Named product properties in generated code

[dlurton]

This does not change the s-expression representation of products but changes the type universe grammar to require names for elements of products. Previously they were only required for records. The names will be used for the Kotlin properties which represent those elements.

The reason for this change is to improve the readability of the generated code. The unreadability of the synthetic names became very apparent in this PR, where for example, the synthetic name "expr2" is used to represent a let-body on a PIG-generated product type.

After this change, product elements are given names which are used only in the generated code.

This represents a somewhat significant change in the grammar of type universe files, even if it is a simplification. Namely, within sum types, the keyword product or record must now be used to indicate the type of variant-tuple. PIG used to detect the type of tuple by the syntax that was in use, but now that both products and records both require their elements to be named, another method was required.

// OLD product defintion:
(product int_pair int int)

// NEW product definition:
(product int_pair (first int) (second int))

// OLD sum definition:
(sum entity
    (slug)
    (android int)
    (human 
           (first_name symbol)
           (last_name symbol)
           (title (? symbol))  
           (parent (? entity))))     

// NEW sum definition:
(sum entity
    (product slug)
    (product android (unique_id int))
    (record human 
           (first_name symbol)
           (last_name symbol)
           (title (? symbol))  
           (parent (? entity))))     

Start by reviewing the changes in README.md, and review the rest in the order of your preference.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.