ACM digital library, maybe start with SIGCOMM as it would be easier to deal with, at a small scale.

A lot of implicit context in artefacts.

Reviewers should be able to ask for more artifacts.

Artifacts should be in Open Access

What problem do we try to address: How can we get the artifacts we need?

  - where/who to store knowing it has to last for long periods, and
    verifiability
      - permanent reference (DOI, link that can be used as indirection so
        if data move, URI doesn't change).
      - provide an archive/snapshot
      - versioning so you can see how it change but keep access to original 

  - how to know what to provide as artifact (level of ?) how to describe it

  - BCP per techno of sharing to be sure it has been used correctly
    (external dependencies).

  - agree on a benchmark that everyone can competes against

  - how to describe the artifact
      - encourage/allow to provide supplementary materials describing
      - SIGCOMM to facilitate the pipeline, not only the camera ready,
        but also supplementary material. It evolves with time, at review
        you may have to change, after publication, ...
      - just human text or just dumps are not enough

  - get feedback from the community to refine and understand what is
    really necessary

  - anonimity

Sharing profile (sharing strategy for the artifact), if reproducible then
discuss it what you put in place for reproducibility, maybe a checklist.

Tags:
Created by Damien Saucez on 2017/08/28 11:41