Rename API methods - terminology changes

[bogovicj]

Block / shards / chunks

The latest alpha version uses "block" to refer to the smallest unit of nesting (that are compressed), and "shard" to refer to larger groupings of blocks. We are considering using "block" to refer to the largest unit (for writing) and "chunk" to refer to the smallest unit (reading). Initial discussion here

Codec types

We currently use these terminology that differs from zarr's jargon, but might consider changing these

• "BlockCodec" (for Zarr's "array -> bytes")
• "DataCodec" (for Zarr's "bytes -> bytes")
• "DatasetCodec" (for Zarr's "array -> array")

[cmhulbert]

Another contender for renaming: LazyRead to me is a bit of a misnomer, as the interface itself defines the active reading part of the ReadData that implements it. Not sure a good alternative, but it throws me off 😅

[cmhulbert]

We have also discussed before renaming ReadData, which I would be open to, but don't currently have a suggestion for alternatives

[bogovicj]

Start of work here #216