Language for the ages

I totally agree. IMHO, it's a lot better to firstly put a 'header' on each routine giving the purpose and the entry/exist variables (these can refer to other comemnts in the program or related programs -- e.g. if you're searching a disk directory, there's no point in giving the directory entry format every time it's used) and secondly to give meaningful comments to instructions. Rather than ADD R0,#8 ; Add 8 to R0 it's more use to have ADD R0,#8 ; Point to next 8-word record I feel the same way about scheamtics. Much of the time spent _usefully_ producsing schematics (particularly when reverse-engineering) concerns giving sensible names to things. There are 3 things that need to be added to the schematic 1) an overall title. Often (if not always) the division of a machine into schematic sections is governed by the physical division of the components amongst PCBs/modules/whatever, and it's often not sane to give an totally precise title to each board (So the HP 09810-66512 is not going to get titled 'CPU clock, microcode repeat and half the I/O logic'). But at least each board can be given a somewhat sensible name 2) Signals need to be given understnadable names (this is akin to using sensible variable names when programming). Some manufacturers diagrams are particularly poor on this and have signals 'A', 'B', 'C', etc. I prefer names like DDIR (Data Direction), MClk (Master Clock), and so on 3) Sections of the schematic should be titlted. Things like 'Microcode sequncer' 'Master Clock' 'Memory address decoder' and so on. Most commercial diagrams are very bad about this, but IMHO it really helps when findiny your way round a machine. -tony