Talk:KD11-E/EA microcode

Accuracy

The list of uinsts was prepared by going through the engineering drawings (I have a hardcopy set of KD11-EA drawings which include the full ucode flow diagrams), and creating an ordered list of uinsts. After this was done, I made another pass, checking the newly created table against the drawings. Finally, I made a third pass, comparing the table against the ucode flow diagrams in the KD11-E drawings, which are, with one exception identical to the KD11-EA drawings (at least, location and label; I didn't compare the actual contents of each uinst). So, the table is probably correct, with those checks.

Early on, in the first pass, a couple of 'crashes' appeared; different uinsts assigned to the same location. When that pass was done, it was relatively easy to work out, for two of the three, what the (typographical) error was. The correctness of the fixes was vouched for by the fact that the corrected location was a gap in the otherwise full list of uinsts. The error in the third wasn't so obvious (it was a long way from its correct location), but use of the dump on BitSavers allowed it to be fixed too. Jnc (talk) 04:25, 13 January 2019 (CET)

For the 'next uinstruction' column, the data was entered from the prints, but not re-checked against them; instead, it was checked against the dump from BitSavers (which was a lot less work, since both lists are in uinst order). Jnc (talk) 02:51, 15 January 2019 (CET)