Quran word glosses and English dictionary alternatives Prepared 2026-11-01 1. BASELINE ENGLISH WORD GLOSSES Downloaded file: r3-wbw.en.json Source: R3GENESI5/quran-bil-quran, app/data/translations/wbw.en.json https://github.com/R3GENESI5/quran-bil-quran https://raw.githubusercontent.com/R3GENESI5/quran-bil-quran/master/app/data/translations/wbw.en.json Git blob SHA: eec442a6bc777b46f2b66dfbb944c09222708806 Source tree SHA: ad76e8e3b0928e300cd24a74a100c745384cc993 SHA-256: 32c58b3b7e2bb7215d2c3b28b568f58cf704d2ea9f16a908acc5a85583dcf3b2 The repository README credits Quran.com API for its word-by-word English glosses. Its download_translations.py retrieves English word translations in verse/word order from the Quran.com v4 API, excluding verse-end markers. The repository README states MIT, but it does not provide a separate LICENSE file, and that statement does not independently establish the copyright or licence of the underlying Quran.com word translations. The upstream licence of the English word-by-word resource remains unconfirmed. Do not describe the upstream translation as MIT, GPL, public domain or Creative Commons. The available README and download script are preserved locally. Official QUL resource examined: https://qul.tarteel.ai/resources/translation/92 https://qul.tarteel.ai/resources/translation/92/docs https://qul.tarteel.ai/resources/translation/92/licence QUL's licence page reports that it does not have copyright information for this resource. It therefore does not resolve the upstream licence question. The downloaded baseline was obtained from the GitHub mirror, not from QUL. Source coverage: 6,236 verses and 77,429 source glosses. The app uses 77,432 word positions. Five documented segmentation adjustments align them; see alignment-audit.json. Surrounding whitespace was removed from 497 glosses. No words in the source SQLite were changed. Locations always mean the tuple (surah_n, ayah_n, word_in_ayah), never a page glyph's word_global value. Segmentation crosschecks from the Quranic Arabic Corpus: https://corpus.quran.com/wordbyword.jsp?chapter=2&verse=181 https://corpus.quran.com/wordbyword.jsp?chapter=8&verse=6 https://corpus.quran.com/wordbyword.jsp?chapter=13&verse=37 https://corpus.quran.com/wordbyword.jsp?chapter=27&verse=20 https://corpus.quran.com/wordbyword.jsp?chapter=37&verse=130 These pages corroborate the source word segmentation/glosses. No bulk Corpus translation data were copied. In 37:130 the gloss Elijah belongs to the combined name; neither separate app fragment has an independent gloss. The output alternatives-by-exact-arabic.json aggregates source glosses only where the app's text_uthmani string is exactly identical, including diacritics and recitation marks. No Arabic normalization, root matching, lemma matching or suffix removal is used. Multiple context glosses are suggestions; their presence does not mean every gloss fits every occurrence equally well. 2. OPEN ENGLISH WORDNET Downloaded official release: Open English WordNet 2025 (2025-12-31). https://en-word.net/downloads https://en-word.net/downloads/english-wordnet-2025-json.zip Attribution: Open English WordNet Community, derived from Princeton WordNet. Licence: Creative Commons Attribution 4.0 International, with retained Princeton WordNet attribution and licence notices. https://creativecommons.org/licenses/by/4.0/ https://wordnet.princeton.edu/license-and-commercial-use Full licence notices are included as OEWN-LICENSE.md and WNDB_License.txt. Those notice files came from: https://raw.githubusercontent.com/globalwordnet/english-wordnet/main/LICENSE.md https://raw.githubusercontent.com/globalwordnet/english-wordnet/main/WNDB_License.txt Modification: wordnet-quran-lookup.json is a subset/index of the official JSON release, limited to English terms or documented/derived base-form lookups found in Quran baseline glosses. Definitions, synonym terms and examples retain their source wording; examples with a source credit append that credit with an em dash so the examples array consistently contains strings. Dictionary matches retain lemma, part of speech and synset ID. Documented English forms and heuristic English base-form lookups are explicitly distinguished from exact English lemma matches. Dictionary alternatives are general English senses, not independently verified translations of the Arabic. The interface should present them by sense with definitions and label them dictionary forms. A dictionary base form must not automatically replace an inflected phrase. For example, created may look up create, but and created must not silently become and create. Custom wording can be entered explicitly by the user. Suggested visible attribution: English word glosses: Quran.com, via the R3GENESI5/quran-bil-quran data mirror. Dictionary alternatives: Open English WordNet 2025, by the Open English WordNet Community, derived from Princeton WordNet (CC BY 4.0). This app subsets and indexes the dictionary and aligns five source verse segments. Link to this source/changes notice and the full dictionary licence files. 3. INDEPENDENT MIRROR USED ONLY FOR VALIDATION https://github.com/MDMOQADDAS/data-quran-sharif https://raw.githubusercontent.com/MDMOQADDAS/data-quran-sharif/master/word-translation/en-qurancom.json https://raw.githubusercontent.com/MDMOQADDAS/data-quran-sharif/master/word-text/uthmani-qurancom.json This repository identifies its data as CC BY-NC-ND 4.0 and credits the Hablullah team/data-quran. Its README and licence are preserved as md-README.md and md-LICENSE.txt. It was used only to crosscheck word ordering and source glosses, not to generate the production baseline or dictionary. Do not bundle this mirror's raw files into the app as reusable MIT data. All 77,429 English glosses agree with the R3 source after surrounding spaces are trimmed. Arabic order was checked after the same five segmentation adjustments. Script-form comparison removes marks and folds documented Uthmani script variants for this validation ONLY; it is never used for application exact-word keys. One residual orthographic variant is 11:13:3: app ٱفۡتَرَىٰهُۖ versus mirror افْتَرَاهُ ۖ, representing the same word at the same location. There were no detected additional positional discrepancies. See mirror-crosscheck-summary.json. 4. GENERATED FILES / REPRODUCTION build_english_resources.py reads the original SQLite in mode=ro and writes: - baseline-by-location.json: every app word position; two nulls are covered by the shared proper-name span rather than invented independent meanings. - shared-name-spans.json: the one shared name span at 37:130:3-4. - alternatives-by-exact-arabic.json: exact Arabic -> [{text,count,example}]. - alignment-audit.json: source positions, app positions and all five changes. - wordnet-quran-lookup.json: lookup index and sense dictionary. - resource-stats.json: counts and content hashes. No generated wording is represented as a source translation. User-saved English choices should be stored separately from the unchanged source data.