Skip to content

TODO.md

Latest commit

 

History

History
748 lines (687 loc) · 44.7 KB

TODO.md

File metadata and controls

748 lines (687 loc) · 44.7 KB
Raw

Open Research Threads

Follow-ups from README.md, roughly in order of expected value. None of these were run down during the initial research pass.

  1. Confirm the Kincaid identification. Section II of the README lays out a strong-but-unproven case that the Patrick Kincaid of 618 Douglas Drive, Mill Valley, CA 94941 (per TODAY.DOC, 1986) is the same Patrick James Kincaid whose obituary is cited there (1947–2020). A Marin County property/voter record or an old Marin-Sonoma PC User Group roster from the mid-1980s could close this gap directly.

  2. Butler's identity is now anchored, but not complete. The arcbbs- hosted TODAY_DOC only credited "Mike Butler." A June 5, 2019 Facebook post from UTMBiology (University of Toronto Mississauga, known as Erindale College in 1981) independently confirmed his full name — Michael Butler — the institution, and the exact date the mainframe original first ran (June 5, 1981, matching the today.6 self-reference S06051981 TODAY ran for 1st time). See README Section I for the full quote and the archived screenshot at screenshots/utmbiology-facebook-2019-06-05.png.

    Update: this is now independently corroborated a second time, and from Kincaid himself. The full TODAY.DOC manual recovered inside TODAY34.ZIP (see README Section I and zips/) has a dedicated "TODAY's History" section, in Kincaid's own words, also naming "Michael Butler" in full — not just "Mike." The Facebook post and Kincaid's own documentation are independent of each other and agree exactly, which meaningfully strengthens this identification.

    Still open: whether Butler was a student, faculty, or staff member at Erindale/UTM in 1981; whether any of the original PL/1 source survives; and how Butler and Kincaid actually connected five years later (personal acquaintance, or did Kincaid just receive a copy secondhand through the PC user group / BBS scene?).

    Already checked, came up empty — don't repeat these:

    • General web search for "Michael Butler" + University of Toronto / Erindale / computing turned up nothing. Several Toronto-area "Michael Butler" obituaries exist (notably one for 1948–2019) but none mention computing or U of T — treat those as false leads, not candidates. Re-ran the same University of Toronto/Erindale/1981/ PL-1/VM-CMS combinations with "Mike Butler" instead of "Michael Butler" specifically (since TODAY.DOC itself only ever says "Mike") — also nothing. Both name forms are exhausted against this specific institutional context; no need to retry either.
    • GitHub code search for "Michael Butler" / "Mike Butler": nothing.
    • UTM's own history pages (the CS department's public timeline, the "It was 50 years ago today" retrospective article): no mention.
    • The Medium (Erindale's student newspaper, digitized on Internet Archive as "Medium II," identifiers mediumiiVVnNNerin): pulled and full-text-searched all 32 available issues from January 1981 through March 1982 (the paper has a summer publishing gap, so June 5, 1981 itself isn't covered) for "Butler," "TODAY," "computing centre," and "VM/CMS." No genuine hits — two "Butler" matches were unrelated people (a comptroller in a job ad, an actor in a theatre review).
    • A University of Toronto "pioneers of Canadian computing" retrospective PDF (individual.utoronto.ca/klyons/files/pioneers.pdf): no mention of Butler or Erindale in a relevant context.

    In progress:

    • Outreach sent to the UTMBiology Facebook page asking how they sourced the 2019 "on this day" post and for more on Michael Butler — awaiting a reply. Whoever wrote that post had some source, which may lead straight to Butler or his original code.

    Not yet tried:

    • Contact UTM Archives & Records Management Services (UTARMS) directly — they hold the official institutional archive, including things a student newspaper wouldn't cover, like computing centre newsletters, course calendars, or IT staff records from 1980–82.
    • Check for a UTM/Erindale "Computing Society" or similar student club, which would be a more likely venue than the general newspaper for coverage of a hobby mainframe program.

    In progress (July 2026): Leo Bicknell has sent a LinkedIn contact request/message to a Michael Butler (linkedin.com/in/strider1) — based in Mississauga, ON, BA in Economics and Political Science from the University of Toronto, 1980–1984. Matches the right city (Mississauga/Erindale) and era; class years start just after the 1981 mainframe debut, so if this is the right person he'd have been a first- or second-year student at the time. Sent, no reply yet; not yet confirmed as the same Michael Butler.

