“…To this end, one either has to improve the upper bound z b = O(bz log n z ) or has to provide a more elaborate series of examples improving the lower bound z b = Ω(bz) from Section 3 (obviously, the examples must deal with non-phrase-aligned parsings). We point out, however, that the tightness of the bound from Theorem 3 would necessarily imply the tightness of the currently best upper bound g = O(z log n z ) [4,18] from Lemma 3 that relates the size g of the minimal grammar generating the string and the size z of the LZ77 parsing for the string (the best lower bound up-to-date is g = Ω(z log n log log n ) [8,10]). Indeed, for a constant b > 1, if there exists a string whose LZ77 parsing has size z and whose b-block contraction can have only LZ77 parsings of size at least Ω(z log n z ), then the minimal grammar of such string must have a size of at least g = Ω(z log n z ) since, by Lemma 3, the string has a phrase-aligned LZ77 parsing of size g, and thus, by Theorem 2, the b-block contraction has an LZ77 parsing of size O(bg), which is O(g) as b is constant.…”