Talk:Łoś–Vaught test

Not a typo
David Eppstein asks while reverting: "none of the words here are typos; why are you picking out those ones?" not a typo is used to tell automated processes that the enclosed passage has been checked by a human and does not need correcting. This passage showed up as a spelling error because it is an unusual situation where math markup is being mixed in with English words. My spell checker ignores math markup to reduce false positives but obviously tries to check English words, and this case just gets confused and marked with an error. I tagged this so it won't show up on future spell check runs and waste another editor's time. -- Beland (talk) 18:08, 13 August 2018 (UTC)
 * Math markup is used with English words like that on many, maybe most, mathematics articles. It's already tagged as math by the template; we don't need to gunk up the markup with additional stuff telling us what anyone competent to read or edit the article would already know. —David Eppstein (talk) 18:11, 13 August 2018 (UTC)
 * In reply to the above and the edit summary "are you going to add this to all math markup on all Wikipedia articles? This is an extremely bad idea that needs to be nipped in the bud." Not at all. Most math markup is ignored automatically. This the first case I've seen where a math template is used inside a word - if the hyphen wasn't there, this would have been ignored cleanly. This template is not a new idea; it's already in use on over 3000 pages. -- Beland (talk) 20:10, 13 August 2018 (UTC)
 * It's not used inside a word, it's used as a prefix of a word. And as I thought I already said immediately above, such combinations of notation-hyphen-word are used all the time in mathematics, in many mathematics articles and maybe even most of them. —David Eppstein (talk) 20:31, 13 August 2018 (UTC)
 * By "word", I mean to include hyphenated compounds. Anyway, I traced through some code and I figured out what went wrong here, and it also required that this be happening inside a piped link (which explains why I don't see widespread problems parsing hyphenated words where part of it is math). Things are happening in the wrong order - $&kappa;$-categorical gets changed to -categorical by template removal, and then that "|-" looks like table syntax, so the spell checker is reporting that "theoremcategorical" is not a word. What's on the left of the | should never be spell-checked, so that's clearly a bug. I will add a test case and fix that so this will not show as a typo in future runs. -- Beland (talk) 21:19, 13 August 2018 (UTC)