  3. Find the original Usenet posting — but note the evidence for when is weaker than it looks. The working assumption that this was posted to comp.sources.misc or alt.sources rests mainly on this directory's own origin story (saved from Usenet), not on anything in TODAY.DOC, which describes Kincaid's actual distribution channel as the Marin-Sonoma PC User Group BBS and mail-in submissions — a regional PC-shareware/BBS model, not the Unix-world sources-group convention. comp.sources.misc didn't exist before 1987, so a posting there can't predate that, but it could plausibly have happened any time from 1987 through 1990 — the only two fixed points we have for the data circulating beyond Kincaid's own hands (Hugo Fiennes having a copy by May 1990; the GNU gcal conversion of "ProLine's today file" dated March 1991) are both consistent with pure BBS/UUCP/ProLine-network propagation and don't require a Usenet posting to have happened at all.

    Update: three later Usenet/Fidonet posts have since turned up showing the DOS executable itself in use (rec.org.sca, Aug 1995; comp.os.msdos, Sept 2001; fido7.moscow-oklahom, Jan 2004 — see README Section I for the version banners each one shows). None of these is the original release posting — they're years-later reposts of program output by end users, not source distribution — but they do confirm real circulation on Usenet/Fidonet from at least 1995 onward, and independently established a version history (1986 → 1988 → 1990 v3.1 → 1993 v3.3 → Nov. 1993 v3.6) that didn't exist in this document before. Hugo Fiennes's direct account (item 9, below) now makes BBS-to-BBS spread — rather than a Usenet source posting — the better-supported propagation story overall; the original-posting question may be somewhat moot.

    The tuhs.org and nycbug archives only cover January 1988–June 1991 — that's just what happens to be digitized there, not a range justified by evidence. It covers the plausible late end of the window reasonably well but has no claim on 1986–1987. If a posting exists and predates 1988, these two archives won't find it. Worth trying Google Groups' own Usenet archive search (broader/different coverage, sourced from Deja News, may reach further back) with more specific terms than were tried earlier in this research — now that we have "TODAY.DOC," "Datalight C," "Mill Valley," and "Marin-Sonoma PC User Group" as more distinctive search anchors than a generic "Patrick Kincaid" + "today" query.

    Update (July 2026) — tried, came up empty; the case for "moot" is now stronger. Prompted by the discovery that PKZIP didn't exist until 1989 (which led directly to TODAY10.ARC/TODAY21.ZIP, item 11 above), the same logic was applied one step further back: .arc only dates to 1985 itself, so an original Unix-side posting from the plausible 1986–1987 window would more likely have used a Unix-native format — .shar (shell archive, the standard comp.sources.misc/ alt.sources posting format), .tar/.tar.Z, or the old ar archiver. A search of discmaster.textfiles.com for TODAY.SHAR, TODAY10.SHAR, TODAY11.SHAR, TODAY20.SHAR, TODAY21.SHAR, TODAY.AR, TODAY10.AR, TODAY11.AR, TODAY.TAR, TODAY10.TAR, TODAY11.TAR, and TODAY.TAR.Z returned zero hits for every single one — a sharp contrast with the immediate, multiple hits TODAY10.ARC/TODAY21.ZIP produced. An archive.org advanced-search API query for today.shar likewise returned zero results, and "patrick kincaid" returned only the two DOS releases already in hand plus unrelated court records. The Google Groups searches suggested above (using "Datalight C," "Mill Valley," "Marin-Sonoma PC User Group," TODAY.DOC/TODAY.JAN as anchors, alone and combined with shar/tar) were also run and returned nothing relevant. Finally, the tuhs.org comp.sources.misc full index for Volumes 1 through 8 (covering 1987–1989, the earliest part of that archive's coverage and the most plausible window for an original posting) was checked directly for any entry containing the word "today" anywhere in its archive-name or subject fields — no match. None of this proves a Usenet source posting never existed (Deja News/Google Groups' indexing of this era is known to be incomplete, and a .shar predating 1987 wouldn't be in comp.sources.misc at all, since that group didn't exist yet), but between this and Hugo Fiennes's account (item 9), pure BBS-to-BBS propagation is now the much better-supported story — this thread is likely not worth further search time barring a new, more specific lead.

  4. Investigate "ProLine." Largely resolved. The GNU gcal data files state their trivia was "extracted from ProLine's today file." This referred to Morgan Davis's ProLine BBS software (an Apple II Unix-like system with UUCP connectivity, popular in the late 1980s/early 1990s).

    Update (July 2026): Leo Bicknell found that Morgan Davis has posted ProLine's disk images, a 335-page manual, and its full source code at morgandavis.net/post/proline. The source archive (proline-3.0-source.zip) includes a man page, man/man/man.G/today.G, for a built-in today command — and it names names directly:

    The ProLine version was written by Jeff Jungblut (jeff@pro-avalon.cts.com). The original database is from a similar public domain program for the IBM PC called TODAY/PC.

