False positives in greek morphemes

[ghost]

Greek morpheme search occasionally has false positives if it cannot find a match for the word in the dictionary

• e.g. heartache finds no match in dictionary but has substring ache inside so is matched to Greek.

This may be a rare and excusable example as although its etymological history is middle english, the modern spelling of 'ache' is due to it being mistaken for Greek in origin (akhos - pain).

Proposed solutions:

• Removing morphemes from Greek dictionary that are too small/common as substrings of other non-greek words to be meaningful.
• Improving Anglo/Germanic/French/Latin/Arabic dictionary (need to locate reliable sources to scrape).
• Adding logic to check if a word is a compound word and seeing if it can be composed from words from our dictionary:
  • find all words that are 0-indexed substrings of our unknown word (he, heart)
  • checking if the remaining substring is also a word in our dictionary(artache, ache)
  • the first pair that matches gets highlighted as separate words (heart + ache)

[four-d-tesseract]

In this case, the program is finding tach, the Greek root for 'fast', inside hear**tach**e. A simpler and more general solution would be to force the Greek roots list to be more specific. When tach is a real Greek root, it's always followed by a "y" or an "o." I can remove tach from the list of matches and add tachy and tacho.

Removing morphemes from Greek dictionary that are too small/common as substrings of other non-greek words to be meaningful.

Already had that problem with oo, the Greek root for egg! I should review the Greek roots list for more.

This is something I'd like to handle myself, which I'll be doing in the next couple of weeks.

[ghost]

Ah yep - that's a much simpler solution!