Well, the best bets are the collections of stuff from the old archives, together with the stuff from tokilili and kulupu pi toki pona and the handful of websites which one or another person has maintained for a time in tp. I don't suppose that the list of these is complete or up to date and the garnering of material from some of these sites is also running a bit behind (nothing new for several months now) but collectively they constitute a corpus of purported tp usage. Some of it is to be rejected out of hand, since it contains, for example, unauthorized words or constructions clearly against the rules . Others are out because they have been universally rejected. Others because they just don't make sense as they stand. The rest are presumably good but all open to revision for both form and content. Ultimately, of course, the decision is jan Sonja's -- for the time being, at least. There are not a lot of tough cases left, incidentally, but writing up the final report remains to be done.