    This is a direct, independent confirmation of Kincaid's exact program name (matching the "TODAY/PC" version banners already found on the recovered DOS executables — item 11) and identifies a sixth independent implementer, previously unknown: Jeff Jungblut, who ported the today display command (not necessarily Kincaid's own code — most likely a fresh implementation reading an imported copy of his data files, in the same pattern as Fiennes/Conley/Howland). jeff@pro-avalon.cts.com is almost certainly defunct (cts.com was CTS Network Services, a San Diego ISP Davis himself co-founded — so "pro-avalon" was a ProLine BBS node hosted on his own network), but the name is a new, searchable lead. The man page itself doesn't include the actual today source code or data files (the ProLine source archive only bundles system/admin code, not per-installation game data — the today databases would have lived in $/games/lib/today/* on a running system and aren't preserved here).

    In progress (July 2026): Leo Bicknell found a likely candidate — a Jeff Jungblut (linkedin.com/in/jeff-jungblut-7170572a1), based in San Diego, CA — a strong match, since cts.com (the domain in jeff@pro-avalon.cts.com) was CTS Network Services, the San Diego ISP Morgan Davis himself co-founded. The profile has very little detail, but it's the only LinkedIn hit at all for this unusual name/spelling combination. LinkedIn contact request/message sent (July 2026), no reply yet; not yet confirmed.

    Update (July 2026): Morgan Davis replied — no new information, but confirms Jungblut as the sole source. Responding to Leo Bicknell's comment on his blog post (morgandavis.net/post/proline), Davis wrote:

    I'm afraid I cannot offer much help here. Jeff Jungblut provided the code and any database that may have been included with ProLine. Hunting down Jeff would be your next step in getting some answers. Good luck on your search.

    For reference, the documentation says: The ProLine version was written by Jeff Jungblut (jeff@pro-avalon.cts.com). The original database is from a similar public domain program for the IBM PC called TODAY/PC.

    Davis himself never had the data files independently — Jungblut supplied everything — so this thread now dead-ends at Davis and points entirely back to confirming the Jeff Jungblut LinkedIn lead, above (sent, no reply yet).

    Resolved (July 2026): Jungblut replied, largely closing this item. Davis forwarded Bicknell's question directly to Jungblut, who replied to both of them, confirmed he is the same Jeff Jungblut found on LinkedIn (explaining the earlier silence there: he no longer uses the site), and gave permission to add his source code, data files, and message excerpts to this repository. Full account, correspondence, and findings now in README.md Section VI and zips/README.md (JUNGBLUT-TODAY.ZIP/JUNGBLUT-TODAY.BAS). Highlights:

    • His code is an independent implementation (Applesoft BASIC, ProLine's own system-call extensions), not a port of Kincaid's C — the same "wrote my own reader" pattern as Fiennes (item 9) and Howland (item 17).
    • His own Feb-1989 data files are a 99.1% verbatim match against this repo's own today.1–.12 (already known to be byte-identical to Kincaid's 1986 v1.0 release), with June matching 100% — a third independent lineage confirming the same original 1986 database, including the self-referential S06051981 TODAY ran for 1st time entry later dropped from the DOS v3.6 release.
    • He does not know, and cannot shed any light on, Michael Butler, Michael Conley, Bruce Goldman, Robert Heckendorn, or Bill Randle — those threads (items 1, 5, 14, 15) remain exactly as open as before.
    • He does not recall where he originally obtained Kincaid's data, and no source earlier than his surviving v1.2 (5/20/92) copy exists — so the specific BBS-to-BBS link between Kincaid and Jungblut, circa 1988/89, is still unknown.

  5. Contact the gcal data's editors. The events1/lives3 headers name Robert Heckendorn (robert@fc.hp.com, contributed "Today in History" dates) and Larry Virden / Bill Randle (billr@saab.cna.tek.com), who converted "ProLine's today file" into CalenTool format in 1991. Any of the three may have first-hand knowledge of where the ProLine file came from, or retain older copies with more attribution than what survived into gcal.

    In progress (July 2026): Leo Bicknell has sent LinkedIn contact requests/messages to a Robert Heckendorn (linkedin.com/in/robertheckendorn) — software engineer at HP, Fort Collins, CO, 1979–1995, matching the robert@fc.hp.com ("fc" = Fort Collins) address exactly — and a Bill Randle (linkedin.com/in/billrandle) — at Tektronix, Beaverton, OR, 1974–2001, matching the billr@saab.cna.tek.com ("tek" = Tektronix) address. Both are strong company/location/era matches. Sent, no reply yet from either; not yet confirmed.

  6. Preserve the Wayback Machine capture — done. The last known-good archive of burgerbecky.com is from January 16, 2026 (link). The whole site (4,077 URLs, last good pre-hijack capture of each) has since been mirrored and preserved at bicknell/burgerbecky.com, so it no longer depends on Wayback access remaining available. Checked directly in that mirror: Heineman's Doxygen-generated burgerlib documentation has no additional history notes about today beyond the source comment already quoted in README Section I.

