3 ) Now use TUid's to identify field type
5 ) fields are now stored in their own streams using TSwizzle technology etc.
7 ) there's now a single concrete header, no more built-in stuff
9 ) there's now a field factory mixin which all apps that want to support fields (apart from page numbers in headers and footers) must inherit from.
11 ) improvements to storage of fields brings a reduction in in-memory requirements and also in code size.
13 ) TRomanNumeral class which converts from denery to Roman Numerals. T_ROMAN testcode tests this
15 ) the full set of fields now exists: page number, number of pages, filename, date/time