Modern email can be built from borrowed parts

Original Article ↗ Hacker News Discussion ↗

Efforts to redesign email on top of modern web standards like HTTP are rekindling long‑standing debates about spam control, identity, and user consent. Commenters explore ideas such as first-contact approval, optional postage or micropayments to make mass mail costly, and tighter integration with chat-style messaging, while noting the immense network effects that keep SMTP in place and the power large providers wield via reputation systems. Many remain skeptical that any technically superior replacement can gain traction without broad community and industry buy‑in, especially given email’s entrenchment and the failure of past “ultimate” spam solutions.

First-contact consent & “requests” inbox

  • Many like the idea of unknown senders landing in a “requests” area before reaching the inbox, similar to modern chat apps.
  • Others note similar patterns already exist: greylisting on servers, manual whitelisting, auto-replies with confirmation links or tokens, and services like Hey/Spark.
  • Reported user reactions are mixed: some say people “hate” extra hoops; others say almost no one complains and are willing to confirm to reach them.
  • Critics argue it just moves the problem: the “requests” folder can become the new spam folder, with legitimate first contacts buried.

Email vs direct messaging

  • Some ask whether an improved email spec could replace WhatsApp-style messaging; DeltaChat is mentioned as an example bridging email and chat.
  • Others emphasize different workflows: email as a mailbox vs chat-like infinite threads, with complaints about difficulty extracting topic-specific threads for things like legal work.

Spam, postage, and consent

  • Several proposals involve “postage” or per-message cost, ideally very cheap at low volume but rising with scale, to make mass spam uneconomical.
  • Micropayments are widely seen as impractical; identity churn undermines volume-based pricing.
  • Alternatives suggested: explicit “Automated: 0/1” headers with enforcement via reputation, mandatory consent for automated mail, and per-sender pre-approval plus quarantine.
  • Reputation and blacklisting at domain/IP level are described as central to today’s anti-spam, with self-hosters often struggling.

Protocols, HTTP, and compatibility

  • Some like building on HTTP and existing tools (MTA-STS, JMAP, Web Key Directory) with incremental evolution of SMTP rather than outright replacement.
  • Others dislike “email over HTTP,” preferring DNS/MX/SRV and DANE, and warning about issues like content-addressed mail breaking mailing lists or unencrypted metadata.
  • A hybrid “NewEmail” overlay that auto-upgrades delivery when both sides support it (like iMessage vs SMS) is proposed, with backward compatibility seen as essential.

UX, accessibility, and broader skepticism

  • Several argue the real problem is UX and clients, not the protocol; others think existing GUIs are already fine and varied.
  • Accessibility issues and distracting UI on the article’s site are criticized.
  • Multiple comments stress history: countless “fix email” and “solve spam” schemes have failed; email is entrenched, “captured” by big providers, and works well enough that radical change is unlikely.