XLE produces a packed representation of all possible solutions as its output. This has a number of advantages over trying to pick the "right" solution:
"They saw the girl with the telescope."
In this sentence, "saw" is ambiguous between the past tense of "saw" (labeled a:1) and the present tense of "see" (a:2). Further ambiguity arises from the PP "with the telescope" as it can either be an adjunct of "saw" (b:1) or "the girl" (b:2). The contexts a:1 and a:2 are mutually exclusive. Similarly for b:1 and b:2. Moreover, a:1 and a:2 are independent from b:1 and b:2. So this feature structure represents four different solutions. XLE uses a contexted feature structure like this for each edge in a chart to represent all of the possible feature structures that the edge can have.