Content Quality: Concise, well-structured News item (883 words, within the 400-1200 range). Opens with the decision and its first-party source, separates what the announcement says, the history, the planned replacement, how earlier independent coverage described the feature, JDK 28 context and an explicit 'What We Don't Know' section. Every statement about the withdrawal is attributed to the Java team's email or Inside Java post.
Source Verification: All 15 sources read from the stored snapshots (sources/2026-11/java-language-team-withdraws-primitive-patterns-after-five-previews-saying-the-feature-will-be-removed-from-jdk-28-and-constant-patterns-is-planned/source-0..14.html.gz; all HTTP 200, no archive fallback, no suspicious_patterns, gunzip sha256 of every file matches manifest). I also fetched every URL independently as raw text with a browser User-Agent; source-0, 2-11 are byte-identical to the live fetch (source-1, the HyperKitty thread, and InfoWorld pages differ only in dynamic page chrome; the email text is identical).
PRIMARY SOURCES. source-1 (mail.openjdk.org thread, sha256 dbfaf7e4863e...): email 'Withdrawing the Primitive Patterns preview feature', sender Angelos Bimpoudis, list shows 'Sender's time: Nov. 9, 2026, 6:28 p.m.' with no time zone, signed 'Angelos & Gavin', 1 participant. Confirms verbatim: 'We have decided to cease development of Primitive Types in Patterns, instanceof, and switch', 'The draft JEP ... will be withdrawn, and the feature will be removed from JDK 28', the five previews JDK 23-27 with JEPs 455/488/507/530/532, the Valhalla rationale quotes ('has been exploring a new vision for Java numerics, including the possibility of user-defined numeric types'; 'would need to support exact numeric conversion, and vice versa'; 'Given that the Valhalla work is still evolving, it would be premature to finalize Primitive Types in Patterns with their current definition of exact numeric conversion'), 'withdrawing a preview feature is unusual' / JEP 12, 'will be removed in JDK 28 and will no longer be available in that release, even with --enable-preview', 'we plan to introduce a different feature, Constant Patterns', the case Point(0, 0) example and 'A draft JEP for Constant Patterns will appear shortly.' The email does not mention JEP 401. source-0 (inside.java post, sha256 c4fd85d241e5...): a 2-sentence notice repeating the opening of the email with a 'More at' link to the thread; the fetched page shows NO byline and NO date (the date 2026/11/11 is in the URL path; byline and date appear in Inside Java's feed.xml, which the article does not cite, where the entry lists Angelos Bimpoudis and Gavin Bierman, published 2026-11-11). source-2 (Jan 9 2026 email, Fri Jan 9 23:08:05 UTC 2026, Gavin Bierman): 'Currently we have one feature in preview - Primitive Patterns'; Constant Patterns 'a subset of constant expressions, including `null`, to appear as nested patterns' - verbatim.
JEPs. source-5..9: JEP 455 (release 23), 488 (24), 507 (25), 530 (26), 532 (27), each 'Closed / Delivered', titles Preview/Second/Third/Fourth/Fifth Preview, so 'five previews' and the JDK versions are correct. source-10: JEP draft 8393282 'Sixth Preview', Status Draft, Created 2026/09/30, Updated 2026/10/13; History text proposes a sixth preview in JDK 28 'by improving the dominance rules for case labels and clarifying how a switch label is selected at run time' and 'This feature remains in preview while we refine its relationship with constant patterns' - matches. source-3 (JEP 12): 'the owner must file an issue in JBS to remove the feature in the next JDK feature release; no new JEP is needed' - matches. source-11 (JDK 28 page): last update 2026/11/2 14:47 UTC, Rampdown Phase One 2026/12/03, GA 2027/03/09, JEPs targeted include 401 and not primitive patterns - matches. source-4 (JEP 401): Status Integrated, Release 28 - matches.
