| Question asked
|
In Section 3.4, the paper writes: "Typical … In Section 3.4, the paper writes: "Typically, filtering A′ creates distinct partitions, and we apply the tableau algorithm to each partition separately. If all of the partitions are consistent, then we are done. Otherwise, we need to check A..." (actually, it will be "we need to check the image of the inconsistent partition, say image(Ap'), in A" according to the rest of the paragraph). IMHO, the reason that we have to double-check image(Ap') in the original A is that the inconsistency in Ap' might be an artificial inconsistency introduced in the process of summarization. So, my question would be: if we can keep track of whether a specific inconsistency is caused artificially or not, does it mean that we don't have to re-check image(Ap') in A? we don't have to re-check image(Ap') in A?
|
