Appeal No. 2006-1729 Application No. 10/107,628 OR gate placed between each input and at least an output wherein each input is coupled to the outputs differently. Now, the question before us is what Ivanov would have taught to one of ordinary skill in the art? To answer this question, we find the following facts: At page 1394, sections 1, 2.1 and 2.2, Ivanov discloses the following: Space compaction other than by MISRs has so far been dominated by parity checkers, which are linear and circuit-independent compactors [4], [5]. The popularity of parity checkers as space compactors follows from their generally fairly effective fault coverage and simplicity. On the other hand, parity checkers are quite costly to implement in terms of area since they are usually implemented with trees of XOR gates and the cost of an XOR gate is typically among the highest for 2-input Boolean functions. A k: 1 parity tree compactor requires k – 1 2-input XOR gates and thus amounts to significant area. Consider the general case of compaction where a matrix of test data D = [dij] of dimensions m x n bits is transformed into a matrix C = [cij] of dimensions p x q bits, where p < m and/or q < n. We denote the transformation operator Φ as a matrix operator such that C = Φ (D). We refer to the ratio m:p as the space compaction ratio and the ratio n:q as the time compaction ratio. The column index of the test data matrix D is referred to as the time dimension since it corresponds to the output bits from a single circuit node (primary or pseudo-output) resulting from the application of different input test patterns to the CUT. Thus, if Φ is such that C has its time dimension q < n, then time 8Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 NextLast modified: November 3, 2007