Posting this here for future reference, as the Gemini documentation on the issue of tilted slits is ambiguous.
Spectroscopic slits for GMOS are designed on an x-y coord system, with the dispersion direction being along the x axis. The slits are described by their tilt angle "slittilt" (in degrees, with slittilt=0 for untilted slits oriented perpendicular to the dispersion direction) and by "slitsize_x" (the width of the slit in the x direction; the resolution of the spectrum varies inversely with slitsize_x) and "slitsize_y" (the height of the slit in the y direction). Untilted slits are rectangles, but tilted slits are parallelograms, with two sides of width slitsize_x in the x direction and two tilted sides of height slitsize_y in the y direction.
* The software "gmmps" displays tilted slits as parallelograms but with the WRONG WIDTH in the x-direction (it uses slitsize_x/cos(slittilt) instead of slitsize_x).
* The document "gmosmaskmakingv104.ps" draws tilted slits as rectangles; that is incorrect.