TX-2 Project

Re-creating the historically important TX-2 computer

TX-2 Software

The project currently has some listings of TX-2 software, but as yet we have no known-correct source code and no known-correct binary.

The main problem is that we only have scans of code and it’s difficult to ensure that these are error-free. See Verifying Listings for an explanation of how we plan to do this.

With accurately transcribed authentic software, we can test the simulator and demonstrate that it is a credible re-creation of the TX-2 machine.

Listings we Have Already

We Need Machine-Readable Code

Known-good machine-readable code would act as a reality-check for our understanding of the TX-2. This would allow us to complete the simulator with come confidence about its correctness.

If we can be reasonably confident about the correctness of the simulator, then it may become tractable to try to reconstruct the illegible portions of the Sketchpad listing. This of course would be a huge task.

If we could obtain a machine-readable copy of the Sketchpad code, this would allow us to make that program run again for the first time in decades.

Having a machine-readable version of the Sketchpad code would by itself be an immensely valuable thing for the study of the history of computing.

Software that Once Existed

We have seen mention of various pieces of TX-2 software, though we don’t currently have a copy of any of these:

  1. Sketchpad by Ivan Sutherland; as noted above, there is a listing, but we would very much like to find a better-preserved one.
  2. Sketchpad-III by Timothy Edward Johnson
  3. The TX-2 BCPL compiler by Martin Richards, Henry Ancona et al.
  4. The LO Text formatter
  5. Bill Sutherland’s program for the on-line graphical specification of computer procedures
  6. The M4 assembler
  7. The APEX time-sharing executive