  7. KitchenSink itself has no other contributors to ask — checked via the GitHub API. It's a single commit, authored solely by Rebecca Heineman (burgerlib is the same story: 273 commits, all hers). The Olde-Skuul org has no publicly-listed members either, so there's no roster to pull directly.

    One narrower lead turned up and then mostly fizzled: her doom3do repo has 10 commits from Heineman and 1 from a contributor called Soviut — identified as Ian Zamojc (GitHub zamtools, email izamojc@zamtools.com, resume at resume.zamtools.com, LinkedIn at linkedin.com/in/ianzamojc). His actual commit, though, was a one-line README typo fix from December 16, 2014 — not a collaboration. His resume shows a 17+ year career as a web/frontend engineer (Vue, Node, design-focused roles at Arternal, Synervoz, Metrist, NPM) with no game-industry background and no mention of Heineman, Olde-Skuul, or retro computing anywhere. Low priority — he's very unlikely to have any insider knowledge, though his contact info is on record here in case it's ever worth a long-shot ask. (Two other org repos, Catacomb and keen, have solo commits from unrelated third parties — Richard Mandel and Javier M. Chavez respectively — but nothing ties them to Heineman or this research thread either; not worth pursuing.)

  8. Track down a copy of TODAY.OWN — done. Found intact inside TODAY34.ZIP (see README Section I and zips/TODAY34.ZIP): a short, unremarkable set of sample reminders (mortgage payment, fictional family birthdays) — no personalization or attribution clues about Kincaid himself. The identical 380-byte file size in the later TODAY36.ZIP listing suggests it never changed across releases.

  9. Contact Hugo Fiennes — done; he replied. See README Section V for his full profile and his verbatim reply (July 2026). Summary of what was resolved and what wasn't:

    • His ARCbbs "on this day" module was his own independent C implementation (c/today in his repo), not a port of Kincaid's code — it just reads the same data-file format.
    • He believes he got the data files from another BBS sysop, or downloaded them from another BBS — he does not know which one, or whether it traces back to Kincaid directly or through intermediate hands.
    • He has no contact with anyone named in TODAY.DOC (neither Kincaid nor Butler) and describes himself as "very much a leaf node in this whole thing."
    • He did not confirm or deny Usenet access in the 1980s; the question wasn't directly answered, so item 3's open question about a possible Usenet posting is still unresolved on that front (though his account generally favors BBS-to-BBS spread over a Usenet source-posting).
    • He's agreed to be invited to the repo.

    New lead from his reply: he suggested contacting the operator of bbsdocumentary.com (Jason Scott, who compiled the BBS Documentary project) as someone who might know more about this program's place in BBS history generally. Not yet contacted.

  10. Consider flagging the domain hijack. burgerbecky.com currently redirects to a gambling site (see README Section IV) with no obvious sign that anyone in a position to fix it is aware. If there's a natural point of contact — GoDaddy abuse reporting, a member of the retro-computing/games-preservation community, or someone connected to her estate or the Olde-Skuul org — it may be worth a heads-up, independent of the Kincaid research. Update: the site itself is no longer at risk of being lost regardless — see bicknell/burgerbecky.com (item 6, above) — but the hijacked domain itself is still worth flagging if a good point of contact turns up.

  11. Track down the actual DOS releases, and place this repo's copy in the resulting timeline. Resolved, far more completely than originally scoped, in two waves. First wave: real, verified copies of v3.1 (TODAY31.ZIP, from archive.org), v3.3 and v3.5 (TODAY33.ZIP/TODAY35.ZIP — both previously undocumented versions, found via discmaster.textfiles.com, a ~1.6-billion-file search index across dumped shareware CD-ROMs/BBS/ FTP archives), v3.4 (TODAY34.ZIP, from bbs.retropc.se), and v3.6 (TODAY36.ZIP, from archive.org).

    Second wave (July 2026): the two earliest releases, TODAY10.ARC (v1.0) and TODAY21.ZIP (v2.1), have now also been recovered, via a deliberate search of discmaster.textfiles.com for TODAY.ZIP, TODAY10.ZIP, TODAY11.ZIP, TODAY20.ZIP, TODAY21.ZIP, and their .ARC/.ARK equivalents (PKZIP didn't exist before 1989, so any release predating that had to circulate as .ARC — this is exactly what turned up). All seven releases are preserved in zips/ — see README Section I and zips/README.md for everything they revealed collectively: the full "TODAY's History" section and Michael Butler's full name direct from Kincaid, present as early as v2.1 (1988) and continuing through v3.4; documentation consistently lagging the version number by one release; registration terms; the Golden Gate Computer Society/Corte Madera/PCBoard details; the full data-format spec; TODAY.OWN's actual contents; byte-exact confirmation of the Usenet version banner (and an even earlier one, TODAY/PC V1.0 (c) 1986, from the v1.0 executable itself); the ALTFN/MUSIC/SPACE special-database feature's actual introduction in v2.0 (March 1988); and confirmation the lunar-phase feature is absent from every recovered release's code and docs alike, including the earliest one.

