Content Quality: A News piece of about 1,043 words (range 400-1200) with a clear lead, a section on what the paper argues, the six options, The Register's report, the Herald's own capture comparison, and a closing list of what the sources do not settle. Every contested statement about named people is attributed to its author (Sutter, Pusz, Fertig, Bott, The Register) and reported as that person's statement, with no motive speculation and no claim in the Herald's voice about who caused the ballot result. The Register's 'sore losers' passage about Fertig's view of Stroustrup is correctly not repeated.
Source Verification: Route: chief:review snapshotted 3 of 5 sources (source-0 The Register, source-1 the isocpp.org PDF stored as raw PDF bytes in source-1.html.gz, source-4 train-it.eu); both Internet Archive URLs returned HTTP 429 to the script (file null in manifest). I read them myself by live curl with a browser User-Agent, retrying with backoff until HTTP 200, then extracted text with pdftotext and diffed. Hashes (sha256 of raw bytes): live isocpp.org PDF and stored snapshot 27e5bb8fb5ca576a7459ec5178516e7e779dd523a3710124dc0373d396412b6e (live Last-Modified: Wed, 11 Nov 2026 15:52:06 GMT); capture 20261111193217 identical to the live file; capture 20261108220233 70cba8a5f755d37e2ee539e6d09dbc1dd78503df0d39dacc0334e503c0d2fef7; capture 20261109115953 d6c04b8c8388e4947b4c6d51cea6f1470486669f72934cc8fc48de10988af3c2. The Register snapshot sha256 b60fba03a4555341071bc4aa28fc52831c6a7d2bb53b90c4227b58a1e3c23fdb; train-it.eu d54d82d000b58703bc32af83ed98e6b1e2a82ca23df2eab26ec35e096cb91214. open-std.org timed out (60 s) for both the papers index and p4454r0.pdf; wg21.link/p4454r0 returned 404. suspicious_patterns: null for all three snapshotted sources; the two archive URLs were not scanned by the script, and I found no instruction-like text addressed to an AI in the PDFs. (1) The Register (Joab Jackson, 'fri 13 Nov 2026 // 07:31 UTC'): confirmed 'The ballot saw 13 votes cast in favor, 9 against, and 12 abstentions, failing to meet the ISO requirement of having no more than 25 percent negative votes' and 'Contracts was the major sticking point for those voting nay'; Fertig ('a member of the C++ standardization committee') quote and the 'trust and respectful collaboration, has been shaken' quote verbatim; Bott, 'CEO of the C++ Alliance', email quote verbatim. The Register never mentions Sutter or P4454 (no 'Sutter' in the page) and gives no source for the counts: both absence claims hold. 13+9+12 = 34 ballots; the article does not state a sum or percentage. The ISO ballot is not public and I found no primary record of the tally: neither the paper nor Pusz's post gives one, and Pusz says ISO policy allows the overall outcome but not how any national body voted, with the full report due by mid-January 2027. The counts therefore rest on The Register alone and are reported as its figures, which the headline also does. (2) P4454R0 (Document Number P4454 R0, Date 2026-10-31, Reply-to Herb Sutter, Audience WG21): current text confirmed verbatim for the tl;dr sentence, 'doesn't try to relitigate...', 'neither proposes that Bjarne become a BDFL nor aims to blame anyone', 'This was decisive', the 'personal communication' quote 'design is fundamentally flawed', the Section 4.2 sentence and the Ed. note on Option B; 'five people' confirmed. Diff of 8 Nov capture vs live text confirmed every difference the article lists: title 'its owners' vs 'steward body'; the 8 Nov tl;dr 'The core decisive issue that caused the DIS to fail is that the language's creator and its owners disagree'; the 8 Nov sentence 'Of those additional NBs who voted Disapprove, if even one had voted in favor and one abstained, with all the others still against, the DIS would have passed' and 'hence Bjarne's dissent was the pivotal factor' (neither in the live text; the footnote instead says 'Without those additional recruited Disapprove votes ... the DIS would have passed'); 'I privately reconfirmed with the key people in two NBs' about P3911 (absent live, replaced by a reference to P4451R0). Both versions show Date 2026-10-31. Headline test: the paper as it NOW reads still argues the point (tl;dr 'The decisive issue...'; 'his voice ... was what changed the outcome'; 'Over half of the NBs that voted Disapprove ... This was decisive'; footnote 4 'Without those additional recruited Disapprove votes ... the DIS would have passed'), so the present-tense 'Argues' is supported by the live text, not only by the 8 Nov capture; the article reports the softened wording separately. The CDX index for the PDF shows 4 distinct content digests (8 Nov 22:02 to 9 Nov 11:14; 9 Nov 11:59; 9 Nov 19:25 to 10 Nov 10:41; 11 Nov 19:32 onward). The 9 Nov 11:59 capture, which I read, already has 'steward body' and lacks 'pivotal factor' and 'if even one had voted', so the article's two-copy comparison is accurate but the intermediate changes are larger than 'other captures in between' conveys (see corrections). The 9 Nov 11:59 capture is not described in the article; the article labels its comparison as the Herald's own. Captures give capture times, not change times, and the article uses capture timestamps only. (3) Options A-F recounted: six, each described accurately and in attributed summary form; the paper says its list 'isn't exhaustive' and the article states 'at least' via 'lists options', acceptable. (4) train-it.eu (Pusz, 25 Oct 2026): 'the overall outcome of a vote', 'not how any national body voted', 'recommend a course', 'from 16 to 21 November 2026', Polish national body membership and 'My preference is to keep them' all confirmed; attributed individual-expert commentary, not independent. (5) Earlier Herald article 2026-11/05 exists at the linked path and its body contains no vote tally: description correct. Independence: one independent outlet (The Register); the paper (author Sutter), Pusz, Fertig (committee member) and Bott (CEO of the C++ Alliance, the C++ Alliance authored P4238R1, covered in the Herald's earlier article, so it is not a neutral party) are attributed voices and are treated as such. Allowlist: isocpp.org and web.archive.org were not listed; I added both with justification comments (see Allowlist decision in recommendations). Anthropic disclosure: not required - Anthropic and its products are not part of the story and not named in the headline, summary or lead.
Factual Accuracy: All quotes verified verbatim on the exact page or capture. Dates and times (2026-10-31 paper date; 13 Nov 07:31 UTC Register; 8 and 11 Nov capture UTC times; 25 Oct Pusz) correct. The Register's counts are correctly framed as The Register's and as unattributed to a primary source. Two subordinate omissions are filed as clarifications: the intermediate archive versions, and the absence of a primary record for the tally. A third point is noted under concerns: 'Stroustrup's own response to the paper; none of the cited sources reports one' is accurate for a public response, but the paper itself relays Stroustrup's comments made while reviewing it, which the article reports elsewhere.
Overall Assessment: The headline, summary and lead are each supported by a source I read: the paper's current text still argues that Stroustrup's disagreement was the decisive issue, the diff of the 8 and 11 November captures is exact, and The Register's article contains the 13/9/12 counts, which the article attributes to The Register alone. Named individuals are treated with strict attribution and no motive speculation. Two clarifications are filed (intermediate versions; no primary record of the tally) and one absence sentence is noted. Featured candidacy: 2 - a failed ISO ballot for C++26 matters to a broad programming audience, and the paper's documented wording changes are distinctive original reporting