Skip to content

docs: add a comparison table to the README - #398

Merged
ppeeou merged 1 commit into
mainfrom
docs/readme-comparison
Aug 8, 2026
Merged

ppeeou merged 1 commit into
mainfrom
docs/readme-comparison

Conversation

@ppeeou

@ppeeou ppeeou commented Aug 8, 2026

Copy link
Copy Markdown
Member

Summary

Adds a "Why FxTS?" comparison table to the README, right after the intro bullets. It positions FxTS on the axes where it actually differentiates — lazy evaluation, AsyncIterable pipelines, backpressure concurrency (concurrentPool), curried data-last composition, early termination across async work, and literal-preserving type inference — against lodash and native Array/Iterator Helpers (ES2025).

Markings are honest: lodash gets partial credit for _.chain laziness and lodash/fp, native Iterator Helpers get full credit for sync lazy evaluation and partial for early termination (sync only, async helpers still a proposal). Rows where everyone ticks the box (deep utils) are included so the table doesn't read as cherry-picked.

Also links the new p-limit/p-map migration guide (#397) for the concurrency row's audience.

Test plan

  • prettier clean; table renders correctly in GitHub markdown preview

🤖 Generated with Claude Code

Positions FxTS against lodash and native Array/Iterator Helpers on the
axes where it differentiates - lazy evaluation, AsyncIterable
pipelines, backpressure concurrency, curried composition, async early
termination, and literal-preserving inference - with honest partial
marks (lodash chain, lodash/fp, sync-only native helpers). Links the
p-limit/p-map migration guide.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@ppeeou ppeeou self-assigned this Aug 8, 2026
@ppeeou
ppeeou merged commit 2a04bab into main Aug 8, 2026
4 checks passed
@ppeeou
ppeeou deleted the docs/readme-comparison branch August 8, 2026 13:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant