Tuples TODO

• tuple field names (aliaes). One can access a field using alias with tuple.name
• default field types, including a default field type for the entire tuple, used for the variable part of it
• indexing of advanced field types: signed integer, double, date, money, geospatial types (point, polygon), case-insensitive indexes
• indexing based on field path, not field number
• association of pack/unpack functions with field types, to be used when converting to/from Lua or the binary protocol.
• space schemas as constraints on what tuples it can accept: number and type of fields, etc. If a schema is set, it could also be used to speed up field access in a tuple.