    This resolves the central open question: line-level comparison against this repo's own month files, across all seven versions, shows this repo's today.1–.12 are byte-for-byte identical, file for file, to TODAY10.ARC's (v1.0, 1986/87) month files — not merely the closest match, as the earlier v3.1-vs-v3.3 comparison had suggested, but an exact, unmodified copy of Kincaid's original 1986 databases. See README Section I for the full comparison table and per-file verification.

    The same discmaster.textfiles.com searches also confirmed TODAY32.ZIP (an unrelated HP48 calculator app) and TODAY11.ZIP ("Today's Callers v1.1 for PC-Express," an unrelated BBS caller-log utility) are false positives, found no evidence at all of a standalone TODAY20.ZIP/ TODAY20.ARC/TODAY2.ARC/TODAY21.ARC/TODAY.ARK, turned up 9 further copies of TODAY36.ZIP, 1 more of TODAY31.ZIP, and 29/26 copies respectively of Michael Conley's TODAYDOR.ZIP/TODAYBBS.ZIP (see README Section I), none individually downloaded — and turned up a fifth independent implementer, Mick Howland's CP/M port (TODAY-CPM.ARC, 1989, see item 17 below).

    Update (July 2026): the v2.1 "Reducing"-compression blocker is resolved. Neither Info-ZIP unzip, p7zip, nor the current official 7-Zip 26.02 (7zz) can decode PKZIP's old "Reducing" method (methods 1–4) — confirmed by testing all three directly. Hans Wennborg's hwzip, a public-domain reference implementation written specifically to document the Shrink/Reduce/Implode legacy methods (source at hanshq.net/files/hwzip/hwzip-2.4.zip, builds clean with a plain clang invocation per its own README), extracts all 16 members of TODAY21.ZIP cleanly, including the 4 that had been stuck (TODAY.APR, TODAY.MAR, TODAY.OCT, TODAY21.EXE). This gave a second executable's version banner directly from the binary (TODAY/PC V2.1 (c) 1986, 1988 by Patrick Kincaid) and let the month-file comparison in README Section I be run against the complete v2.1 dataset rather than 9 of 12 months (result: unchanged, 99.6%).

    Update (July 2026): the LZEXE-compressed v3.4/v3.5/v3.6 binaries are now decoded too. All three carry the LZ91 signature (LZEXE v0.91). mywave82/unlzexe, a POSIX-ported build of the classic UNLZEXE decompressor, compiled cleanly with a plain clang invocation and decompressed all three without issue. This gave clean version banners directly from each binary — TODAY Version 3.4 01/05/92, TODAY Version 3.5 07/12/92, TODAY Version 3.6 11/14/93 — matching the already- known v3.6 release date exactly and pinning v3.4/v3.5's for the first time. Confirms (via a strings/diff comparison of all three decompressed binaries) that no lunar-phase code exists in any of them, and that v3.4→v3.5→v3.6 are near-identical builds differing mainly in the version string and minor internal offsets — consistent with the "pure bugfix release" pattern already established from the doc files. Every one of the seven recovered TODAY.EXE/.COM binaries has now had its compression scheme defeated and its strings read directly; none has actually been run yet (DOSBox or similar) — see below.

    Also resolved: TODAYQ1.ZIP and TODAYWIT.ZIP retrieved and compared (found via discmaster.textfiles.com, since neither was directly downloadable through pcorner.com's Wayback captures — see zips/README.md for full provenance). Both confirm the "far more expanded, separately-maintained" hypothesis: TODAYQ1.ZIP's Jan/Feb/Mar files run 5–7x longer than this repo's own, with real but small verbatim overlap (confirming shared lineage, not just a shared format) and neither Kincaid's nor Fiennes's self-referential entries present. TODAYQ2.ZIP–Q4.ZIP remain unrecovered — zero hits on discmaster.textfiles.com, and pcorner.com's actual downloads were never captured by the Wayback Machine (only its /INFO/ description pages were).

