Version	Date	Description
1A	1978	Original Macro-10 shared memory test.
2A	1978	Macro-10.
3A	1979	BCPL (Essex MUD, CompuNet MUD).
3B	1986	BCPL/Fortran 4 for CompuServe (BL).
3C	1987	Fortran for PR1MEs.
3D	1999	As 4E.
4A	1985	Apricot MS-DOS in Turbo-Pascal (development only).
4B	1985	VMS Pascal for VAXen.
4C	1988	Pascal under OS9 for the MUDbox (development only).
4D	1989	C for Archimedes (development only).
4E	1991	C for Unix (originally SVR4, but now many derivatives).

Version 1A was called MUD, although the main comment in its code said "MUDD - MULTI-USER GAME OF ADVENTUROUS ENDEAVOUR", a typographical error. It was conceived and written by Roy Trubshaw, with helpful programming suggestions by Keith Rautenbach and others, and encouragement/praise from Nigel Roberts and Richard Bartle. Version 1A was not programmed as a playable game: it was a test to see whether the shared memory system Trubshaw envisaged would work (it did). The date on the earliest surviving listing is 1979, but actually it was first run in 1978.

Version 2A was called MUD, but expanded the acronym into 'Multiple User Dungeon'; however, the more informal 'Multi-' was always used, and this became the standard in version 3A. Version 2A was coded entirely by Trubshaw, and its database was run-time programmable by privved players; Roberts and Bartle again chipped in with suggestions, and created rooms/objects in the system (the 'ox', still in the database today, got in that way). Programmability turned out to be something of a Bad Thing: it used too much memory/disc space, and people kept adding things which were not in character (sense 5) with the rest of the game. The database-design module was therefore removed in later versions, and the task of permanent game design done separately.

The heart of version 3A was Trubshaw's, amounting to perhaps 25% of the code (the hardest 25%!); the rest was added incrementally by Bartle over the next 3 years. Brian Mallett and Ronan Flood each provided useful hacks to the low-level code. Although the program was begun in 1979, it wasn't in a playable state until Easter 1980, and therefore this later date is more often quoted concerning the program's beginnings. The alterations to version 3A which made it 3B were Bartle's. 3C was a recoded 3A by Mike Arnautov for the PR1ME system at Glaxo.

Version 4A was Bartle's work, and the interpreter made it to 4B, with Trubshaw coding the FE and the inter-process communication. Bartle converted the lot for 4C, and Jon Thackray did the transliteration into C for 4D, working from 4A and 4B; Thackray's database compiler is still used, with minor additions by Bartle. Bartle did the transliteration to C for 4E from 4C, with help (ha!) from a program called 'ptc'. 3D was a stripped-down 4E, with modifications to make it look and feel like 3B, by Chip Hayes.

See also SUD.