Closed Bart-Ericsson closed 2 years ago
Additional information from Bart: An example of this you can find in the Commonalities WG folder, in the “deliverables” subfolder, there are documents on the Reference architecture, AuthN-AuthZ-concept, etc, which I believe are not deliverables, but “contributions”. They could eventually become deliverables, if agreed and approved, but since there is no process, it is hard to distinguish.
Agree. We should have 4 subfolders (contributions, working, deliverables, supporting documents) and describe the process in the ProjectStructureAndRoles file:
Anything to add?
I think it is good structure, but on which level should this exist, should it be part of documentation and code? Or should approved API code also be moved in deliverables?
I would create the subfolders on top level of each sub project or working group, having documentation, API_code and API_definitions below. In this way it is easier to manage releases and publish it.
I'll create a slide for approval of this change in the next steering committee meeting.
Steering Committee voted to use branches instead of different sub directories. New section will be added to ProjectStructureAndRoles.md with a description how branches shall be used to reflect the status of an artefact.
In the repositories it is not possible to differentiate between: ◦ “a contribution” to a WG, which is not yet discussed; ◦ “a contribution” to a WG, for which the WG has discussed, agreed and endorsed; and ◦ “a deliverable of the WG” We need a standard structure to deal with this, since today the documents are a mix of the above, and connected to this, who is allowed to move contributions from one stage to another?
Should there be a higher level for deliverable approval? e.g. the steering group?
Perhaps we can learn from other project setups, e.g. 3GPP but apply it to their github way of working. See: https://www.3gpp.org/DynaReport/21900.htm but I'm sure this is also dealt with in other open source projects