Max Jones

A status update: - The conversion of the landing page, quick-start, user-guide, developer guide, and release notes are complete - The API documentation has been [structured for mkdocs](https://zarr--3118.org.readthedocs.build/en/3118/api/) - The...

@zarr-developers/python-core-devs this PR is now ready for review. I recommend reviewing it without striving for perfection because many people share the opinion that the current documentation is difficult to use,...

We're also using an extra private xarray method `from xarray.backends.common import _find_absolute_paths` to handle relative paths in [open_virtual_mfdataset](https://github.com/zarr-developers/VirtualiZarr/blob/014e9623b56df14166adb6cc19fc9833479f02bf/virtualizarr/xarray.py#L151) which isn't great.

> > we will still need the `fs_root` argument, because some pre-existing kerchunk files come with only relative paths. > > Actually could this just be a kwarg to the...

> this is a great point -- what API would you like to see here? What makes this kind of interesting is that the "cache the absence of a key"...

Thanks for all your work on Zarr Python 3.0! I know this is truly the last quarter mile (or km) before v3.0 and apologize for being so last minute, but...

@d-v-b do you have thoughts on how downstream libraries would interact with Zarr's concurrency limit? Would those set there own concurrency approaches (e.g., global or per function) or could there...

> I am concerned that the additions to the codec API in this PR will be disruptive to people who implemented custom Zarr V3 codecs, e.g.., anyone who defined a...

> - all the codecs in zarr.codecs.numcodecs produce JSON like {name: "astype", "configuration": {"dtype": "int8", "astype": "uint8"}}. Note two things: the name does not have the numcodecs. prefix, and the...

> But this PR makes a breaking change by ensuring that the v3 JSON form of such a codec uses the zarr v3 data type metadata. I agree with making...