Technical Team/Minutes/2012-02-14

Attendees

• Gary O’Neall
• Bill Schineller
• Kirsten Newcomer
• Jack Manbeck
• Savino Sguera
• Peter Williams
• Marshal Clow
• Ed Warnicke
• Rana Rahal
• Brandon Robinson

Agenda

• Large license text update
• 1.1 date
• Discussion on SPDX package definitions
• Discuss proposal for hierarchical supply chain (bug 818)

Update on license text

• Bug 1003 – could just keep the HTML formatting and update the pretty printer to interpret the tags
• Will implement the license text as external text files encoded in UTF-8. Spreadsheet will contain a link to the file.

Collab. Summit

• Agree we should have a tech team meeting at the Linux Collab. Summit in April

SPDX package definitions

Hierarchical supply chain

• Ed has been updating the Wiki page with additional details on 1.0 and 1.1 comparisons to proposed model.
• Discussion and review or Ed’s proposal
  • Enhanced model for Annotations
  • Mapping table of the SPDX 1.0 sections to the proposed model
    • Package file name – should this be mandatory? In the case of the archive file containing the SPDX file itself, does it makes sense to include the filename in the SPDX?
    • Package originator – may not required with the supply chain model. Discussion about different scenarios (such as there is no original SPDX file). May depend on people generating SPDX files for the code they bring in.
    • Discussion on annotation types – adding specific types to annotations will allow for compatibility with earlier versions
    • Discussion on alternative using subclasses for packages
      • Are the properties specific to a package or something more general?
      • Conflict with specificer model
        • Are we overloading the specifier with specifying the properties?
        • Does RDF already provide the mechanisms for concrete and references to the concrete data?
        • Agree we should call SPDX Element SPDX Licensable

Gary will do some homework on mapping the properties into a model where SPDX Package is subclassed from SPDX Licensable. Will also do some thinking about how to implement the specifier mechanism.