Auto correct and no apostrophes for contractions

Hi all,

I just carried out a test to see the auto correct, as I am used to typing fast without adding apostrophes most of the time and let my software add them in, but Scriv 3 is having a hard time.

The expected result is that Scriv 3 should add the missing apostrophes, but instead, it auto corrects the words to something completely different, as shown below.

wasnt - wast
youre - your
dont - font
didnt - dint
Ive - vie
hed - hes
itd - its
weve wee
youd - yous
whos - whews
itll - till
aint - aunt
whod - dhow
theyre - there

There may be more, but these are just some of the ones I have discovered.

Just add the custom corrections you want and Scrivener will insert them.

I can do that, but that doesn’t address the issue. I don’t have to do that in Scrivener 1.9.16, which auto corrects the words.

Perhaps it’s only an issue for a small number of users? I have never typed like that, so never needed any of these corrections. However, if anyone wants them, they’re easy to add. I’m just suggesting that it may not be a bug, just L&L deciding to do something differently.

I doubt it’s a bug. Scrivener’s auto-correct system has always been asinine. I turned it off years ago because I got tired of having my typos “corrected” to the least likely possibility, and even to non-words. Accidentally typing han instead of hand got me Han. Leaving out the space in onthe resulted not in on the or even just a red underline, but Anthe. It’s not surprising that the OP got dhow out of whod. The most obvious correction is who and even the hoped-for who’d is more likely, so of course Scrivener auto-corrects it to dhow.

I had hoped that v3 would show improvement in this area, but no such luck.

I believe that the dictionaries, which presumably include the basic autocorrect entries, are all third-party (from the open source project Hunspell, I believe). Obviously, on the Mac (where Scrivener began), the OS includes system-level dictionaries; supporting dozens of languages on Windows is more challenging – a small team like L&L could hardly develop their own (so I felt that “asinine” was a little bit harsh).

However, as Hunspell is open source, I imagine there’s some way for users to contribute to improving their dictionaries, etc. You could try checking http://hunspell.github.io/ as a starting point, if you’re interested.