I am interested in the extension of DINA to the system of item selection scenario. In item selection scenario, a set of items are provided to examinees to select a part of items from the set. I think a smart examinee would select the item which required the attributes she or he has. If the item is satisfied at examinee's attributes, examinee will select it but there is a probability s of non-selecting it (like slip parameter) because another satisfying items were selected before. if the requirement of the item cannot be satisfied by the examinee, he or she won't select the item but there is probability g of selecting it (like guessing parameter) because no other satisfying items in the candidate of items exist. the g parameter can have the subscript of item and person. A smarter examinee will select the two kinds of items. One is the item which requires attributes close to what he has if no item satisfying him in candidate items. It should be modeled in g parameter so the parameter should be involved to the liklihood of zeta for expressing the pattern of satisfying items. The other is the item satisfying him which is easier than other items also satisfying him. It should be modeled in s parameter. pattern of satisfying items also should be included. However, the details for modeling this concept are still orgnized.