Google returns links that never existed
Google, in its present incarnation, returns URLs that, to paraphrase Apocalypse Now!:
do not exist... nor will they ever exist.
Because there are no web servers in the .sj domain, google has listed http://sj/ without consulting robots.txt. Or?did they try, but was told the domain has no IP? If so, they should also know that http://sj/ is an invalid URL?
The link to http://sj/ may actually work in browsers with the stupid auto-search feature, e.g. Mozilla FireFox. Not a good thing, because the URL visited may change unexpectedly.