My project is a code perfect implementation of an IBM 1130 that I first programmed in 1970. It runs the entire OS including Fortran and the Macro Assembler. The original was a very cool computer - the first a person could see over.
I started the project on a Spartan 3 Starter Board and have moved it to a Nexys2. Next I will move it to the Artix or Basys 3 if there are enough cells and it looks like there are.
In addition to the CPU and Compact Flash disk drive, I also have serial ports for the Console Typewriter, Console Keyboard, Printer and Card Reader. The Plotter is emulated in an mbed with the optimized output fed over TCP to my LaserJet. It even has the IBM Electronic Circuit Analysis Package (ECAP) that I used in my undergrad studies.
The real machine had only 8k words and mine has the max of 32k words. The original ran at about 400 kHz with lots of steps per instruction (not to mention an adder that didn't implement carry logic but instead repeatedly applied carry operations until none remained) and mine runs at 50 MHz.
I have puttered with this thing for about 10 years. Nothing has changed (or improved), just migration from board to board.
I have the original 2 volume Algol 60 books that include the source listings. I had to write a tiny Algol compiler in grad school ('76) and the only machine I could get at ran time-shared BASIC. It's hard to write clean code in BASIC! It would be kind of fun to build a CPU that was more or less optimized for Algol or even Pascal. I wrote most of an FPGA core to implement Wirth's CDC 6400 Pascal but I got hung up when it came to system calls. I now know how to do that.
For many years, I have wanted to build BLUE, an elementary 16 bit computer in Caxton C Foster's "Computer Architecture". The CPU has nothing to recommend it except that it is conceptually simple. Al Williams made a hyped up version that is over at Open Cores
http://www.drdobbs.com/embedded-systems/the-spartan-blue-cpu-in-verilog/228700593That Basys3 board will be perfect so that's the next project.