For long term archiving, what metadata format(s) are most likely to be readable in 10, 20, or 30 years? Currently I archive photos by directory in a tree structure starting with year at the top level. The structure is YYYY/MM/DD/image.format. So it looks like something like this: ├── 1999 │   └── 01 │   ├── 01 │   │   └── eb2a.jpg │   └── 02 │   └── piEc3.JPG ├── 2000 │   └── 01 │   ├── 01 │   │   ├── IMG_0001.JPG │   │   ├── IMG_0012.JPG

The advantages of this system include:

  • No single directory will contain too many files and cause slowdowns due to filesystem limitations.
  • It's easy to locate photos by date.
  • It's simple and there are lots of tools that can sort photos this way based on EXIF data.
  • This system will most likely still be readable in 30 years as long as one regularly transitions the data to the latest filesystems.

The disadvantages include:

  • There is no way to tag/organize photos other than by date.
  • Images that have incorrect or missing EXIF data have to be sorted manually.
  • You have to remember dates in order to find anything.

In terms of longevity, I can think of no other system that is comparable in terms of suitability for archiving. Plain text files with metadata may work, but any current software to create/update such files may not be around in the future.

That being said, are there any metadata formats that can be used in conjunction with this system that are highly likely to still be around? Propriety formats are out of consideration on the grounds that there is no guarantee that any company will be around and continue to support any particular proprietary format.

The ideal metadata format should have properties such as:

  • Easy for humans to read without the use of special software (i.e. a plain text format with simple markup/formatting)
  • Wide support among current software
  • Based on a standard (e.g. something like EXIF)

Does anything like this exist? If not, where should I look for inspiration to create such a format?


1 Answer 1

  1. Keep your current date-based organizational scheme for files
  2. Add XMP metadata to the files, including keywords, titles, and descriptions
  3. Use a database to collect this centrally, and allow search and presentation based on the metadata

This has all of the organizational advantages you cite, with few disadvantages. XMP is XML-based, so not quite human-readable, although it's okay in pinch — and there is a huge amount of software support, and it is standards-based. It is at least as futur-proof as your image files themselves.

You can even do basically this with existing commercial software like Lightroom.

  • \$\begingroup\$ What kind of database do you suggest? I use Darktable which creates XMP files upon importing a directory. I also use Digikam which can recognize faces and tag photos. It uses XMP plus a database. I'm not sure if Digikam's database is compatible with anything else. \$\endgroup\$ Sep 8, 2014 at 6:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.