| OCR Text |
Show 95 scheme, as the background language for analysis. The work would then be to provide a better set of analysis primitives to install into the language, instead of making both a terse language and the primitives. The sixth one is that of recognizing more features of the layout, particularly active devices, so that ACRE could provide a complete SPICE netlist instead of a partial one. Currently it needs to be used in conjunction with a tool that can recognize the rest of the layout. The seventh one to consider is to provide a means to handle path abutment when paths meet; currently the tool ignores this. This would take some cross linkage between paths that could be considered when measuring the perimeter of a path. The next one to consider is a language issue, but the language needs to be truly compiled if it is to become as efficient as possible. It is presently interpreted, and the compilation would eliminate the need to work with userdefined extensions for speed. The language exposes more possibilities for measurements than are presently implemented, and foremost is the measurement for parallel paths where there is a coupling besides the common length, a measurement from one corner to the other side for corner fringing. Much of the simplification of circuits is done in explicit programming now, and a future enhancement would be to move some of the simplification primitives into the language as features at the point of establishing circuit connectivity. |
