Maintainer
Thomas E. Dickey has recorded every change since 1.9.9e, and maintained ncurses from 1.9.9g through 4.1 and again from 5.0. He has revised every manual page, which Eric S. Raymond first wrote.
ncurses's own test programs, 10 seconds each. Click or focus the terminal to take over.
ncurses and its documentation are the work of its maintainer and contributors. All the text on this site is theirs.
Maintainer
Thomas E. Dickey has recorded every change since 1.9.9e, and maintained ncurses from 1.9.9g through 4.1 and again from 5.0. He has revised every manual page, which Eric S. Raymond first wrote.
Principal authors
Juergen Pfeifer, Eric S. Raymond, Alexander V. Lukyanov, G. Branden Robinson, Philippe Blain, Sven Verdoolaege and Nicolas Boulenguez, in the order of their contributions in the project’s AUTHORS file.
Originators
Pavel Curtis originated ncurses, and Zeyd M. Ben-Halim was its original maintainer.
Guides
Writing Programs with NCURSES by Eric S. Raymond and Zeyd M. Ben-Halim, A Hacker’s Guide to NCURSES by Eric S. Raymond, and the NCURSES Programming HOWTO by Pradeep Padala.
Many more people contributed patches, reports and analysis. The project’s NEWS file records each of them. See Credits for the full list and sources.
This edition is moving every ncurses document into Merlon Docs shards, then reviewing each one against its sources. It counts two steps. Sharding counts documents: each guide chapter, manual page and contributor document once. Review counts tokens, so a long manual page weighs more than a short one; the tokenizer picker below changes them. Done gives half for the documents sharded and half for the tokens reviewed.
65% done
15 of 15 chapters sharded; 70,313 of 70,313 tokens reviewed. The HOWTO, Writing Programs with NCURSES and the hints, each chapter checked against its sources and the library.
133 of 133 pages sharded; 18,717 of 322,862 tokens reviewed. Every page in man/, split into one troff fragment per section. The editorial review hasn't started.
4 of 4 documents sharded; 44,315 of 44,315 tokens reviewed. The hacker's guide, INSTALL, the Ada binding and the announcement.
What a reader, or a language model, has to take in: the documents this edition started from (the upstream snapshot v6_6_20261107) and the shards that replaced them, each read whole as stored. Tokens depend on the model's tokenizer.
Exact for GPT-4o, GPT-4.1, the o-series and GPT-5.
| Track | Before | Now | Tokens before | Tokens now | Change |
|---|---|---|---|---|---|
| Guides | 2 files, 40,067 words | 82 files, 35,653 words | 103,003 | 70,313 | −32% |
| Manual pages | 133 files, 101,203 words | 1,187 files, 103,334 words | 311,693 | 323,673 | +4% |
| Contributor and install docs | 9 files, 25,974 words | 50 files, 24,533 words | 59,853 | 44,315 | −26% |
| All | 144 files, 167,244 words | 1,319 files, 163,520 words | 474,549 | 438,301 | −8% |
Before: the guides' HTML (the HOWTO and Writing Programs with NCURSES), the troff pages in man/, and the hacker's guide, INSTALL and READMEs, the Ada page and the announcement. Now: their shards under docs/. Words leave out markup. Measured bydoc-sizes.py; rerun it after a change to the docs.
How the tokens are spread over files: a shard is read whole, so its size is what a reader, or a language model, takes in at once. The tokenizer above switches the chart too.
| Track | Files | 25th percentile | Median | 75th percentile | Largest |
|---|
One agent session did the work, and Betsalel Williamson, who runs this edition, reviewed and merged each pull request. The ncurses maintainer, Thomas E. Dickey, hasn't reviewed this work. Figures are as of November 18, 2026.
