NewEmdrosSchema

Discussions with Ulrik have led to a complete reconsideration of the Morph emdros database schema. The present implementation is:

  [Book
    [Chapter
      [Verse
        [Word
          [Morpheme]
        ]
      ]
    ]
  ]

This hierarchy means that the monad range of Morpheme is contained in Word, which is contained in Verse, and so on to Book. However, the current schema also includes object types which match the parts of speech present in the database: Noun, Verb, Suffix, etc. Objects of these types have exactly the same monad range as the Morpheme they are associated with.

The basic insight here is that the parts of speech are characteristics of Morpheme, i.e., alternate labels of Morpheme, and so are more properly represented as features of the Morpheme object type. In fact, the most efficient representation would be to take the entire parsing string of the legacy Morph record and assign it to a "Parsing" feature of Morpheme. The consequence of this would be a vastly decreased number of object types, hence a reduced number of instantiated objects of those types, along with a reduction in the number of enumeration constants.

Further, the Verse object type has the features book, chapter and verse. This was done so that a query:

  [Book