This PR adapts many of the suggestions from my private "sparse memories" doc, which I think merge cleanly with the ideas and rationale currently put forward in the overview. Concretely:
Adds an explicit mappable mode to memories (alternative to maplength, subject to discussion)
Adds core spec instructions for anonymous mappings
Adds "Suggested Implementation" section to clarify implementation details on major operating systems
Some of the ideas in my additions conflict with ideas already in the overview, but I think we can work these out quickly in discussion.
This PR adapts many of the suggestions from my private "sparse memories" doc, which I think merge cleanly with the ideas and rationale currently put forward in the overview. Concretely:
mappablemode to memories (alternative tomaplength, subject to discussion)Some of the ideas in my additions conflict with ideas already in the overview, but I think we can work these out quickly in discussion.