TANGOES Rex Games, Inc 2001 California Street San Francisco CA 94109 (415)931-8200 (c) 1991 -------------------------------------------------------- | \ / | | \ / | | \ / | | \ / | | \ / | | \ 6 / | | \. / | | |\ / | | | \ / | | 4 | \ / | | | \ / | | | \ / | |_ | \_,/ | | \ | 1 /\ 7 | | \ | / \ | | \ | / \ | | \ | / \ | | \ | / \ | | \ | / \ | | \/ 3 X | | \ / \ | | \ / \ | | 5 \ / \ | | \ / \ | | \ / 2 \ | | \ / \ | *--------------------------V---------------------------* All angles are intended to be mutilpes of 45 degrees. In particular, each edge will be parallel to one of 4 lines. The lengths of the sides are as follows [followed by area]: 1) 1, 1, sqrt(2) [1/2] 2) same as #1 3) 1, 1, 1, 1 [1] 4) sqrt(2), 1, sqrt(2), 1 [1] 5) sqrt(2), sqrt(2), 2 [1] 6) 2, 2, 2sqrt(2) [2] 7) same as #6 Total area is 8 in these units. Typical problem and solution: Form the pieces into two squares. Solution: We must have 2 squares of area 4, hence length 2 on a side. Note that the lengths on a side can only add to 2 if the horizontal lengths are rational; in particular, _all pieces must be rotated by 45 degrees_ from this initial position (followed by rotations which are multiples of 90 degrees) So for example, the square #3 is aligned with the big square. The medium triangle #5 must have diagonal parallel to the square's edge; with #3 in place, this requires the diagonal _be_ and edge, so #3 and #5 meet in the center. This leaves a triangular hole with area 1.5, which then must be filled with a small triangle #1. The remaining space has area 1-1/2, so must include the other half-integral area (#2), and then the piece (#4) of area 1 must take up the remaining space. Trying the several edges of length sqrt(2) against that edge of #5 leaves only this case (up to reflection and rotation of course): +-----------+ | \ 5 / | | 1 \ / | +-----+ 4 + | | / | | 3 | / 2 | +-----+-----+ Each of #6 and #7 must be half the other square. This is the arrangement in which the pieces are shipped. (Actually one ought to _prove_ that no big triangle #6, #7 is in the same square as the square piece #3. If it were, it would have to be the half above the main diagonal, leaving room only for #1 and #2. But placing #7 similarly in the other square nd aligning #5 along an edge of that square leaves room only for an area-1 piece _in the shape of #5_, rather than #4. Thus this is impossible.) So goes a typical proof: given a desired shape to make, scale it so that it has area 8, then calculate the lengths of various edges, each of which would typically have to be of the form a sqrt(2) + b (a,b integral); there are only finitely many unions of edges with such an edge length total.