3 ms·
The most difficult issue I've encountered is identifying a format to use. Here are some specifications, which I think are widely applicable: * Readable 100 yea
by hackuser 10y ago
The most difficult issue I've encountered is identifying a format to use. Here are some specifications, which I think are widely applicable:
* Readable 100 years from now. If that sounds crazy, consider whether your family currently has photos 75-100 years old? Ours does, and I don't think it's unusual. Come to think of it, possibly the factor that determines the age of a family's archive is how long they have had access to cameras.
* Maximum fidelity: Complete and correct, over time. This involves resolution, color gamut and some sort of color correction (I'm not quite sure; I'm no color expert), and error correction for the ravages of time.
* Metadata: Stored in the same file as the image (otherwise the relationship between image file and metadata file will almost certainly be lost in 100 years), readable 100 years from now (UTF-8 seems the obvious choice), machine-readable and parsable for building databases, user-defined fields, and editable (to add and update information after the image is created).
* End-user control of data: That online service you're using might not be around in 100 years, or even in 1.
Maybe I'm missing one or two specs here, but does anyone know the recommended solutions? What do professional archivists use?
- moonfern 10y agoPdf/A and xml based formats in theory. And in general use open source software. In practice use the Odf format because I support that and it does what you want. And don't use pdf/A because too many implementations are pretending they produce pdf/A while they don't.