    Update (July 2026): v3.1 has now actually been run, in DOSBox. Leo Bicknell installed dosbox-staging and ran TODAY31.ZIP's TODAY.EXE directly, with the DOS date set to June 28 — Kincaid's own birthday. The live output matches the strings-based reading exactly, and confirms the program's actual on-screen formatting and color scheme (white/yellow/cyan text on blue, red highlighted years): the "Happy Birthday to..."/"On this day..."/"Thought for the day..."/"And remember..." sections, ending on the same "Pay the mortgage!" reminder and Kincaid's own birthday entry ("In 1947 Patrick Kincaid, send him a card, he'll love it.") already known from the data files — see screenshots/today31-dosbox-1947-birthday-1.png and screenshots/today31-dosbox-1947-birthday-2.png. The DISK/NOPAUSE options (documented in TODAY.DOC) also work as described, writing an identical transcript to TODAY.PRN. Also confirmed directly from the doc: the manual's own example date, "0628," is explained by "you will see why I used it" — Kincaid was pointing readers at his own easter egg.

    Update (July 2026): all seven recovered DOS releases have now been run in DOSBox — v1.0, v2.1, v3.1, v3.3, v3.4, v3.5, and v3.6, each on June 28, confirming a clear evolution in the program's UI across versions:

    • v1.0 and v2.1 (1986/87 and April 1988) run in plain monochrome text — no color, no screen-clearing at all. Neither has a DISK option (that's a v3.x addition), and neither pauses by default — both need the WAIT option to hold the screen open, confirmed directly from each release's own .DOC (v1.0's option set is just MMDD/HELP/CLEAR/WAIT/ONCE; DIR/SET overrides and NOPAUSE don't exist until v2.1). v1.0 also turned up a real quirk: it has no relative-directory support at all — it expects its data files at the hardcoded path C:\TODAY\, and fails with a misleading "Not enough memory to read file \TODAY\TODAY.JUN" if run from anywhere else (the message is generic, not an actual memory problem — the fix was simply mounting a drive where the data really does live at C:\TODAY). v2.1 has no such issue, matching its documented DIR/SET override support. Both show the birthday entry in its short original form, "send him a card." — without the "he'll love it." that v3.1 later added. See screenshots/today10-dosbox-1947-birthday.png and screenshots/today21-dosbox-1947-birthday.png (both fit on one screen, no pagination needed).
    • v3.1 (1990) is the first to add color, a drawn border, and screen-clearing — see the screenshots linked above. Ends simply with "Press ENTER to end Today and restore your screen."
    • v3.3 (1991) replaces that ending with an interactive loop, "Enter a new date as MMDD or just <ENTER> to quit," — new in this release. It also gave a live, behavioral confirmation of this document's own hypothesis about TODAY.OWN: the "And remember..." reminder section is completely absent from v3.3's output, exactly because TODAY.OWN really is missing from this release (confirmed directly: ls on the extracted archive shows no TODAY.OWN file). Per Leo Bicknell's direct observation running it, v3.3 does not clear the screen before showing the paginated continuation, unlike v3.4 onward. See screenshots/today33-dosbox-1947-birthday-1.png and -2.png.
    • v3.4, v3.5, and v3.6 (1992–93) all clear the screen before showing the continuation page (the difference from v3.3), and all have TODAY.OWN restored — "And remember... Pay the mortgage!" is back in all three. Version banners read live from each, matching the unlzexe-decoded strings exactly: 01/05/92, 07/12/92, 11/14/93. See screenshots/today34-dosbox-1947-birthday-1.png/-2.png, screenshots/today35-dosbox-1947-birthday-1.png/-2.png, and screenshots/today36-dosbox-1947-birthday-1.png/-2.png.

    Every recovered DOS release has now actually been executed and visually verified, not just read as strings — this bullet is fully resolved.

    Still open:

    • TODAYQ2.ZIP–Q4.ZIP (on pcorner.com) remain unrecovered — see update above.

    • This repository's own today.c still has no version banner, no copyright string, and no lunar-phase code at all — unlike every DOS release documented so far, including the v1.0 release its own data files are an exact copy of.

      Update (July 2026): a plausible explanation, from Leo Bicknell himself. He was briefly into the BBS scene, but this archive's files date from his college years, spent entirely on Ultrix/NetBSD machines — so he most likely went looking for (or ran across on a newsgroup) a Unix port of the software, rather than the DOS original. That would explain a bare Unix C reader with none of the DOS-specific version banner/copyright strings: it'd be a from- scratch reader written against the data-file format (the same pattern as Fiennes's ARCbbs module, Howland's CP/M port, and Jungblut's ProLine port — see Sections III/V and TODO items 4/17), not a direct transcription of Kincaid's own DOS source, even though its data files are an exact copy of Kincaid's 1986 originals. Not confirmed — the exact source (which newsgroup, whose port) is lost to time — but it reframes this from an unexplained anomaly to an expected consequence of this being yet another independent reader implementation, consistent with how widely this data format was reimplemented across platforms. (He also allows a small chance he wrote the reader himself, though he doesn't believe that's the case.)

