This page has a backlog that requires the attention of willing editors. Please remove this notice when the backlog is cleared.
The moss project seeks to find and remove the furry green typos that have been growing on Wikipedia articles. It uses a python script named moss and written by User:Beland to automatically find misspellings, mistakes in English grammar, violations of the Wikipedia:Manual of Style, and confusing or broken wiki markup.
The moss spell checker is run against a recent set of database dumps, which are generated on the 1st and 20th of every month (but take a few days to process). All the articles in the English Wikipedia are examined. The following are ignored:
Text inside references, templates, tables, quotation marks, sections like "External links" and "Works", and some other weird places.
Capitalized words (which are presumed to be correctly-spelled proper nouns)
Words that appear in titles in the English Wiktionary (which has definitions of all words in all languages, excluding proper nouns and systematic words like chemical names and large numbers)
Words that appear in titles in the English Wikipedia (which explains some things that don't appear in the dictionary)
Words that appear in titles in the Wikispecies (which has many technical words that don't appear in the dictionary or encyclopedia)
Many mistakes are not (yet) caught:
Improper addition of 's (possessives are not added to Wiktionary, so these are excluded systematically)
Incorrect capitalization
Incorrect multi-word phrases
Wrong word used in context
Non-English language words not tagged with ((lang)) or where an English misspelling happens to be the same as a word in another language. (These are counted as correct spellings if they are in the English Wiktionary, which lists words in all languages – only the definitions are restricted to English.)
AI thinks it's trying to be a non-English language (Foreign to English Wikipedia), sorted by language (e.g. TF+el)
TS
Missing or extra whitespace or dash (or new compound). Currently included if there is a period (TS+DOT), comma (TS+COMMA), or extra space (TS+EXTRA). Missing bracket (TS+BRACKET) needs code improvements to be reliable, and the remainder of TS need sorting.
U
URL
Z
Decimal fraction missing leading Zero
I
Definitely not English (International) due to accents or mixed with punctuation (other than hyphen)
MI
Probable coMpound, non-English (International) in English Wiktionary (both A-Z and non-ASCII characters, with and without dash)
ML
Probable coMpound, transLiteration
MW
Probable coMpound, found in non-English Wiktionary
R
Regular word (A-Z only) not near a common English word
T2
Edit distance 2 from common English word
T3
Edit distance 3 from common English word
W
Not in English Wiktionary, in non-English Wiktionary
red = Probably need to fix
yellow = Unsorted - need code improvements to sort into likely vs. unlikely typos or subtypes that can be usefully processed.
blue = Probably OK (but may need to verify)
bold = actively working on fixing
grey = no longer used
Instructions for editors
Just like a regular spell checker, sometimes a word that's highlighted is really a misspelling and should be changed, but sometimes it is a correct spelling that needs to be added to the spell checker's dictionary (which in this case is the English Wiktionary and Wikispecies). For the below lists, here's how you can help:
For spelling mistakes: Click on the links to the individual Wikipedia articles, and edit them to correct the misspelling. Make sure this is actually a misspelling, and not a technical term that needs to be better explained, or an alternate spelling (possibly from a different regional variety of English).
For non-English words (including words from Old English and Middle English, since they are pronounced differently): Edit the article and use the ((lang)) or ((transl)) templates to mark all non-English passages. Template contents are ignored, so they will not show up in the next report. If you can define the word, it would still be helpful to add the non-English word to the English Wiktionary or the same-language Wiktionary if you speak that language. As of the March 20, 2019 dump, only words not found in any Wiktionary are reported by moss as misspellings. (The "home" Wiktionary for Old and Middle English words is the modern English one.)
For languages that don't have an ISO 639 code (often happens with historical languages), you can use an IETF language tag instead. Failing that, use the miscellaneous code "mis" and add an HTML comment indicating the language. For example: ((lang|mis|sharbe do kin ratz))<!-- Old Runish -->
For incorrect spellings in direct quotes:
These shouldn't be picked up by the spell checker, as text in double quotes ("") is ignored. The article probably has incorrect punctuation.
