Doesn't save changes

[mrmikel]

Thanks.

[BetterRed]

Originally Posted by mrmikel

Maybe a keyboard shortcut for add word to dictionary would be helpful. Go through the list and hit a key and then only the bad stuff would be left.

@mrmikel - Maybe I'm misunderstanding but, Alt/d adds current word in list to user dictionary, and Alt/i adds current word in list to ignore list.

Also words in the list have a context menu, in there you can select which user dictionary, good if you want domain specific dictionaries - maritime, geoscience etc.

BR

[mrmikel]

I didn't know that, so I may make use of that. I don't care whether they are in the user dictionary or main, just so long as they are eliminated from the list.

I tried it and it works fine. For information of others, it works its way up the list, so you want to start at the bottom, given a choice.

[AnotherCat]

Originally Posted by mrmikel

...What you are working on is probably pretty decent text, but what some of us work on is OCRed junk, with hundreds of problems, so putting it off just means you lose your place.

In those cases with many errors I, for myself, find it much the easiest and quickest to run through the book with "Show misspelled words underlined in the code view" turned on rather than use the spellchecker window. A right click on the underlined word brings up suggested correct spellings for left click overwriting of the incorrect one, and then F8 (or selection from the right click context menu) will take one to the next spelling mistake; but if it is a case of a book stricken with many errors visible in one CV page I scroll manually to the next scanning visually for homophones and other trickies on the way (something one has to do anyway, sooner or later).

This is even my approach when using MSWord where one can edit in and out of the spell checker window with no issues at all; even there I find it very much quicker to run through an error strewn document via the squiggly red underlines than to traipse through using the spellchecker window.

In saying the above I recognise that everyone finds their own convenient zone for such things so only mention it in case it is of some use. Otherwise just regard or ignore it as being just information as to what someone who has been writing and editing documents on PCs for over 30 years now has found most convenient for himself, but has probably missed some shortcuts and developed bad habits along the way too.

[BetterRed]

Originally Posted by AnotherCat

and then F8 (or selection from the right click context menu) will take one to the next spelling mistake;

But if I then call up the context menu with the menu key I don't get the spelling error context menu, I get the 'vanilla' one that offers copy, paste etc, but not context menu with suggested words etc.

I think the 'problem' is that F8 also selects the current XHTML start/end tags and not just the misspelt word.

Is there a way of stopping that? If not could F8 be made to work as F4 does in Sigil?

FWIW - I'd like to have an editing view where all the XHTML tags are hidden so that I can focus on content - not a wysiwyg view, a plain text draft view, i.e a Joe Friday view

BR

[AnotherCat]

Originally Posted by BetterRed

But if I then call up the context menu with the menu key I don't get the spelling error context menu, I get the 'vanilla' one that offers copy, paste etc, but not context menu with suggested words etc.

I think the 'problem' is that F8 also selects the current XHTML start/end tags and not just the misspelt word.

Is there a way of stopping that? If not could F8 be made to work as F4 does in Sigil?

FWIW - I'd like to have an editing view where all the XHTML tags are hidden so that I can focus on content - not a wysiwyg view, a plain text draft view, i.e a Joe Friday view

BR

I may be misunderstanding you BR but if I go from error to error with the F8 key I still get the context menu and spelling suggestions with a right click of the mouse, and it ignores spelling in the HTML.

There may be a keyboard shortcut to the context menu which is what you are referring to and which goes astray as you describe, but that is outside my experience (except for the few widely used exceptions, cut/copy/pastes, F5, F3, etc. I do the mouse bit).

Apologies if I have misunderstood.

I would find hiding the HTML markup during searches to be very helpful (one can use the Preview to search in but means extra effort reverting to the CV when a correction is to be made; in saying that I am not promoting editing in the preview ).

[BetterRed]

Originally Posted by AnotherCat

I may be misunderstanding you BR but if I go from error to error with the F8 key I still get the context menu and spelling suggestions with a right click of the mouse, and it ignores spelling in the HTML.

Apologies if I have misunderstood.

What I wrote was -

Originally Posted by BetterRed

But if I then call up the context menu with the menu key I don't get the spelling error

My keyboards have a key immediately left of the right Ctrl key labelled Menu with a graphic that (imagination permitting) looks like context menu - I call it the menu key. I find using it faster and much more accurate than using a mouse. To make it even easier I reconfigured F8 in calibre editor and F4 in Sigil to Ctrl/Shift/Down arrow

See attachments for difference between pressing the menu key in calibre & sigil after pressing F8/F4 to go to next misspelt word

Observe that sigil treats the word 'bank-ruptcies' as one word whereas calibre treats it as two. This presents a problem with valid hyphenated compound words such as helter-skelter, helter is not a 'valid' word by itself, but when hyphenated with skelter it is.