Every count above comes from git, at build time. The history below keeps the same counts for each merged pull request, taken at its merge commit, inmigration-history.json. Which pull request reviewed each document is listed inmigration-units.csv. To check both against the repository, clone it and run:
node site/scripts/migration-history.mjs verifyIt recounts every row from git and fails on any difference. It also checks each review claim: the pull request merged, changed that document, and recorded its corrections in the errata. CI runs it on every pull request. The token and cost figures are the exception. They come from the agent's session API, and the repository keeps each response's numbers insite/data/session, along with the session's time totalled by day, which the time figures are the sums of.node site/scripts/migration-stats.mjs verify checks the figures above against those files.
| Pull request | Merged | Done (documents) | Guides | Manual pages | Contributor docs | Errata |
|---|---|---|---|---|---|---|
| #84 | November 18, 2026 | 58% | 15/15 of 15 | 133/6 of 133 | 4/4 of 4 | 432 |
| #82 | November 18, 2026 | 58% | 15/15 of 15 | 133/5 of 133 | 4/4 of 4 | 425 |
| #81 | November 18, 2026 | 58% | 15/15 of 15 | 133/4 of 133 | 4/4 of 4 | 423 |
| #80 | November 18, 2026 | 57% | 15/15 of 15 | 133/1 of 133 | 4/4 of 4 | 410 |
| #79 | November 18, 2026 | 57% | 15/15 of 15 | 133/1 of 133 | 4/4 of 4 | 407 |
| #78 | November 18, 2026 | 57% | 15/15 of 15 | 133/1 of 133 | 4/4 of 4 | 403 |
| #77 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 395 |
| #76 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 391 |
| #75 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 387 |
| #74 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 380 |
| #73 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 378 |
| #72 | November 17, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 362 |
| #71 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 357 |
| #70 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 348 |
| #69 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 345 |
| #68 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 342 |
| #67 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 336 |
| #66 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 304 |
| #65 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 303 |
| #64 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 302 |
| #63 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 336 |
| #62 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 302 |
| #61 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 302 |
| #60 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 302 |
| #59 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 302 |
| #58 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 298 |
| #57 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 298 |
| #56 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 291 |
| #55 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 275 |
| #54 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 272 |
| #53 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 262 |
| #52 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 4/4 of 4 | 261 |
| #51 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 3/3 of 4 | 254 |
| #50 | November 15, 2026 | 56% | 15/15 of 15 | 133/0 of 133 | 3/3 of 4 | 254 |
| #49 | November 15, 2026 | 55% | 15/15 of 15 | 133/0 of 133 | 2/2 of 4 | 245 |
| #48 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 1/1 of 4 | 221 |
| #47 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 1/1 of 4 | 221 |
| #46 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 1/1 of 4 | 221 |
| #45 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 1/1 of 4 | 221 |
