There’s been a ton of work on this in IPLD.

The problem as it relates to “collections” have been broken down into many conceptual data structures https://github.com/ipld/specs/blob/master/data-structures/multiblock-collections.md

We have schema’s for a hashmap and vector that are partially implemented and will be used on some real data in the next month or so.

And finally, the layout problems in representing large binary have been turned into a schema and implementation that is being used in unixfsv2 which is being used in reg.

Thank you! Also, note to self: reference ipld/specs#211

Is this meant to be closed or merged?

@hsanjuan it is meant to get done :) I didn't want to loose track of the previous convos.