Also note that had that been 'bank-ruptures', then neither calibre or sigil would have picked up the error. But I would have found it via Sigils ability to list hyphenated words.

BR

[AnotherCat]

Ahhh, the Menu or Application key - the least worn key on keyboards, I wasn't aware that anyone used it so it didn't click . As I said I am not a shortcut key user. It does not as a matter of course mimic the mouse right click context menu so its behaviour may be different.

Regarding hyphens I am aware that Sigil and Editor spell checks work differently for hyphenated words. Given all the trickies such as picking up words that are not hyphenated but should be, I, in editors, normally do a search on "-" and assess each case (a search I find I normally need to do anyway to rationalise "-"s, ems and ens) rather than use the spellchecker.

As an example of trickies, and taking your example of the word "helter", is that helter is in fact a valid word by itself which will be come across if editing old text (it is an old form of "halter"). When editing text (such as OCR'ed) there are lots of others similar such as old spellings and hyphenations of words and the author's own rendering as was originally published, and so when doing hyphen checks I also have the big OED running as well. Also I am often unclear as to what are the usages of hyphens for a particular even modern word combination and spellcheckers just give some of the alternatives in a list, whereas the OED descriptions and etymology allows judgement as to the best choice to be applied according to the usage in the particular prose.

So, for example, as can be seen above and in other posts, I have a habit of using the word spellchecker when writing informally, whereas the OED tells me that usage is "spell-check" and "spell checker". I don't use Sigil anymore but if one tries putting "spellcheck" "spell checker" "spell-checker" "spell-check" into a MSWord (2007 in my case) document one will find that it only finds, apparently incorrectly, one of those words to be incorrect. In my experience most spellcheckers will allow almost any hyphenation of two valid words as being correct. For example MSWord (2007) accepts "tall-short" and "green-sky" as being valid, the OED does not, however there is no reason why an author could not just invent and use them in their prose (perhaps to come into later common usage) in which case the ebook should follow that.

I think perhaps where I am differing is that I see the finding of the likes of hyphenated words as being trivial (after all, even a simple search for "-" finds them most reliably no matter what context they are used in) and so really don't care how a spellchecker does it. But the identification of whether the word should be hyphenated or not, or even concatenated in the particular context of the prose is not trivial and is something that no spellchecker can solve without being a full blown dictionary (such as the OED); and finding words that are not hyphenated, but should be, is another challenge (and so is finding the likes of homophones).

Note, I am not averse to, nor am I intending to denigrate, any effort towards improving Editor's or any other spellchecker, nor criticising the methods of other users, but am rather just stating how I go about some of the matters that have arisen or matters I encounter which complicates and reduces their competence when using them.

Whew, I reckon I need a tinny after that. You might too .

[BetterRed]

Originally Posted by AnotherCat

Ahhh, the Menu or Application key - the least worn key on keyboards, I wasn't aware that anyone used it so it didn't click . As I said I am not a shortcut key user. It does not as a matter of course mimic the mouse right click context menu so its behaviour may be different.

As far as I'm concerned the menu key (which I use where ever possible) always works exactly the same as a right mouse button click. Elsewhere in calibre, eg the book list, metadata editor, comments editor, the tag browser etc. And in windows, my file manager, imaging editing tools, text editor etc. And in Sigil's code view window after F4 (Next spelling error), and Notepad++'s editing window after Alt/N (Next spelling error).

The only instance I can think of where it doesn't work like a right button mouse click is after F8 (Next spelling error) in the calibre-editor code view window.

BR

[kovidgoyal]

https://github.com/kovidgoyal/calibr...1638abce0c6751

[AnotherCat]

Originally Posted by BetterRed

As far as I'm concerned the menu key (which I use where ever possible) always works exactly the same as a right mouse button click. Elsewhere in calibre, eg the book list, metadata editor, comments editor, the tag browser etc. And in windows, my file manager, imaging editing tools, text editor etc. And in Sigil's code view window after F4 (Next spelling error), and Notepad++'s editing window after Alt/N (Next spelling error).

The only instance I can think of where it doesn't work like a right button mouse click is after F8 (Next spelling error) in the calibre-editor code view window.

I see it is changed now .

Yes, perhaps my "It does not as a matter of course mimic the mouse right click context menu so its behaviour may be different" was a bit short.

