Microsoft open-sources "the earliest DOS source code discovered to date"

Original Article ↗ Hacker News Discussion ↗

Overall reactions to the DOS source release

  • Many see it as valuable historical preservation and thank Microsoft for releasing it.
  • Some argue it came “decades too late” to be maximally relevant.
  • A few dismiss it as minor or “too little, too late,” or joke it’s just more training data for code assistants.

Hopes for Windows and other source releases

  • Strong desire for early Windows source, especially Windows 2000, which several call the best or favorite version.
  • Skepticism that newer Windows will ever be opened due to third‑party code, licensing entanglements, and massive review costs.
  • Leaks of NT4/XP/2000 are mentioned, but their illegality limits use.

Nostalgia for early computing and low-level understanding

  • Multiple comments praise the era when a small team could understand the whole machine and write a few thousand lines of assembly to ship a product.
  • Some lament that many modern programmers start at high levels (JS/Python/AI tools) and never learn the “whole stack.”
  • Others counter that no one truly understands the full modern stack and even assembly is “high level” compared to modern CPUs.

Technical nature and relevance of DOS

  • DOS is noted as single‑user, all ring 0, with no networking; some argue this makes the concept of “vulnerabilities” almost moot, aside from physical access or floppy infection.
  • Primary modern value is seen as historical: understanding design choices that influenced x86 and OS structure.

Preservation, OCR, and paper vs digital

  • Discussion of the recovery effort from aging dot‑matrix listings, including OCR difficulties and manual “global substitutions” to clean recurring character errors.
  • Debate over whether this is a “win for paper”:
    • One side: printed listings outlived original digital media.
    • Other side: printouts were barely legible and required heroic recovery; digital was there originally.
  • Broader reminiscences about punch cards, paper tapes, and why code was once routinely printed.

Microsoft BASIC and language history

  • Some argue early Microsoft’s more important contribution was BASIC, not DOS.
  • Microsoft BASIC’s lineage from Dartmouth BASIC via BASIC‑PLUS is discussed.
  • Popularity of Microsoft’s interpreted BASIC variants is said to have effectively derailed ANSI efforts around more structured “SBASIC,” leading instead to True BASIC, which ultimately faded.

AI, leaks, and legal minefields

  • Speculation about using LLMs to reconstruct Windows (e.g., ReactOS) is called a legal minefield, especially since leaked Windows source is accessible online.
  • Questions raised about whether LLM training on that source would taint outputs; comments joke about “fair use,” “derivative work,” and indemnification policies, but the legal status is portrayed as risky and unclear.
  • One commenter mentions already using an LLM to help rebuild leaked Windows 2003 components but finds it hard to “wash” from leak contamination.

Industry history and business pivots

  • Recounting of the IBM–CP/M–Microsoft story: IBM wanted CP/M; negotiations with its vendor allegedly failed; Microsoft stepped in with DOS.
  • Another commenter notes many details of that narrative are disputed, but agrees CP/M’s lack of a ready 16‑bit version hurt its chances.
  • The Altair BASIC story is held up as Microsoft’s real early technical achievement, with emphasis on constraints (no Altair hardware, tiny memory budgets, last‑minute bootloader).

Entrepreneurship, competition, and changing landscape

  • Envy that a small assembly codebase once sufficed to start a major software company.
  • Others argue code was never the main barrier: product, distribution, and business execution mattered more, then and now.
  • Debate over how much easier or harder competition is today, with comments that modern tools (including LLMs) can clone features quickly but that real differentiation still lies beyond raw code.

Broader reflections: digital archaeology

  • Several commenters frame this work as “digital archaeology”:
    • Recovering lost source from decayed media.
    • Reconstructing historical toolchains and formats.
    • Preserving context for future historians analyzing early software, standards, and system design.