Back in 2021 I announced the rediscovery of a text adventure named Castlequest (Mike Holtzman and Mark Kershenblatt, 1980). This month another “lost game” has been found: David Long’s 751-point Adventure!

Background

Around the time (in 2012) that I was developing and Kickstarting Colossal Cave: The Board Game, as a promotional extra associated with the game, I ported Donald Knuth’s advent.w from CWEB to ANSI C99, thus creating a version of Crowther and Woods’ Adventure that was playable both in the terminal and on the Z-machine.

Then I ported David Platt’s 550-point Adventure 3, Doug McDonald’s 551-point Adventure 6, and finally Peter Luckett and Jack Pike’s 440-point Adventure II. All of my C ports are playable online.

Adventure 6 was the most interesting of the bunch. The 551-pointer I ported is due to Doug McDonald circa 1984 at the University of Illinois Urbana–Champaign (hence its taxonomic name MCDO0551); but McDonald created it by adding only a few treasures and puzzles to a 501-point version (LONG0501) created by David Long circa 1978–1979 at the University of Chicago Graduate School of Business (GSB). LONG0501 had billed itself as Adventure 5, so it makes sense that when McDonald added his treasures, he bumped the version number to Adventure 6.

However, David Long himself had not finished modifying Adventure; he added a great many more rooms, puzzles, and treasures to the game and in January 1980 presented his own Adventure 6. It received major updates through March 1980 before stabilizing at a grand total of 751 points. It’s easy to find references to this LONG0751 version of Adventure; it was playable on CompuServe for years under the title New Adventure or Enhanced Adventure.

However, Long was secretive with the source code; Rich Alderson writes, “There was a time when Dave wouldn’t let more than 1 person at any particular institution have access to his version.” With the demise of CompuServe, it seemed this game had been lost forever.

But in fact it had not! …At least, not the executable.

Rediscovery

A lost-game enthusiast known as LanHawk contacted me in late 2024 to ask if I still needed help finding LONG0751. I said yes. About fifteen months went by, with occasional communications on various Adventure-related topics, and then… LanHawk sent some screenshots of a KL10 simulator running a program he’d found on a tape image from bitsavers.org.

Somewhere nearby is Colossal Cave, where others have found fortunes in treasure and gold, though it is rumored that some who enter are never seen again. Magic is said to work in the cave. I will be your eyes and hands. Direct me with natural English commands and I will tell you where you are and describe what is happening around you. My normal language is Fnordish, so if I don’t always understand what you tell me, be patient and try to say the same thing some other way.

[…]

Using 5 turns, you have reached a score of -5 points, out of a possible total of 751. Your lamp has 750 dwergs of power left.

It was an executable copy of LONG0751 in the flesh! Sadly, it was not source code; the source code remains lost. But the executable is real and playable!

How do I play it?

First, you’ll need Richard Cornwell’s pdp10-kl simulator. I built it from source (on a fairly modern MacBook):

brew install pcre
git clone --depth=1 https://github.com/rcornwell/sims
mkdir sims/build
cd sims/build
touch ../LICENSE.txt
cmake ..
make pdp10-kl
sudo ln -s `readlink -f ../BIN/pdp10-kl` /usr/local/bin

Then, you’ll need the files which LanHawk helpfully packaged up for me. Please note that I have very little idea what exactly this stuff is doing; it’s all magical incantations as far as I’m concerned. Buyer beware. But basically, as I understand it, LanHawk transferred the Adventure programs from the original tape image onto a disk image also containing the TOPS-20 operating system; the dsk/ directory contains this disk image, and the bts/ directory contains bootstrap code to boot it. Finally, pdp10-kl.ini is a script that tells the simulator to load those things into the right virtual places.

git clone --depth=1 https://github.com/Quuxplusone/Advent
cd Advent/LONG0751
unzip LONG0751.zip
cd LONG0751
pdp10-kl ./pdp10-kl.ini

Once the terminal settles down, hit Enter once, then type LOGIN OPERATOR TEST. (You don’t have to set caps-lock; all input on this system is automatically upper case.) This is a username/password pair, so, as you type TEST, you won’t see any echo to the screen. That’s normal. Hit enter at the end of that line. You should soon see a command prompt @. Type

R GAME:ADVENTURE

A little more detailed information (from LanHawk) is in the file running_Long0751.txt inside that zip file. Also note that there’s an executable of LONG0501 there as well: R GAME:ADV501 for that one.

Timeline

“Mid-1977”: Long receives the source code of WOOD0350 from Don Woods, according to one version of his “History of Adventure” text.

1978 to 1980: David Long expands WOOD0350 into “Adventure version 5” (LONG0501).

1980-01-19, according to the billboard: Robert Silverman is the first to set foot in version 6’s brand-new Castle of Aldor. (In 2017, Silverman wrote to me: “I was a friend of Dave Long at U. of C. from 1978 to 1980. […] I acted as a player/tester while Dave expanded the dungeon. That ended in spring of 1980 when I finished grad school and moved back East.”)

1980-03-07, according to the billboard: Long finishes coding version 6.4, although he still characterizes the game as “far from complete.”

Spring 1980: LONG0501 reaches Russel Dalenberg at the Illinois Institute of Technology. (See Dalenberg’s own timeline.)

1980-11-11: Dennis Donovan draws a fanciful, yet very accurate, map of LONG0751, which CompuServe sells as a poster for several years.

1982-01-14: Date printed by the unearthed LONG0751 executable (version 6.1/3).

1982 through 1996: Carl Ruby plays LONG0751 on CompuServe and is inspired to create his own “sweded” version of the game in BASIC. Without access to Long’s source code, his game is incomplete: it records only the parts he’s personally visited.

Spring 1995: Enhanced Adventure (LONG0751) remains playable on CompuServe at least this late.

2014-03-04 through 2014-03-25: I email Carl Ruby and receive a copy of his BASIC code (which see).

2016-01-06: I post my “In Search of LONG0751” webpage.

2016-02-08: Snatching at James Powell’s recollection that U of C Adventure was developed by “Dave Long and Michele Medanski and others,” I email Michele Madansky — who, it turns out, was merely “a precocious 12-year-old in 1978”; it was her father Al who was a professor at the GSB. He is very helpful in tracking down David Long’s current whereabouts (namely Portland, Oregon).

2016-09-03: I email what turns out to be David Long’s correct address. Four days later, Long responds! He claims to have “all the original Fortran-7 code, tho don’t know if it has been perfectly preserved”; but, while he’s reasonably forthcoming with anecdotal details (see below), no amount of gentle nagging ever succeeds in getting that source code out of him.

2017-01-29: I buy a copy of the Dennis Donovan map poster from a guy who emailed me after finding it at an estate sale.

2017-08-18: Long’s last email to me.

2024-09-05: LanHawk’s first email to me, asking if I still need help finding LONG0751.

2025-12-06: LanHawk informs me he’s found what looks like an executable for LONG0751 on a tape image at bitsavers.org.

2025-12-26: I confirm that the unearthed game indeed matches Carl Ruby’s BASIC version as far as it goes, and matches a walkthru distributed by PC-SIG in April 1990, and matches Dennis Donovan’s poster. I upload the executable to GitHub.

Anecdotes from David Long

Nine years on, I figure it’s fair to release these snippets to the world at large. This is pretty much all the historical information I ever got out of my email correspondence with David Long.

2016-09-03: […] the 751 game contains a small number of UofC in-jokes, of which my favorite is the “Blackened Shoals” location by the seaside, taken from a widely-taught finance theorem created by two then-UofC professors Black & Scholes. Can’t recall if that was original with me or if it came from one of the game’s fans with whom I occasionally met (and furnished w/beer) at the Eagle to spec out new caverns.

The billboard text here credits “Eric Weber for the wonderful play on words found at the seaside entrance.”

2016-09-11: Both Crowther & Woods did their work on PDP-6/10/Tenex/20 lineage of mainframes from Digital Equipment Corp, as did I, starting at Computer Corp. of America in Cambridge, Mass., in the early 70’s. (CCA was the 12th or 13th node on what is now the Internet.) When I went to Chicago in 1977, UofC had just purchased two DEC-20’s: one in the Graduate School of Business; the other for general use by the college. Since I tended to work 50–60 hours a week on GSB stuff, no one cared if I spent another 10–20 hours on Adventure. Turns out, Compuserve was entirely DEC-10/20 based, so all they needed was the executable, not the source. Easy-peasy. Earned me all of a thousand bucks […]

2016-11-23: I think I initially had a 500-point version that morphed into 501 points. The only entity to whom I can recall sending my code was CompuServe, and I think that was the later version. On the other hand, I might have handed out the executable more than once. It was an EXE for DEC-20 mainframes, which were once widely dispersed throughout academia. It would have been a straightforward exercise for anyone competent in dealing with binary files to decompose all the data structures, although the coding of all the action verbs would not have been easy. […]

I’ve never spelunked, although I’ve been in a number of caves around the country as a tourist. Like Carlsbad Caverns, Lava River Cave here in Oregon, Cave of the Winds in Colorado and others I can’t recall. Most of my additions were suggested either by fantasy stories or the many geology texts I’ve read.

A few more anecdotes from Charles Richmond, Rich Alderson, and James Powell are found in a thread from alt.folklore.computers (1999-09-23 through 1999-10-07).

Is this LONG0751’s final form?

It seems that LONG0751 went through a speedy expansion in early 1980 which ended (as Bob Silverman said above) with the graduation of the playtesters. That is, LONG0751 never got any bigger than what we’ve now unearthed: for example, the castle staircase was never completed. Still, I see at least two tantalizing hints that Long may have had, and/or completed, bigger plans for his cave. Both hints are found in the RARE BOOK, which upon reading turns out to contain a “History of Adventure.”

(1) In OSKA0501, which remember is Long’s version 5, the History of Adventure includes a “teaser” for version 6:

[...] To whet curious appetites,
major version 6 is projected to appear sometime around the summer
of 1979.  It will contain a whole new wing to the cave (natch!)
plus a much enlarged surface area, including the Great Serbonian
Bog, the Castle of Aldor and the terrifying Passage of Fire.
Stay tuned!

The unearthed LONG0751 has the Castle of Aldor; and it has a new swamp, which is close enough to a “Great Serbonian Bog”; but it has nothing remotely identifiable as a “terrifying Passage of Fire.” Did Long abandon this planned addition to the cave?

(2) On 1996-05-02, Ken Hargreaves posted in comp.sys.ibm.pc.games.rpg a version of the History of Adventure different from the one in the unearthed game. The exact diff of the middle portion (modulo whitespace) is:

The latest additions were done throughout 1978-80 by David Long
at the University of Chicago, Graduate School of Business.
-Long's additions include the seaside entrance and all of
+Additions include the seaside entrance and all of
the cave on the "far side" of Lost River (Rainbow Rm - Crystal
-Palace - Blue Grotto - Rotunda - beyond Joshua's wall, etc., etc.).
+Palace - Blue Grotto - Rotunda - Joshua's wall, the Gothic Cathedral, etc.).
The surface has also been greatly increased to include a much
more varied landscape containing swamp, marsh, seashore and
meadowland areas.  Most recent additions include the great Castle of
-Aldor, the Elephants' Burial Ground, Leprechaun Rock and more.
+Aldor, the Elephants' Burial Ground, Sham Rock, the Lost Silver
+Mine, the helicopter, the Secret Garden, Bigrat and more.
-The current cave is more than double the size of the Woods model,
+The current cave is nearly triple the size of the Woods model,
and moreover the puzzles and treasures are somewhat more "dense",
(and more difficult!) in the current version.  During the expansion
process, the code was almost entirely rewritten to permit more
generalized handling of objects and to interpret a more complex
natural English syntax.

This other History is slightly more boastful: “more than double” becomes “nearly triple.” But does this indicate any actual expansion of the cave? The unearthed game already includes all of the features mentioned here. Was Long just massaging his messaging?

Reward for lost games’ source code

If you happen to have the source code for any flavor of LONG0751, I’m quite willing to pay a bounty of, say, $1000 for it. Send me an email!

And while I’m at it, I’ll offer the same bounty for source code to the CompuServe/GEnie game Blackdragon (see here) and to the Colorado State University expansion of Adventure (see here).

Both of these games remain completely lost, as far as I know — not even an executable of either of them is extant. So I’ll also happily offer $500 for a runnable executable of either of them, too.

Other lost games include Dor Sageth, The PITS, and anything on this list. If you have any information on any of these, please email me! And Jason Dyer is looking for even more lost games: about 24 titles here and another 15 here.