• Using modified 2012 Bangor data, with improved suggestions.
• Some improvements in spellcheck correctness as well: generated letter usage frequency from CorCenCC word frequency data.
• Package is now about a third the size of v1.2 (thanks to new 2012 data being smaller than 2004).
• All in a single dictionary (merged the temporary extra dictionary 'mwy' into 2012 data 'cy_GB').
• Improved tools for working with the dictionary ('teclyn').

Problems:

• Still working on issues regarding 'r 'n 'u 'i 'o 'ch 'th etc SFXes.