Unverified Commit 3bc8b067 authored by Loïc Dachary's avatar Loïc Dachary
Browse files

remarks from NLnet

parent ad69f945
Hello Loïc,
I've a few remarks regarding the MoU. Apart from the remarks, it looks good.
The links to the SH forge are fine, but the plan should be clear without them.
Some descriptions such as 'IO throttling' are too short.
'SWH: Learning the codebase'
This is not a deliverable that we can check. Every part of the plan should
have a deliverable such as code or documentation. This is how we who to the
funders what was done.
Time spent learning a codebase and other tasks that do not result in public
deliverable should be taking into account in the tasks that do result in
public results.
'create the {PostgreSQL,Ceph} cluster' and 'Self-host Software Heritage on
grid5000'. These are a setup tasks and not a result that the public can reuse.
'IO throttling' I assume this will result in scripts that people can run
locally. Perhaps the require some VMs to be setup. Scripts to set up such an
environment can be a deliverable if they are documented and tested to be
runnable without access to SH internals.
'User Research on Gitea export' Is the deliverable a report of the user
research? Or perhaps .feature files documenting the expected user experience?
Please make that explicit.
'User Interface for Gitea export' appears twice. Perhaps give them version
numbers.
Can you be more specific in the recurring tasks? Will you write a series of
blog posts? What will be the deliverables?
Best regards,
Hi [redacted],
Thanks for the guidance. Although I was careful to have a deliverable (code, documentation, etc.) for each subtask, it was sometime not trivial to find. I kept the same structure for the workplan, exactly, but reworded the subtasks for clarity. I also failed to realize it had to be self-sufficient (i.e. not rely on external content).
The deliverable for "learning the codebase" was a contribution to the existing Software Heritage codebase in order to prepare the ground for the implementation of the object storage. In my mind the most valuable part of the work is learning the codebase intimately. But since this is expected to be integrated in other deliverables, I renamed it to focus on code contribution instead.
'create the {PostgreSQL,Ceph} cluster' and 'Self-host Software Heritage on grid5000' deliverables are Ansible modules published in a repository and that can be run against physical machines or virtual machines (convenient for CI).
'IO throttling' is actually an optimization of the object storage that is added after the initial implementation. It is a second iteration of the codebase and the deliverable is an updated version of the codebase. I clarified that in the substask.
'User Research on Gitea export' is indeed a user research report and I clarified that.
'User Interface for Gitea export' appears twice and I followed your advice to give them version
numbers. They need to be implemented a few months apart to allow for user feedback.
The recurring tasks are done monthly and the deliverables include a blog post with a detailed summary of the work done in the past month, the recording of a public videoconference organized to discuss the report with the community, merge requests, and blog posts with codewalks on the codebase. I clarified each aspect in the description.
Here is the updated MoU.
Cheers
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment