THE HISTORICAL BBS LIST 108,078 listings, about 103,110 boards, and the world before the web ------------------------------------------------------------------ Before there were websites there were boards. You dialled a telephone number with a modem, one caller at a time, and reached a computer in somebody's spare room. This is the directory of them - every dial-up bulletin board system its compilers could account for, across 268 North American area codes. There is one thing in here that no web archive can help with: these have telephone numbers, not addresses. Nothing in this list can be looked up in the Wayback Machine, because none of it was ever on the web at all. It is the generation immediately before. Listings 108,078 Distinct boards (see below) ~103,110 Area codes 268 With a named sysop 80,433 (74.4%) With years of operation 105,619 (97.7%) With the BBS software named 28,096 (26.0%) Carrying notes or magazine excerpts 4,257 Earliest and latest listed 1978-2020 The percentages are shares of listings, not of boards. A duplicated listing duplicates its sysop and its note along with it, so dividing those counts by 103,110 would invent a number rather than correct one. A CORRECTION: 108,078 IS NOT THE NUMBER OF BOARDS ------------------------------------------------------------------ I have been quoting 108,078 boards. That is right as a count of listings and wrong as a count of boards. All 268 area code pages state their own totals, and all 268 agree with what was scraped. That check holds. But in the 1990s the telephone companies split area codes. Detroit was 313 until 810 was carved out of it. Toronto was 416 until 905 appeared. The same machine on the same line kept ringing with a new number in front of it, and the BBS lists recorded both. Isle Net in New Jersey is listed at 201-495-6996, 908-495-6996 and 732-495-6996. Three entries. One modem. Measured: 4,947 telephone numbers appear more than once under area codes that really split, accounting for 5,041 duplicate records. That leaves roughly 103,110 distinct boards. My first attempt was worse. I wrote the list of splits from memory, got 53, and came out low by about half. The corpus can be asked instead - a real split leaves hundreds of identical seven-digit numbers whose board names match, and a coincidence leaves a handful with names that do not. Asked that way it returns 94 pairs. One of the 94 is not a split. 313 and 314 share 209 numbers and the names match on 177 of them. 313 is Detroit, 314 is St. Louis. That is an error in the 2001 source list, not in the scrape, and it files 245 Detroit boards under Missouri. I have counted them separately and left the records alone, because silently correcting somebody else's list is not my business. 143 pairs carry different names on the same line. Renamed board, or number reassigned to a stranger? I read them: about seventy are the same board, about seventy-three are not. So the figure sits between 103,037 and 103,180. A NOTE ON THE SPELLINGS ------------------------------------------------------------------ Some entries show a corrected name with the original beneath it, marked "listed as". That is deliberate. This list was assembled from many separate 1990s BBS lists, and each one carried its own typing errors - Inveriont Factory for Invention Factory, Manhatten Transfer for Manhattan. Those mistakes are evidence of which source an entry came from, so they are never overwritten. 995 names carry a correction, and every one of them still shows what it actually said. Corrections are made a single word at a time. Invention Factory Node #3 is not silently turned into Invention Factory - that would lose which node it was, which is information, not a mistake. And a board numbered #1 through #4 is four boards, not four misspellings of one: a typo does not politely increment. WHICH OF THEM BECAME WEBSITES ------------------------------------------------------------------ That is the question this collection was built to ask, and it has two answers that disagree with each other. Matching board names against a catalogue of 30,164 early websites proves seven. Reading what the sysops themselves wrote about their boards finds ninety-five more. The gap between those two numbers turned out to be the whole finding: a join between two collections can only surface what both of them already contain, and 96% of the addresses the sysops named were in no catalogue at all. That ninety-five was ninety-three until recently. I had said there were three audit files nobody had read. Months later I read one. It contained two crossings that the rules had cancelled out on - Abingdon Online, whose sysop wrote "Moved to the web in 1996 and eventually disconnected our phone lines", and Adult Fantasy BBS, whose home page was printed in Boardwatch in January 1996 in the same sentence as its telnet address. Neither was hard to see. They were in a file I had already told you I had not opened. AND A CORRECTION ABOUT THE ONES THAT SURVIVED ------------------------------------------------------------------ This page used to say that forty-four boards never became websites and are still answering a telnet port thirty years later, and it named four of them. I dialled them. On 12 August 2026 I opened one connection to every address of that kind in the list - no login, no keystrokes, read whatever the far end sent unprompted, and hang up. Real addresses settled 173 Answered 9 boards Connection refused 20 Hostname no longer resolves 142 No answer either way (firewalled) ~105 Nine, not forty-four. And of the four I named, Sursum Corda and Bandit are gone - their hostnames do not resolve at all. The Basement Underworld is still there, answering on port 9000. The ones that answered: The Keep, running since 1983 and still greeting callers with "Auto-sensing..."; The 13th Floor, on a Commodore 64 in Nashville, answering on two addresses at once; Part-Time; Slumberland; Synerchat; Christian Fellowship, which moved hosts and survived when its old address died; The Basement Underworld; and two more. The hundred and five that gave no answer either way are not deaths and I have not counted them as such. Modern firewalls drop packets silently rather than refusing them, so a machine behind one looks exactly like a machine that is gone. I know this because SDF - a public-access UNIX system that began answering in 1990 and is certainly running as you read this - timed out when I dialled it. If I had counted timeouts as deaths I would have buried it. What I got wrong here was not the arithmetic. Forty-four was what sysops wrote between 2001 and 2008, and I published it as though it described today. A sentence in the present tense has a date on it whether you write one or not. WHERE THIS CAME FROM ------------------------------------------------------------------ Every entry belongs to the TEXTFILES.COM Historical BBS List, compiled since 2001 by Jason Scott and a great many contributors, from sources listed on that site: http://bbslist.textfiles.com/ What is added here is the search, the cleaning, and the cross-reference. The searchable directory - by board name, sysop, town, area code, software or year - is on the web version, where every row links back to the area-code page it came from: https://ge97.com/bbslist.html The notes, memories and magazine excerpts are reproduced as published there, and belong to the people who wrote them. If you want to argue with any of these numbers, I would rather have the argument than the number. The numbers are long dead. Please do not ring them. Khalid Alshaikh, 7Z1FP https://ge97.com/ .