Improvements/aipreadme

User story

As a repository manager, I would like AIP's to be as self describing as possible, so that future users, with little or no information about Archivematica or what an AIP is, will be able to understand the structure and contents of the AIP's I produce now.

Status

2017-06-01 - New Proposal

Interest

If you'd like to get involved in this development, please feel free to contribute to this wiki page or start a discussion on our user forum.

Analysis:

Currently, Archivematica AIP's are structured as a Bag (https://tools.ietf.org/html/draft-kunze-bagit-14) and contain a METS file, which describes the contents of the AIP. Details about the Archivematica AIP structure are here: https://www.archivematica.org/en/docs/archivematica-1.6/user-manual/archival-storage/aip-structure/

METS files are machine readable, but are not human friendly formats.

Adding a human readable index or description into an AIP would improve the chances of a future user understanding the structure.

Archivematica structures AIP's in a specific way, but that is not documented within the AIP. Adding more explicit documentation about the structure would help users test that AIP's are valid, and help them to understand the structure.

There is a similar proposal outlined here: https://github.com/UTS-eResearch/datacrate

Use case: Add a README to each AIP

In the data/ directory (beside the mets file) add a README.html or README.md file. This would be intended as the first file to be opened by a human being trying to examine an AIP.

The README file would include

some boilerplate text, describing what an AIP is
links to the Archivematica documentation, to METS documentation, to PREMIS docs, etc.
a link to the METS file
optionally a link to a CATALOG.html file, that includes more detailed information about the contents of the AIP.

Use case: Create and Use a Bag Profile

https://github.com/ruebot/bagit-profiles

Archivematica could define a bag profile and reference this in the AIP's it produces. This would help make AIP's more easily machine readable.