| #44 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 0/0 of 4 | 195 |
| #43 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 0/0 of 4 | 192 |
| #42 | November 15, 2026 | 54% | 15/15 of 15 | 133/0 of 133 | 0/0 of 4 | 192 |
| #41 | November 15, 2026 | 50% | 15/15 of 15 | 122/0 of 133 | 0/0 of 4 | 189 |
| #40 | November 15, 2026 | 35% | 15/15 of 15 | 75/0 of 133 | 0/0 of 4 | 186 |
| #39 | November 15, 2026 | 29% | 15/15 of 15 | 57/0 of 133 | 0/0 of 4 | 185 |
| #38 | November 15, 2026 | 26% | 15/15 of 15 | 48/0 of 133 | 0/0 of 4 | 184 |
| #37 | November 15, 2026 | 26% | 15/15 of 15 | 48/0 of 133 | 0/0 of 4 | 184 |
| #36 | November 15, 2026 | 24% | 15/15 of 15 | 42/0 of 133 | 0/0 of 4 | 183 |
| #35 | November 15, 2026 | 17% | 15/15 of 15 | 23/0 of 133 | 0/0 of 4 | 182 |
| #34 | November 15, 2026 | 13% | 15/15 of 15 | 9/0 of 133 | 0/0 of 4 | 181 |
| #33 | November 15, 2026 | 12% | 14/14 of 15 | 9/0 of 133 | 0/0 of 4 | 179 |
| #32 | November 15, 2026 | 12% | 13/13 of 15 | 9/0 of 133 | 0/0 of 4 | 172 |
| #31 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 162 |
| #30 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 162 |
| #29 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 162 |
| #28 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 162 |
| #27 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 161 |
| #26 | November 15, 2026 | 11% | 12/12 of 15 | 9/0 of 133 | 0/0 of 4 | 161 |
| #25 | November 15, 2026 | 10% | 11/11 of 15 | 9/0 of 133 | 0/0 of 4 | 152 |
| #24 | November 15, 2026 | 10% | 11/11 of 15 | 9/0 of 133 | 0/0 of 4 | 152 |
| #23 | November 15, 2026 | 10% | 10/10 of 15 | 9/0 of 133 | 0/0 of 4 | 143 |
| #22 | November 15, 2026 | 10% | 10/10 of 15 | 9/0 of 133 | 0/0 of 4 | 143 |
| #21 | November 15, 2026 | 10% | 10/10 of 15 | 9/0 of 133 | 0/0 of 4 | 143 |
| #20 | November 15, 2026 | 10% | 10/10 of 15 | 9/0 of 133 | 0/0 of 4 | 143 |
| #19 | November 15, 2026 | 10% | 10/10 of 15 | 9/0 of 133 | 0/0 of 4 | 142 |
| #18 | November 15, 2026 | 9% | 9/9 of 15 | 9/0 of 133 | 0/0 of 4 | 133 |
| #17 | November 15, 2026 | 9% | 9/9 of 15 | 9/0 of 133 | 0/0 of 4 | 133 |
| #16 | November 15, 2026 | 8% | 8/8 of 15 | 9/0 of 133 | 0/0 of 4 | 127 |
| #15 | November 14, 2026 | 8% | 7/7 of 15 | 9/0 of 133 | 0/0 of 4 | 119 |
| #14 | November 14, 2026 | 7% | 6/6 of 15 | 9/0 of 133 | 0/0 of 4 | 112 |
| #13 | November 14, 2026 | 7% | 6/6 of 15 | 9/0 of 133 | 0/0 of 4 | 109 |
| #12 | November 14, 2026 | 6% | 5/5 of 15 | 9/0 of 133 | 0/0 of 4 | 100 |
| #11 | November 14, 2026 | 6% | 4/4 of 15 | 9/0 of 133 | 0/0 of 4 | 94 |
| #10 | November 14, 2026 | 5% | 3/3 of 15 | 9/0 of 133 | 0/0 of 4 | 79 |
| #9 | November 14, 2026 | 4% | 2/2 of 15 | 9/0 of 133 | 0/0 of 4 | 59 |
| #8 | November 14, 2026 | 4% | 1/1 of 15 | 9/0 of 133 | 0/0 of 4 | 39 |
| #7 | November 14, 2026 | 4% | 1/1 of 15 | 9/0 of 133 | 0/0 of 4 | 39 |
| #6 | November 14, 2026 | 4% | 1/1 of 15 | 9/0 of 133 | 0/0 of 4 | 33 |
| #5 | November 14, 2026 | 2% | 1/1 of 15 | 5/0 of 133 | 0/0 of 4 | 0 |
| #4 | November 14, 2026 | 2% | 1/1 of 15 | 5/0 of 133 | 0/0 of 4 | 0 |
| #3 | November 14, 2026 | 2% | 1/1 of 15 | 5/0 of 133 | 0/0 of 4 | 0 |
| #2 | November 14, 2026 | 2% | 1/1 of 15 | 5/0 of 133 | 0/0 of 4 | 0 |
| #1 | November 14, 2026 | 2% | 1/1 of 15 | 5/0 of 133 | 0/0 of 4 | 0 |
Each track shows sharded/reviewed of its total. The history counts review in documents, since tokens are measured only for the current shards.
Edited by betsalel-williamson with Merlon Docs: the existing documentation, reorganized so each fact appears once. The manual pages are generated from the same troff sources the project ships. This is an unofficial edition. The official ncurses site is invisible-island.net/ncurses.