1 = Open Adventure Maintainer's Notes =
4 In which we explain what has been done to this code since Don Woods
5 authorized us to ship it under an open-source license. There's a
6 separate link:history.html[history] describing how it came to us.
10 The principal maintainers of this code are Eric S. Raymond and Jason
11 Ninneman. Eric received Don Woods's encouragement to update and ship
12 the game; Jason signed on early in the process to help. The assistance
13 of Peje Nilsson in restructuring some particularly grotty gotos is
14 gratefully acknowledged. Petr Voropaev contributed fuzz testing and
15 code cleanups. Aaron Traas did a lot of painstaking work to improve
20 This project is called "Open Adventure" because it's not at all clear
21 to number Adventure past 2.5 without misleading or causing
22 collisions. Various of the non-mainline versions have claimed to be
23 versions 3, 4, 5, 6, 7 and for all I know higher than that. It seems
24 best just to start a new numbering series while acknowledging the
27 We have reverted to "advent" for the binary to avoid a name collision
28 with the BSD Games version.
32 Extreme care has been taken not to make changes that would alter the
33 logic of the game as we received it from Don Woods, except to fix
34 glitches that were clearly bugs. By policy, all user-visible
35 changes to gameplay must be revertible with the -o (oldstyle) option.
37 It is a goal of this project to exactly preserve the *intended
38 behavior* of 430-point Adventure, but the implementation of it is fair
39 game for improvement. In particular, we are concerned to move it to a
40 form that is (a) readable, and (b) friendly to forward translation to
41 future languages. It has already survived a move from FORTRAN to C; a
42 future as a Python or Go translation seems possible, even probable.
44 == Functional changes ==
48 * Reading the relocated Witt's End sign in the endgame didn't work right.
50 * Behavior when saying the giant's magic words outside his room wasn't
51 quite correct - the game responded as though the player were in
52 the room ("...can't you read?"). The new message is "Well, that was
53 remarkably pointless."
55 * Attempting to extinguish an unlit urn caused it to lose its oil.
57 * "A crystal bridge now spans the fissure." (progressive present) was
58 incorrect most places it appeared and has been replaced by "A crystal
59 bridge spans the fissure." (timeless present).
61 * Under odd circumstances (dropping rug or vase outdoors) the game could
62 say "floor" when it should say "ground" (or "dirt", or something).
64 By default, advent issues "> " as a command prompt. This feature
65 became common in many variants after the original 350-point version,
66 but was never backported into Crowther & Woods's main line before now.
67 The "-o" (oldstyle) option reverts the behavior.
69 There is a set of standard one-letter command aliases conventional in modern
70 text adventure games; 'l' and 'x'; for 'look' (or 'examine'), 'z' to do nothing
71 for a turn, 'i' for 'inventory', 'g' for 'get', and 'd' for 'drop'. The 'd'
72 alias collides with 'd' for 'down', but the others have been implemented.
73 The "-o" (oldstyle) option disables them.
75 Unrecognized words are no longer truncated to 5 characters and
76 uppercased when they are echoed. The "-o" (oldstyle) option restores
79 A "seed" command has been added. This is not intended for human use
80 but as a way for game logs to set the PRNG (pseudorandom-number generator) so
81 that random events (dwarf & pirate appearances, the bird's magic word)
84 A -l command-line option has been added. When this is given (with a
85 file path argument) each command entered will be logged to the
86 specified file. Additionally, a generated "seed" command will be put
87 early in the file capturing the randomized start state of the PRNG
88 so that replays of the log will be reproducible.
90 Using "seed" and -l, the distribution now includes a regression-test
91 suite for the game. Any log captured with -l (and thus containing
92 a "seed" command) will replay reliably, including random events.
94 The adventure.text file is no longer required at runtime. Instead, an
95 adventure.yaml file is compiled at build time to a source module
96 containing C structures, which is then linked to the advent
97 binary. The YAML is drastically easier to read and edit than
98 the old ad-hoc format of adventure.txt.
100 The game-save format has changed. This was done to simplify the
101 FORTRAN-derived code that formerly implemented the save/restore
102 functions; without C's fread(3)/fwrite() and structs it was
103 necessarily pretty ugly by modern standards. Encryption and
104 checksumming have been discarded - it's pointless to try
105 tamper-proofing saves when everyone has the source code.
107 A -r command-line option has been added. When it is given (with a file
108 path argument) it is functionally equivalent to a RESTORE command.
112 The 2.5 code was a mechanical C translation of a FORTRAN original.
113 There were gotos everywhere and the code was, though functional,
114 ugly and quite unreadable.
116 Jason Ninneman and I have moved it to what is almost, but not quite,
117 idiomatic modern C. We refactored the right way, checking correctness
118 against a comprehensive test suite that we built first and verified
119 with coverage tools (there is effectively 100% code coverage). This is
120 what you are running when you do "make check".
122 The move to modern C entailed some structural changes. The most
123 important was the refactoring of over 350 gotos into if/loop/break
124 structures. We also abolished almost all shared globals; the main one
125 left is a struct holding the game's saveable/restorable state.
127 The original code was greatly complicated by a kind of bit-packing
128 that was performed because the FORTRAN it was written in had no string
129 type. Text from the adventure.text file was compiled into sequences
130 of sixbit code points in a restricted character set, packed 5 to a
131 32-bit word (it seems clear from the code that words were originally
132 *6* chars each packed into a PDP-10 36-bit word). A command noun or
133 verb was one of these words, and what would be string operations in a
134 more recent language were all done on sequences of these words.
136 We have removed all this bit-packing cruft in favor of proper C
137 strings. C strings may be a weak and leaky abstraction, but this is
138 one of the rare cases in which they are an obvious improvement over
139 what they're displacing...
141 We have also conducted extensive fuzz testing on the game using
142 afl (American Fuzzy Lop). We've found and fixed some crashers in
143 our new code (which occasionally uses malloc(3)), but none as yet
144 in Don's old code (which didn't).
146 The code falls short of being fully modern C in the following
149 * We have not attempted to translate the old code to pointer-based
150 idioms (as opposed, in particular, to integer-based array indexing).
151 We don't need whatever minor performance gains this might collect,
152 and the choice to refrain will make forward translation into future
155 * There are a few gotos left that resist restructuring; all are in the
156 principal command interpreter function implementing its state
159 * Linked lists (for objects at a location) are implemented using an array
160 of link indices. This is a surviving FORTRANism that is quite unlike
161 normal practice in C or any more modern language. We have not tried
162 to fix it because doing so would (a) be quite difficult, and (b)
163 compromise forward-portability to other languages.
165 * Much of the code still assumes one-origin array indexing. Thus,
166 arrays are a cell larger than they strictly need to be and cell 0 is
169 * The code is still mostly typeless, slinging around machine longs
170 like a FORTRAN or BCPL program. Some (incomplete) effort has been made
171 to introduce semantic types.
173 We have made exactly one minor architectural change. In addition to the
174 old code's per-object state-description messages, we now have a per-object
175 message series for state *changes*. This makes it possible to pull a fair
176 amount of text out of the arbitrary-messages list and associate those
177 mesages with the objects that conceptually own them.