For exam­ple, a well-known mis­con­cep­tion among stu­dents learn­ing about loop­ing is the expec­ta­tion that iter­a­tion can end in mid-loop, the instant a ter­mi­na­tion con­di­tion becomes true. And yet in this doc­u­ment I will present a lan­guage with loop­ing con­structs that work in just this way, are accepted as per­fectly rea­son­able by inex­pe­ri­enced pro­gram­mers, and are used by them to build effec­tive robot behav­iors. Does this mean that the old “mis­con­cep­tion” can now be dis­carded as another arti­fact of the arbi­trary require­ments of an imma­ture medium? Actu­ally, in this case and many oth­ers, I think some­thing much more inter­est­ing hap­pens. What used to appear as a rel­a­tively iso­lated mis­con­cep­tion is trans­formed into a part of a more gen­eral learn­ing chal­lenge with clearer con­nec­tions to the rest of the world. In this case, for exam­ple, chil­dren still need (even­tu­ally) to mas­ter the dif­fer­ences between the two kinds of loop­ing and the appro­pri­ate uses of each, along with par­al­lel dis­tionc­tions through­out the lan­guage. But in the mean­time, they are free to use a con­struct that matches their intu­itions and meets their needs.

Source: (Han­cock, Unpub­lished Doc­toral Dis­ser­ta­tion, p. 17)