I’ve searched through the threads in this forum, and I don’t think anyone’s mentioned this before. This was a problem in Scrivener for Windows that I’d hoped would be resolved in this new version.
When using the contextual menu for a misspelled word or in the Spelling dialogue box, I get a list of mostly useless words, often including letter combinations that were NOT words. I don’t understand how a spell check engine can suggest nonsense words as possible correct words. (The word typed below is “Seh” which was supposed to be “She.” Which is not in the list below.)
This wouldn’t be as much of a problem if the word I wanted was at or near the top of the list choice, as seems to be the norm elsewhere. Not in Scrivener. I thought the list was alphabetical, but it’s not even following that logic. So I have to look through the list to find the word I wanted. Worse, in the context menu, the list offered is shorter, and my word may not even appear, so I have to use the Spelling from the menu. And then, Spell Check starts running - from the NEXT word. So I still have not gotten my word corrected. I have to stop it and place the cursor earlier.
Which means I might as well correct everything manually. (Don’t get me started on autocorrect. If you’ve found a way that we can choose which word a misspelled word is changed to, please let me know. In the last version I couldn’t, making it worse than useless, since it would make ridiculous changes, such as “adn” to “Adan” - what?). Please fix this. I love using Scrivener. But as it stands, Spell Check is useless in Scrivener.
No, sometimes it shows the right word, but it happens with enough words to be an issue, especially when those words are common or frequently misspelled.
My concern is that the older version (with rare exceptions) shows the proper correction on the top of the list, while the beta often shows a list with entirely useless suggestions. Don’t they use the same spellcheck engine?
Have in mind that v1.9 uses Aspell, and Scrivener v3 uses Hunspell, which has much richer dictionaries. Larger dictionaries, more proposal(not only the basic variants) are available to show. Still we do not use the biggest dictionaries available. I know there are even richer ones. In Scrivener v1.9 we got a lot of complains about the incomplete dictionaries, now we get complains about being too rich. Which one should it be?
You can always download smaller hunspell dictionaries and replace the ones bundled within Scrivener.
I also have to add that Hunspell and Aspell have different algorithms to provide proposals. This is what bothers you most in this case. Hopefully Hunspell will improve their proposals engine in the following versions. As much as I am aware Google Chrome is also using the Hunspell engine, so it should be constantly improving.
Looking further at this problem, I can confirm that this is a “well known” Hunspell problem. If you find some resources/tricks that will help improving the Hunspell suggestion list order, please let us know.
I was just going to chime in to say that when I’d been researching the differences between Aspell and Hunspell, this was a welll-known issue. There are a lot of pros that switching to Hunspell brings to the table, though.
Sorry to not check back in sooner! I gather from the thread, if I understand, you’re saying it’s the engine used that’s causing the problem with the order of words offered? Except this was a problem in the version I’m using (1.9), not just in the beta. Or are you saying the beta is using the old engine?
But that’s only one of the problems - if you check the screenshots, you’ll there are also suggestions that aren’t words. Why would those be suggested? If those were eliminated, the actual right word would be closer to the top. Half the problem solved.
I guess it would be a bit much to ask about the issue of losing the autocorrect option we had in a previous version? (Am I wrong about this? I vaguely remember it. I haven’t used the spell check in years.)
I thought I’d check in to say I do see an improvement in the spelling suggestions. The correct ones are more often at the top, and non-words aren’t making many appearances. So, yay!
So to my next question: autocorrect. I have a vague memory that in an old version we could adjust or add to the autocorrect list, but I could be wrong. There doesn’t seem to be a way to correct a spelling and make that correction automatic. The misspelling “somethign” does not autocorrect to “something” and there doesn’t seem to be a way to add that correction to the autocorrect list. So I have to manually correct it every time. Is there a reason this can’t happen? Or do I have to throw each file into Word to get that?
A couple suggestions: first, remove all entries containing spaces from the list, and second, alphabetize the list. Additionally, add a vertical scroll bar and key-input (to jump in the list) to the functionality.