Most recent mention I can find:
June 2016 [webmasterworld.com]
IP: 202.180.34.abc
UA: ICC-Crawler/2.0 (Mozilla-compatible; ; http://ucri.nict.go.jp/en/icccrawler.html)
Headers: humanoid
robots.txt: YES
Request pattern--selected pages within a single directory--suggests a shopping list provided by an outside source. Several very short visits over a period of 1-2 days, each beginning with robots.txt, all from the identical IP.
Some details of the UA have changed over the years, but the distinctive semicolon pattern--which would get them blocked on some sites (looking at you, wilderness)--remains the same. And, unlike behavior reported in 2016, they DO ask for robots.txt. Since all their requests were for files in one directory, I can’t absolutely swear that they are compliant; I tend to doubt that the shopping list included other material from roboted-out directories which they decided against. But I generally haven’t had much trouble with Japanese robots.
