release early / often, git branching model, CONTRIBUTING.md, (semantic) versioning about pixz

[vasi]

I think semantic versioning would be a reasonable idea.

A more complex branching model might be reasonable, but I think will only be necessary once the pace of development and number of contributors is higher. At the moment, I think it might just discourage contribution? You are welcome to try whatever branching technique you like for your own code, of course :)

[wookietreiber]

I do not think it will discourage contribution. For PRs, contributors only need to base their PRs on the develop branch instead of the master branch, because the master branch contains only tagged, production ready releases. IIRC if you set the default branch at GitHub to develop instead of master, even this will be done automatically whenever you are creating a new PR. So the added complexity is generally rather low.

[justinmayer]

Are there problems with how Pixz releases are currently handled that would be solved by a more complex branching model? If not, I'm just wondering whether this might fall in the "premature optimization" category.

[andreasstieger]

As a distribution package maintainer, semantic versions are good, maintain(ed|able) releases are even better.

As for the development model.. don't think a project is only viable if features are added. For a really good parallel compressor, there is only so much you can do. Don't make it brew coffee.

[cicku]

Ignore my previous comment here

[wookietreiber]

Closing this for now.