there are some things in PLR which have never been rewritten since I first wrote them 1-2 years ago. This means by rewriting we can likely improve those parts a lot.
As an example, @CamilloMoschner has ‘fixed’ the plate definitions, and now they are as good as done.
Another example is @JKH and I recently fixed iswap movements involving rotations on STAR. Highlighting: where we think the api looks good, we need to think about a good test.
We need to start focussing on stability soon, but we can’t be stable with code that is not good. I feel like when we have discussed & tested a part of the code, it’s usually rational and quite stable.
There will obviously be changes we need to make, we are never truly “done”, but they are usually not urgent if the existing version is rational and of good quality.
To track what parts we have ‘finished’ in this sense, what parts are wip, and what parts we need to do still, I created a gsheet:
Let’s add to this the following week and execute asap, targeting the end of the month.
When we have all these parts done, we can release version 0.2 (version 0.1 being ill-defined, dating aug '22-now, oct 5 '24). For changes after, we will keep track of how breaking they are and schedule them for releases in cycles.
Until the end of the month, we maintain high velocity (move fast and break things) to get ready.