jkunze / bagitspec

31 stars 11 forks source link

Proposed changes for 1.0 #17

Closed acdha closed 6 years ago

acdha commented 7 years ago

Since we have a fair amount of discussion history in https://github.com/loc-rdc/bagitspec/pull/1 here's a pull request against your main repo to track any discussion now.

Please feel welcome to create issues at https://github.com/loc-rdc/bagit-conformance-suite if there are any edge-cases or error scenarios which should be better tested across implementations.

jkunze commented 7 years ago

Thanks!

On Fri, Feb 24, 2017 at 4:38 PM, Chris Adams notifications@github.com wrote:

Since we have a fair amount of discussion history in loc-rdc#1 https://github.com/loc-rdc/bagitspec/pull/1 here's a pull request against your main repo to track any discussion now.

Please feel welcome to create issues at https://github.com/loc-rdc/ bagit-conformance-suite if there are any edge-cases or error scenarios which should be better tested across implementations.

You can view, comment on, or merge this pull request online at:

https://github.com/jkunze/bagitspec/pull/17 Commit Summary

  • first cut at bagit 1.0 spec
  • updated ed summers, and justin littman addresses
  • Remove legacy comment about old DNS results for xml.resource.org
  • Trim trailing whitespace
  • Remove long-commented “other file metadata” section
  • Strengthen guidance on UTF-8 BOMs in tag files
  • Recommend modern hash algorithms
  • Update recommendations related to Windows filenaming
  • Update interoperability disclaimer
  • Discourage the use of manual bag creation
  • Add filename-related normalization discussion
  • Update "Payload-Oxum" documentation
  • Consistent number of spaces after a period
  • Upgrade must/must-not text to entity references
  • Strengthen wording: all manifests MUST list all files
  • Better references syntax
  • Update transfer recommendations
  • Consistent indentation in Terminology
  • Clarify the format for tag manifest algorithm names
  • Update interoperability reference
  • Update terminology for bag checksum algorithms
  • Consistent indentation for terminology list
  • Terminology: simplify “tag file” definition
  • Terminology: simplify “valid” definition
  • Convert numbered list in Structure section to
  • Remove Serialization section
  • Cherry pick: first cut at changes from recommendations by Dave Crocker
  • Use
    for ABNF diagrams
  • Minor
    consistency cleanup
  • Terminology: whitespace cleanup
  • Convert tabs to spaces
  • Terminology: update definition of “complete”
  • fixed some wording of sentences and removed section on Disk and network transfer since it doesn't really belong.
  • Remove stale references
  • Remove trailing whitespace
  • Remove stale inline TODO items
  • Normalize XML declarations against current IETF template
  • Prose review for section 1
  • Fix formatting for Section 2 (“Structure”)
  • Fix formatting for “Bag Declaration”
  • Copy-editing for “Payload Directory”
  • Copy editing for “Payload Manifest”
  • Remove injunction against reusing payload manifest algorithms
  • Copy-editing for “Payload-Oxum”
  • Simplify “Other Tag Files” text
  • Spelling
  • Copy-editing for “Complete, Incomplete, and Valid bags”
  • Update version number in example bag
  • Update authors
  • Clarify wording about directories in manifests
  • Remove duplicate wording
  • Update “filesystem” reference in “complete” requirements
  • Update “special directory characters” prose
  • Note that bag-info.txt fields are intended for human consumption
  • Update Makefile and add a format target
  • XML formatting

File Changes

Patch Links:

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/jkunze/bagitspec/pull/17, or mute the thread https://github.com/notifications/unsubscribe-auth/AAGaPEqQha1vxNLeUaoBDe8a-EYI4P3Vks5rfweNgaJpZM4MLaW4 .

acdha commented 7 years ago

One issue which came up in https://github.com/reeset/bagit_sharp/issues/1 is case-sensitivity on the bag-info.txt headers. Since the reserved names are specified as case-insensitive I'm inclined to say that implementations MUST preserve case as authored but SHOULD have insensitive access. Any thoughts?

cc: @johnscancella @edsu

acdha commented 7 years ago

@jkunze I updated this PR to restore fetch.txt

acdha commented 6 years ago

Closing this PR in favor of #19 reflecting the migrated source repo