    • Both recovered v3.3 copies are missing TODAY.OWN (present in every other version) — likely a repackaging omission rather than something genuinely dropped by Kincaid for that one release, but not confirmed either way.

    • A real, independently-archived v2.0 or v3.2 release, if either ever existed on its own (as opposed to v2.0 only being known from v2.1's changelog), has not surfaced.

  12. Golden Gate Computer Society vs. Marin-Sonoma PC User Group — same BBS, two names? Resolved. The TODAY34.ZIP manual (README Section I) gives Kincaid's distribution board as the "Golden Gate Computer Society Bulletin Board, (415) 927-1216" — the exact phone number the arcbbs-hosted TODAY_DOC calls the "Marin-Sonoma PC User Group BBS." A June 16, 1994 Usenet post by Walter Olson to ba.internet ("Marin: GGCS Internet SIG Meeting") settles it: "The Golden Gate Computer Society (formerly the Marin / Sonoma PC Users Group) is starting a new special Interest Group (SIG)..." — a straightforward renaming of the same user group/BBS, not two groups sharing one system. (Provided directly by Leo Bicknell.)

  13. Chase the MUSIC/SPACE/SCIENCE specialized-database lead. TODAY34.ZIP's changelog mentions a MUSIC database, in the same file format, "found on the GGCS BBS" and readable via TODAY's new ADD option. TODAY31.ZIP's doc reveals this was one instance of a broader, Kincaid-encouraged pattern: "create specialized databases, (SPACE.mon, SCIENCE.mon etc) to share with the rest of us." Update (July 2026): TODAY21.ZIP's doc (README Section I) shows this feature — the ALTFN option — actually dates to v2.0, March 1988, two years earlier than previously documented, and was already described the same way ("special interests such as MUSIC and SPACE"). None of these — MUSIC, SPACE, SCIENCE, or others that may exist — have been located yet.

  14. Contact Bruce Goldman, and chase his DAY9nn-*.ZIP series and NEWDAY/PC. Per TODAYBBS.DOC (see zips/README.md), Bruce Goldman was, by 1993, distributing his own expanded TODAY.nnn-compatible data files (series "907") through The Charlatan's Cabin BBS, for use with both TODAYBBS and a program called NEWDAY/PC.

    Update: a June 8, 1996 Usenet post by Michael Conley to alt.bbs.wildcat ("TODAYBBS files") confirms Goldman was still actively maintaining this data at least three years later, now as series "910" (DAY910-1.ZIP–DAY910-5.ZIP, plus a revised DAY910A6.ZIP), distributed via ftp.mustang.com/bbs.mustang.com (Mustang Software, maker of Wildcat! BBS) as well as Charlatan's Cabin BBS. The distribution model is described as "char-ware" — no registration fee, donations to charity encouraged instead.

    Goldman is now the most direct maintainer of the data found in this research (as opposed to the several independent program authors), so the single most valuable question to ask him, if he can be found and reached, is where he first got the TODAY/DAY9nn data from — directly from Kincaid, from another BBS, or from some other intermediate source. Neither the DAY9nn series nor NEWDAY/PC itself has been tracked down yet; Mustang Software's old FTP/BBS archives (if they survive anywhere) are a new avenue worth trying for that.

    In progress (July 2026): Leo Bicknell has found a likely candidate — a Bruce Goldman (linkedin.com/in/bruce-goldman-0a9535), Tools S/W Architect at Oracle, 1990–1995, Redwood City, CA. The profile itself has very little detail, but the era and location are a reasonable fit: Redwood City is the same Bay Area region as Kincaid's own Mill Valley base and the Golden Gate Computer Society BBS, which would make sense if Goldman got the TODAY data first-hand from that same regional BBS scene before redistributing it (as an expanded, separately-numbered series) through Michael Conley's Los Angeles-based Charlatan's Cabin BBS. LinkedIn contact request/message sent, no reply yet; not yet confirmed.

  15. Contact Michael Conley, author of TODAYDOR/TODAYBBS. A third independent implementer of a reader for Kincaid's data format (see README Section I and III, and zips/README.md), alongside Kincaid and Hugo Fiennes. Ran The Charlatan's Cabin BBS in Los Angeles, CA in the early 1990s. Worth the same kind of outreach that worked for Fiennes (item 9).