For correct spellings that belong in the dictionary: Click on the word to add it to the English Wiktionary. Remember the word might not be English (though the definition must be) and be sure to check capitalization!
For correct spellings already in the dictionary: Delete from the list. These have been added in the meantime since the database dump by other editors. They do not automatically turn red as internal Wikipedia links do.
For correct spellings not appropriate for Wiktionary:
If there is no Wikipedia article, you can either ((chem name)); for example:
((chem name|poly(1-phenylethene)))
This should not be used for chemical formulas such as H2O, for which ((H2O)) or ((chem2)) may be appropriate. For some common compounds there are specific templates available such as Template:CO2.
For DNA sequences, add ((DNA sequence)) around it.
For terms that are only relevant to one Wikipedia article (and for which the article makes clear the definition) consider creating a redirect to the article. As long as the "typo" word is in the title (as a whole word), it won't show up as a mistake in future spell checks.
For bird calls: Treat these as foreign-language words or words-as-words and put them in italics, following MOS:ITALICS. Put the call inside ((not a typo)) so it won't show up on moss spell check reports. (It doesn't matter if the double apostrophes that make the italics go inside or outside the template.)
Anything else, add ((not a typo)) around it (for example, nonsense series of letters used as examples in puzzles).
Correct or incorrect, when finished delete the entry for the word from the lists on this page (or subpages), so work won't be duplicated. (There is no longer any need for strikethru.)
If an article or section has generally bad grammar, and you don't have time to fix the whole thing, just add ((copyedit)) at the top of the article or ((copyedit|section)) at the top of the affected section. If it's just a sentence or two, ((copy edit inline)) or ((incomprehensible inline)) can go at the end of the problem passage.
If you see errors being reported from footnotes or bibliographies, check to make sure the section is titled with a standard name following MOS:APPENDIX conventions. Standard end-matter sections like "References" and "Further reading" and "Works" are ignored.
If it helps to leave a message on the article's talk page asking if the word is correct or incorrect, you can use Template:Typo help like this when editing the bottom of the talk page (leave the section header blank; it will automatically be added):
((subst:typo help|PUT WORD HERE)) -- ~~~~
Don't worry if you miss something; it will reappear in a future report if there are still mistakes.
Suggested edit summaries
If you want to help publicize this project, you can copy-and-paste these into your edit summary, if appropriate.
For Wikipedia edits:
Fix misspelling found by [[Wikipedia:Typo Team/moss]] – you can help!
Tag non-English text found by [[Wikipedia:Typo Team/moss]] – you can help!
Tag correct text as ((not a typo)) for automated spell checkers (including [[Wikipedia:Typo Team/moss]])
Fix mismatched quote marks found by [[Wikipedia:Typo Team/moss]] – you can help!
For Wiktionary edits:
Add word identified by [[w:Wikipedia:Typo Team/moss]] – you can help!
Wiktionary cheat sheet
Need to add a word to Wiktionary? The Wiktionary cheat sheet has copy-and-paste templates that make it easy for the types of words commonly encountered here, even if you've never done it before.
The most efficient list to work on if all you want to do is fix misspellings. These listings try to list all the typos from a given article, so they can be fixed all at once. It also tries to only show typos that legitimately need fixing. It's not perfect, so a few words found need to be added to Wiktionary or tagged as not English, not a typo, etc. Only a few letters are updated on each run, to avoid stale listings as the whole list takes far longer than two weeks to work through. (This also avoids duplicating recent work when listings are refreshed.)
Due to length and an increased number of false positives, typo reports for dumps 2020-05-20 and later don't include T2+, T3+, and TS+BRACKET+.
Possible typos by length
(Updated from 2022-12-20 dump.)
Longest or shortest in certain categories are shown, sometimes just for fun and sometimes because they form a useful group. Feel free to delete articles that are fixed or tagged.
Likely chemistry words
These need to be checked by a chemist and marked as ((chem name)).
Chemical formulas should be written with HTML subscripts or ((chem2)); these listings identify those that incorrectly just use regular numbers.
Chemical formulas that use Unicode subscripts (which is against MOS:SUBSCRIPT) will be detected automatically by moss_entity_check.py.
