Informal spec proposal to replace HTML with a simpler, strictly versioned, no-script document format capped at 1.44 MiB compressed.
Key Takeaways
Spec size would be capped at 1.44 MiB compressed tar.gz, mirroring Dillo’s floppy-disk constraint to keep browser implementation costs low.
Semantic versioning (e.g. 1.2.3) replaces the current rolling-live HTML spec; published versions never change, enabling printed reference implementations.
Strict formal grammar with zero error-recovery: non-conforming pages are rejected outright, eliminating HTML’s error-handling complexity.
No scripting; interactive needs are delegated to native clients via typed protocol links (e.g. Geo URIs for maps).
Goal is text-first document exchange, not a feature-complete web clone; CSS, JS, and app-platform capabilities are explicitly out of scope.
Hacker News Comment Review
The strict-grammar/reject-invalid-pages proposal draws immediate XHTML comparisons; commenters note XHTML failed not on technical merit but because it offered no new developer capabilities, so compliance had no incentive.
Several commenters argue scripting is unavoidable: without it, browsers must bundle every possible viewer (maps, etc.), shifting complexity from JS runtime to native client proliferation.
A recurring counter-proposal is a strict, implementable subset of existing specs rather than a clean-room fork, which could achieve graceful degradation while leveraging the existing web’s network effects.
Notable Comments
@rodarima: Author concedes the XHTML upgrade-path problem but argues strict parsing is viable from the start, citing JPEG/PNG as formats that fail cleanly rather than degrade.
@sylvinus: Proposes targeting a “web-lite” strict subset covering 95% of sites, giving existing sites an incentive to officially support it if adoption grows.