What I was alluding to is the context menu that appears is dependant on the message or call that is used (in the Windows API case) so they are not naturally going to be the same. It was just a remark and was not a judgement as to what was correct or most desirable (although the Windows GUI recommendation is that they be the same, though I only think that is so because I am no programmer). Nor was I judging nor lobbying as to how it should work in Editor as because of my approach it is irrelevant to me.

On checking, under the F8 circumstance that you outlined in Editor, I see that SHIFT-F10 does not bring up an extended context menu either. So Windows keyboard maestros may find that a bit of an inconvenience too.

Regarding the practice in other programs, I have checked and I do have some other non Windows GUI applications where the mouse and Application/Menu key context menus are different, and even at least one that the context menus do not exist in either case at all (that one being the OED Dictionary). However, I think for most, if not all, Windows GUI applications I regularly use (Office Suite, etc., etc., etc.) the context menus are probably the same for both mouse and key.

John

[BetterRed]

Originally Posted by kovidgoyal

https://github.com/kovidgoyal/calibr...1638abce0c6751

BR

[Tex2002ans]

Originally Posted by AnotherCat

I think perhaps where I am differing is that I see the finding of the likes of hyphenated words as being trivial (after all, even a simple search for "-" finds them most reliably no matter what context they are used in) and so really don't care how a spellchecker does it.

I don't have any stats on hand from before I requested the Spell Check Tool be added into Sigil, but I spent HOURS doing it the manual method, and it was BRUTAL.

It just took gods damn forever, and it is nowhere near as accurate. (After seeing the same words again and again, you fall into this monotonous brain-dead/vegetative state after doing it for so many hours straight. ).

Doing it the manual way, there is no way to "ignore" a word that you already know is correctly hyphenated.

If "long-term" shows up 183 times, you had to Next 183 times.
If "short-term" shows up 111 times, you had to Next 111 times.
If "Irvington-on-Hudson" shows up 34 times, you had to Next 68 times.

Now, since all unique words are shoved into the list ONCE, this really saves the amount of time your eyeballs have to work + how many times you have to ignore/fix mistakes.

Just those three words, you have wittled down 362 clicks into a quick look at a list.

With the Spell Check List, I typically just do two passes. (One with the "Show Only Misspelled Words" checkbox on, one with it off). You might also want to fiddle around with the Frequency sorting, because just visualizing the data in a different way helps you sort through and catch things much faster (and errors you may have previously glanced over).

For example, many of these hyphen -> em dash errors only occur once, so you can focus more on the words in the list with <3. Something that occurs ~>8 times in a book is probably not a typo, and can just be quickly looked over.

Side Note: This latest journal I am working on digitizing (~4400 pages, ~2 million words), there were ~18800 hyphens before -> 18428 after fixing (this means ~2% of the hyphens were a mistake.)

That would have taken fracking FOREVER to do one-by-one (it already took me 12 hours to do it the Spell Check List way (including all the time double-checking/fixing the source material, plus doing some code cleanup + other spelling corrections). ).

It would be QUITE interesting to gather these stats in the future. How many words have hyphens compared to total words, how many hyphens were wrong (and what %)... I will be sure to keep it in mind in the future. Maybe add it to all my fancy stats I am gathering (Preview here):

https://www.mobileread.com/forums/sho...8&postcount=52

Another thing to toss on my "growing pile of things I should do but don't have the time because I have to digitize more stuff" list.

Originally Posted by AnotherCat

Note, I am not averse to, nor am I intending to denigrate, any effort towards improving Editor's or any other spellchecker, nor criticising the methods of other users, but am rather just stating how I go about some of the matters that have arisen or matters I encounter which complicates and reduces their competence when using them.

Definitely recommend using the Spell Check Tool for hyphens.

Also for accented words + misspelled names + (Now with Calibre's Case Sensitive Search) OCR errors of letters<->numbers:

https://www.mobileread.com/forums/sho...51&postcount=6
https://www.mobileread.com/forums/sho...08&postcount=7

It is the greatest thing since sliced bread!

[mrmikel]

My sympathies on checking a 4400 page document.

I hope you added the correct ones to your dictionary, though it might be easy to wait until done and add the residue to the dictionary all at once. However, easy in this context is a relative term!

[phossler]

@kovidgoyal

It never occurred to me that people would want to both keep the spell check dialog open and manually edit files at the same time.

It might just be the way I'm used to doing things, but after SC finds what it thinks is mis-spelled (spelt?) I have to see it in context.

For ex SC might flag 'err' and suggest 'her'.

In one place 'err' might be correct but in another it might not: "ah, err, ummm -- give it to err (=her) to hold"

If I fix the first 'err' above by editing the html file with SC dialog open, will your change keep me out of trouble?