INDEPENDENT SOURCES. source-13 (InfoWorld, Paul Krill, Sep 15, 2026): 'This feature is in a fifth preview.' source-12 (InfoWorld, Krill): March 17 JavaOne presentation by Dan Smith, 'senior developer in the company's Java platform group'; 'these features may change or go away instead'; primitive types in patterns 'undergoing its fourth preview in JDK 26'. source-14 (InfoWorld New Tech Forum column, Oct 5, 2026, by Simon Ritter, Deputy CTO of Azul): 'one feature of particular interest to me', 'delivered without change in JDK 27', JEP 401 and JEP 539 'the first real parts of Project Valhalla'. All quotes verbatim. Independent-outlet count: ONE (InfoWorld; two news items by one reporter plus one contributed column count as one outlet). The Ritter column is a contributed piece by an Azul employee (a JDK vendor with a stake in Java) and is NOT counted as independent; the article does not present it as such, and cites it only for what it says. All three InfoWorld items predate the withdrawal and none can support it. News category has no independence minimum, and the withdrawal itself rests entirely on the first-party email, which is the correct primary source; the article says so ('They predate it').
ALLOWLIST. inside.java and mail.openjdk.org were not on config/source_allowlist.txt. Applied 'Source allowlist and unreachable sources': inside.java is the Oracle Java Platform Group's own site (linked from oracle.com/java and dev.java; About page says it is produced by members of the Java Platform Group at Oracle, with the caveat that posts do not necessarily represent an 'official view'); mail.openjdk.org is the OpenJDK project's own list host (linked as 'Mailing lists' from openjdk.org and from openjdk.org/projects/amber/, which links the amber-spec-experts list and archive). Both read as raw text (HTTP 200). Both added as bare hostnames with justification comments, for first-party statements only, never as independent corroboration.
Anthropic disclosure: not required - the article does not mention Anthropic or an Anthropic product (Ritter's column mentions 'Anthropic's Claude Mythos' only in a passage the article does not use).
Factual Accuracy: Claim-accuracy tests run. (1) Headline: 'after five previews' correct (JDK 23-27, JEPs 455/488/507/530/532); 'Saying the Feature Will Be Removed From JDK 28' uses the email's own wording ('will be removed from JDK 28', 'removed in JDK 28') and attributes it to the team; 'Constant Patterns Is Planned' tracks 'we plan to introduce' and the body gives the hedge (no release given, draft JEP 'will appear shortly'). (2) Absence claim 'None of the independent sources cited here reports on the withdrawal; they predate it' is scoped to cited sources and holds: the three InfoWorld items are dated Sep 15 and Oct 5, plus a report on a March 17 presentation, and none mentions the withdrawal. The 'one participant' statement matches the archive ('participants (1)'). (3) Valhalla rationale is attributed to the email throughout; the JDK 28 section places JEP 401 beside it but states explicitly that the email does not mention JEP 401 and that no cited source links them - no causal link is implied. (4) Dates and time zones: email dated Nov 9, 2026 with sender's time 6:28 p.m. and no zone, stated as such; Inside Java post Nov 11 (URL/feed); Jan 9, 2026 email is UTC date; JDK 28 page timestamp in UTC; all correct. (5) Derived figures: five previews and JDK 23-27 recomputed from JEP release fields; 'Closed / Delivered' on each JEP verified; draft JEP dates verified. (6) Previously reported link /article/2026-09/16-java-27-reaches-general-availability-... exists and says GA September 15 and fifth preview. (7) 'The Herald reads this to mean that source code using the preview feature would need changing' is labelled as the Herald's reading and is consistent with JEP 12 and the email. Two imprecisions of attribution found (see concerns): the byline/date credit and the 'Java language team' shorthand.
Overall Assessment: Accurate, well-scoped coverage of a first-party announcement; every statement about the withdrawal, the five previews, the JDK 28 removal wording and the Constant Patterns plan traces to the email, the Inside Java notice and the JEP pages, and the hedges are preserved. Two attribution clarifications are filed as a corrections record; neither touches the substance of the headline. inside.java and mail.openjdk.org added to the allowlist. Featured candidacy: 1 - unusual withdrawal of a long-running preview feature, of solid interest to Java developers, but narrow in audience and sourced only to the Java team itself