (replicated, of course). MP3's can be easily tagged with online, automatic services. Google Video even employs some speech-to-text services to index video - but there currently are no automatic ways to index images. GPS and other geotagging tools will help, but when will we see them for real?
--JanneJalkanen, 31-Dec-2007