spellcheck should recognize the language
Title:
spellcheck should recognize the language
Area:
entries, comments
Summary:
The site's spellcheck should be smarter. It should make a best guess of what language it likely is from the text entered, and apply the right kind of dictionary.
Description:
I know some browsers have a spellcheck of their own, but I always use the site's spellchecker. Only it is not as smart as a spellcheck should be. Usually I type English comments, and that's mostly fine (though the dictionary could use some upgrade too), but today I hit spellcheck as I always do automatically, but happened to have typed a German comment.
The output was a ridiculous, annoying wall of red. But computers these days are not so bad at recognizing common languages (at least if I let that google translator guess on a language bit rather than choosing, it seems to guess sensibly and the translations actually have gotten less ridiculous than a few years ago too), so why couldn't the DW spellchecker see that the comment clearly wasn't English, recognize the language, and pick the correct dictionary automatically?
I don't know how much effort it would be technically to get text based language recognition, and I wouldn't want the dictionary be chosen dumbly based on location settings or language settings in the browser, because then I might get the red wall for English comments and that would be worse. But a spellchecker would be much more convenient if it was smarter and did things it is supposed to do automatically.
I don't see any downsides, because if the language recognition should fail for some languages, you are not worse of than now, as I assume then it would just default to English as the site language.
This suggestion:
Should be implemented as-is.
24 (44.4%)
Should be implemented with changes. (please comment)
4 (7.4%)
Shouldn't be implemented.
4 (7.4%)
(I have no opinion)
22 (40.7%)
(Other: please comment)
0 (0.0%)

no subject
no subject
no subject
no subject
no subject
no subject
(My own experience is that it's annoying to be told that I've misspelled "colour" when no, actually, I've spelled it exactly as I've always been taught. So if it could manage to autodetect British vs USan English that'd be awesome. But having everything underlined would be even more annoying so seems the main priority.)
no subject
no subject
no subject
For bonus shiny, there might be a link to the definition in the other language, and a list of suggested translations into the primary language. That might be useful in the case I've seen every now and then, where someone's talking in one of their non-primary languages, and the word for what they're thinking about in the language they're talking isn't coming to mind readily, but they know the word they're thinking of in a different language. With a tool like that (which is probably more of a language-focused-company type tool, a little beyond Dreamwidth's scope) the word in the language they're talking in that they knew they were looking for might be offered, leading to less headpounding when communicating.
no subject
no subject
So what I was trying to describe up there was a sort of compromise between the two: a spellchecker that would tell you when you've used a word in a language that didn't match the main body of text you were writing in, but also wouldn't call it misspelled. It would merely label it as in another language, and let you determine whether it was a misspelling or deliberate.