Chemical formulas that use <sub>...</sub> are allowed by MOS:CHEM, but may show up in the main typo listings above. They can be converted to use ((chem2)) to be accepted by the spell checker, and ((chem2)) is also the way to fix listings of partial formulas.
Any "possible" listings that aren't chemical formulas can be cleared from this list by adding a redirect to an appropriate target (like Dy4 Systems). Most "known" listings that aren't chemical formulas can be fixed with ((proper name)).
Articles with a large number of chemical formulas triggering the spell checker are listed here (manual check on 2022-06-20 dump; counts include potential typos other than formulas, mostly compound names):
Parsing problems (where noted) are probably resulting in words showing up in debug-spellcheck-ignored.txt that shouldn't. -- Beland (talk) 03:09, 27 December 2022 (UTC)[reply]
12/11 - Al2Si2O5 → Parsing problems? Might be leaking out of Al2Si2O5(OH)4 or <chem>...</chem>?; see Kaolinitefind all
8/3 - Fe7C3 → Form of Iron carbide - parsing problems?
8/5 - Mg3Al2 → From silicate mineral Mg3Al2(SiO4)3[1] and its compositional variations; see Pyrope - possible parsing problems
For rhyme schemes, they probably need to be re-styled to follow Wikipedia:WikiProject Poetry#Style for rhyme schemes. If this ends up making them all-caps, they won't show up here on the next run. For mixed-case rhyme scheme notations, use ((not a typo)) after making sure dashes, commas, and spaces follow the recommended style.
(All fixed as of 2022-12-20 dump!)
False positives
Is there a word that is correctly used in an article, but which shouldn't be added to Wiktionary? List it here, and Beland will fix the problem.
Is there a misspelled word in an article mentioned here that was not reported? Feel free to list it below and Beland will try to improve the code if appropriate.
These are currently over-ignored, but could be used to suggest correct spellings:
Wiktionary entries that are known misspellings (e.g. wikt:anticiliary)
In cases where there are variant spellings of the same word or phrase, Wikipedia should probably pick one and stick to it except to mention the variants. This happens with:
Compound words - whether to use a space, dash, or nothing, as in "junebug" vs. "june bug" or "email" vs. "e-mail".
Words with multiple transliterations from another language (often there are multiple systems, no particular system, or a modern system different from historical systems).
Looks like this was because of redirects with "enterpreneur" in the title. I have tagged them all ((R from misspelling)), but I'll have to change the code to ignore those, as noted above. Thanks for catching that! -- Beland (talk) 23:52, 18 October 2018 (UTC)[reply]
1 - Johan Christian Drewsen - wikt:cultication -> "Rogether" instead of "Together", at the start of a sentence. "Copenahgen" instead of "Copenhagen". They obviously didn't get picked up because of capitalisation, but thought I'd list them here anyway just in case it helps. -Xurizuri (talk) 11:09, 13 February 2021 (UTC)[reply]
Dictionary writers needed! And speakers of languages other than English!
Many words (English and otherwise) detected as potential typos have been manually triaged as legitimate words that need to be added to Wiktionary, and are listed at Wikipedia:Typo Team/moss/For Wiktionary. (Moved from this page due to length.) Many of the subpages under the misspelling main listing also have long lists of words to add to Wiktionary, which are sometimes bundled up and moved to the "For Wiktionary" subpage.
Wiktionary aims to have definitions for all words in all languages (with some exceptions), and acts as the primary database for the moss spell-checker.
Highest-frequency words missing from dictionary (a-m)
(updated 2022-12-20)
Good candidates for words to add to the English Wiktionary (which
provides English definitions for words in all languages, including all
compound words), as it seems English Wikipedia readers will frequently
encounter them. For each run, only words from half of the alphabet
are shown, to avoid duplicate work from when new dumps are being
processed.
Most of the words are not from English. To get them off this list,
you can either add an entry to the English Wiktionary (which provides
English definitions for words in all languages) or tag all instances
of the word on the English Wikipedia with ((lang)). Wiktionary
does not accept Romanizations for some languages, so those cases must
be tagged as ((transl)) or ((lang)).
