Happy New Year, hope you all had a great Christmas.
Over December and the first half of January, we’ve been working through the comments and our email discussions to produce a new iteration of the specification for review. It’s still by no means a finished work, so there’s plenty more scope for alterations, additions and removals!
The new version is here: https://docs.google.com/document/d/1eQL1Guv0ihfxPJIIceLJk4l22cRpGbXTsNVSNqcEdOM/edit
The following summarises the changes that we’ve made:
- Expanded all the protocol operations, especially to include their error codes, and reformatted them to remove the confusing table layout
- Removed the “Add File to Content” protocol operation, as no longer necessary
- Added support for concurrency control, using ETag and If-Match headers.
- Added some definitions for the URLs and Document types for ease of use
- Expanded all the JSON-LD context headers for the 4 document types
- Provided field documentation for all the document types, and extended those fields as needed based on the comments/discussion on the first iteration of the spec.
- Expanded the section on Authentication and Authorisation based on our email discussion
- Expanded the section on File Segment Upload based on comments in the documentation and further analysis of the problem
- Expanded the section on By Reference Deposit to bring it into line with similar behaviours to the File Segment Upload
- Expanded the Packaging section to define the 3 core formats: Binary, SimpleZip and SWORDBagIt. For SWORDBagIt, we evaluated the RDA spec, and have produced something similar, but not complaint with, that profile.
- Added an example on how to deposit metadata formats other than the default sword format.
There’s a fair bit to look at, so I’ll leave that to sit with you for a week or so, and for you to add any comments you like to the doc, and then as before we’ll pick some key topics for discussion on this list.
To avoid any confusion, I have now disabled commenting on the old version. Most comments on the old version were resolved, and I’ve carried forward any comments which are still relevant.
All the best,
Richard
