"How quiet--
locust-shrill
pierces rock"

That's Stryk's translation of a famous Basho haiku. I would say it's excellent to have in the mix. And yet...

"Shizukasa ya
iwa ni shimiiru
semi no koe"

Here's my reconstruction, based on my reading of a dozen other versions of this poem and my very limited Japanese:

Quietness, silence [emphatic particle/cutting word "ya"]
Rock [particle "ni" = into] penetrate, seep
cicada [posssive particle "no"] voice

The cicada's voice has to come LAST in this poem, just like the "mizu no oto" (water's sound or "splash") has to come last in the most famous poem by this author. In both poems you have a silence being broken, and the last element in the poem is the noise itself. And a cicada, not locust, is the consensus translation for "semi." So I'd say Stryk's version is quite stryking, but I'd rather have it as one among many versions. He is far from authoritative.

He often switches the order of the images; this may result in a better poem in English, but it ignores something significant for readers such as myself. There are certain formulaic phrases in Japanese poetry that tend to go either in the first or last line. A translation should have certain transparency in this respect. If the first or last line is something like "kusa makura" (grass pillow, the traditional metaphor for traveling) or "hototogisu" (cuckoo) the reader should be able to see this from the translation, without even having the Japanese text handy.