    Update (July 2026): Leo Bicknell found 5 candidates named Michael Conley on LinkedIn, all in the greater LA area and old enough to fit the BBS-scene timeline — but none of their profiles mention BBSs, TODAY, or anything else that clearly identifies the right one:

    Tried to narrow it down via an email address and came up empty:

    • His June 8, 1996 Usenet post to alt.bbs.wildcat ("TODAYBBS files") has "show original"/raw-source view greyed out in Google Groups' current UI, and the classic ?dmode=source raw-NNTP-source URL parameter doesn't expose it either — Google Groups anonymizes email addresses for this group entirely, not just in the UI.
    • The Wayback Machine couldn't be queried directly for an old capture of the same Google Groups URL (fetch tooling can't reach web.archive.org in this environment).
    • Did turn up one corroborating (but not new-identity-narrowing) detail: an old dial-up BBS phone list, usbbs152.lst, lists 213-656-6955 CA Los Angeles Michael Conley ... Charlatan Cabn — confirming the name/BBS pairing from an independent source, with a different phone number than the 1996 post's (213) 654-7337 (consistent with a long-running board that changed lines over the years). No email or other disambiguating detail came with it.

    Not yet tried: messaging all 5 LinkedIn candidates with a short note naming "Charlatan's Cabin BBS," "TODAYDOR"/"TODAYBBS," and Wildcat! BBS specifically — distinctive enough that the right one should self-identify immediately.

  16. Contact Jason Scott / bbsdocumentary.com, per Hugo Fiennes's suggestion (item 9) — he may know more about this program's general place in BBS history, or have leads on the sysop(s) between Kincaid and Fiennes.

  17. Mick Howland, author of the CP/M port — identified, but deceased. A fifth independent implementer of Kincaid's data format (see README Section I and III, and zips/README.md), alongside Kincaid, Hugo Fiennes, and Michael Conley — this one on a third platform (CP/M, via Turbo Pascal), released May 1989. Known contact details were thin and decades-old: amateur radio callsign VK6ZMH (Western Australia) and "c/o Perth Omen," presumably a BBS name, from his own doc file.

    Update (July 2026): Leo Bicknell confirmed his full name — Jean-Michel "Mick" Howland, VK6ZMH, of Perth, Western Australia — and that he is listed as a "Silent Key" (the amateur radio community's term for a deceased operator) in the 2008 entry of the Wireless Institute of Australia's Silent Key database (wiaawards.com, verified directly: "Jean‐Michel Howland VK6ZMH" appears under the "2008 (65)" heading). This closes the identification but forecloses direct outreach — any further leads on his port would have to come through family, estate, or someone from the Perth "Omen" BBS scene who knew him, not from Howland himself.

    Update (July 2026): his CP/M port has now actually been run. Leo Bicknell found davidly/ntvcm ("NT Virtual CP/M Machine," a cross-platform 8080/Z80/CP/M 2.2 emulator), built it on macOS via its mrmac.sh script, and ran TODAY.COM directly. It hit a real gap in the emulator along the way: ntvcm didn't implement the Z80 IN r,(C)/OUT (C),r instructions, and Howland's Turbo Pascal runtime executes IN L,(C) during startup, so the first attempt hit ntvcm's "not-implemented z80 instruction" hard exit. A ~15-line patch (IN r,(C) returns a fixed 0xFF, since ntvcm has no real I/O hardware to query, and updates flags per the documented instruction behavior; OUT (C),r is a no-op) fixed it. A pull request with this fix has been opened upstream: davidly/ntvcm#32, explaining the gap and linking back here for context.

    With that fix, TODAY.COM 0628 runs cleanly and produces exactly the expected output — the short-form birthday joke ("send him a card.") and the same June 28 entries already known from the DOS v1.0/v2.1 releases, confirming Howland's port really did reuse Kincaid's original databases unmodified. Transcript saved at screenshots/today-cpm-ntvcm-1947-birthday.txt (plain text, since this is a console-only CP/M program with no color or graphics to screenshot). This means all eight recovered program releases — seven DOS versions plus this CP/M port — have now actually been executed and visually/textually verified, not just read as strings.

  18. Contact the Golden Gate Computer Society directly — it's still an active organization. Per item 12 (now resolved), GGCS is the renamed continuation of the Marin/Sonoma PC User Group BBS Kincaid used to distribute TODAY, and per its own About page, it's been running continuously since 1981 as a 501(c)(3) — the same founding year as Butler's mainframe original, though the site gives no indication whether that's a coincidence or the group's founding is itself tied to this research in some way. Unlike every other lead in this document, this is a live, current organization with a working contact address (sec@ggcs.org, membership@ggcs.org, help@ggcs.org, P.O. Box 150624, San Rafael, CA 94915-0624) — worth asking whether any long-time members remember Kincaid, the "TODAY Conference" message area on their old PCBoard BBS, or have any surviving BBS-era records/archives.

    In progress (July 2026): Leo Bicknell has emailed info@ggcs.org asking about Kincaid and the BBS's history. Awaiting a reply.