Errors in punctuation (mostly quotation marks) and wiki markup generally cause confusion for readers, and also prevent the spell checker from running on these articles.
The word "manned" and related forms like "unmanned" are used in many articles, but is not gender-neutral as required by MOS:S/HE and the NASA style guide. Gender-neutral alternatives include:
Titles of sources and quotes should remain unchanged.
If the term itself is being discussed, for example to say that "manned spaceflight" is another way of saying human spaceflight.
There seems to be consensus on unmanned aerial vehicle that this and related phrases (like unmanned aerial system) should remain intact, since it is much more frequent than "uncrewed aerial vehicle" at the moment. However, when using Wikipedia's voice it is preferred to describe a UAV as "uncrewed" when not using the whole phrase.
Non-article pages that are retained for historical interest shouldn't be modified if they won't be visible to readers.
Redirects with this title should be left alone if they are redirecting readers to a gender-neutral title
If the word is found the names of articles and categories (except those with names directly related to UAVs), those should be renamed, and the links changed. Many articles have already been renamed, and the links just need to be updated. (Remember that to rename a category, all the articles in that category must be edited to change their pointers.)
Coming soon: moss report on "manned" that ignores references, page titles, proper nouns, and consensus-OK phrases.
These may need to be discussed before being changed.
Manned Venus flyby - Based on the NASA style guide, NASA probably would now refer to this as "human Venus flyby" but historical sources say "manned Venus flyby" so that's what the majority of editors commenting on the talk page currently favor. There is some question as to whether the scope of the article concerns a specific mission or this type of mission in general, which is related to the proper name exception (but then the title would be "Manned Venus Flyby"). Compare Colonization of Venus and Human mission to Mars. -- Beland (talk) 19:41, 21 May 2019 (UTC)[reply]
MOS:NOTE says to avoid the following phrases when they address the reader directly. Not all instances are problematic, such as those in direct quotations.
rhetorical questions, especially in headings - find all questions in headings (some cases, like the names of works, are not problematic)
Internationally comprehensible spelling and vocabulary[edit]
MOS:COMMONALITY advises the use of vocabulary and spellings that are shared across national varieties of English, where possible. This section collects instances where an unshared term is being used which could be improved. For proper nouns and direct quotes, a translation or re-spelling into another dialect may be helpful.
For the UK, Irish, Australian, New Zealand, and South African pound, ₤ should be changed to £
₤ is OK to use with Italian lira. Changing e.g. ₤100,000 to [[Italian lira|₤]]100,000 will prevent legitimate uses from showing up in automated reports, and also help readers understand that this is not British pounds. (Mentions of Italian lira are increasingly rare because it has been replaced by the Euro.)
Caution: Not all problem pages show up reliably; if you do a search, fix all the pages in the results, and then do another search, you will probably get a fresh batch of problem pages. It may also take a minute or two for fixed pages to disappear from the results, due to lag updating the search index.
Work is in progress on detecting and fixing other MOS-related issues with numbers and currencies.
Remove, repair, or convert the HTML markup to wiki markup yourself.
Tag the article ((cleanup HTML)) and it will show up under Category:Articles with HTML markup but not on this list. Use the "tags" parameter to indicate which tags are present on the page; many editors find it hard to locate the offending HTML. For example: ((cleanup HTML|tags=table, cite))
Many of these can be replaced by ((var)) (for text to be replaced) or ((angbr)) (e.g. for linguistic notation). Enclose in <code>...</code> for inline software source code.
"Most likely misspellings by articles" should always have work to do (if not, ping Beland to add more from the current dump). Some of the other sections are occasionally waiting for a new dump to get a useful list, either because they are ranked by frequency or a code change has been made to clean up noise in the next run. New runs are generally posted twice a month. The database snapshot from the first day of the month generally takes about 9-13 days to process, and the snapshot from the twentieth day of the month might take 4-6 days until it can be posted.
All that said, if you want to get a ping when results from a new dump are posted, you can add your name to the list below. If you are only interested in a particular section, include a note to that effect.