[GUI Plugin] Count Pages

[davidfor]

Originally Posted by JSWolf

I used Calibre to convert to text. I loaded the text file into Notepad++ and used the Word Count function of the TextFX plugin and I get a count of 94750.

I will leave it up to you to say if this difference is enough to warrant any more changes.

So, now we have three different counts from three different algorithms or applications. Is there any reason why you think the Notepad++ TextFX plugin is the correct one? Can you point to where it defines what a word is? I can't find anything and I have no reason to think that it is more correct than the two built into calibre.

If you want an example, than look at the post YOU made last year that started all this. Your complaint was about the following being counted wrong:

except…if
except—if
except–if

You stated they should be six words, not three. I have just tested with all three algorithms; ICU, older count pages and TextFX. Only the ICU counts them as separate words.

So, are you going back on your original claim and you think these should be considered as one word?

And I can't work with "a book I have has a problem". I need to see the book and know what the problem is. Or a chunk of text and a description of how it is being counted wrong.

[Katsunami]

Does stuff like this really matter?

Does it matter if an algorithm is off by +/- 1000 words or so in a 100.000 word novel? Does it even matter if the page approximation is +/- 10% off as compared to a paper version?

I've been using 2100 characters a page for a long time; 2400 before that, but I found books were too 'short' compared to paper ones, so I adjusted this. As all books are counted the same, one can get a very clear indication of how 'long' a book is, and 2100 chars/page seems to be close enough to the paperbacks I've owned in the past, so I can also approximate how 'thick' a book is.

Personally, I like the count pages plugin fine as it is

Now that I have an original cover for the KA1 that actually works, there's a fair chance I'll be switching to the KA1 as my main reader. I've put the count pages count into the book's title, while the reader counts on its own. I'm now reading a book which Count Pages sets at 302, with 2100 chars/page, while the reader sets it as 341 pages (screen turns, it seems, as the page number increases as I turn a page).

Good enough for me. I can see how far in the book I am on the KA1, in the form of "page x from 341", while the Kindle will show "page x" (from 302, but it doesn't state that), and I can see my custom page count (302) to compare this book to others, even across readers.

[Terisa de morgan]

Originally Posted by Katsunami

Now that I have an original cover for the KA1 that actually works, there's a fair chance I'll be switching to the KA1 as my main reader.

Congrats! Glad to know the cover works.

Originally Posted by Katsunami

I've put the count pages count into the book's title, while the reader counts on its own. I'm now reading a book which Count Pages sets at 302, with 2100 chars/page, while the reader sets it as 341 pages (screen turns, it seems, as the page number increases as I turn a page).

That's my method too.

[BetterRed]

Originally Posted by Katsunami

Does stuff like this really matter?

It matters to angels who dance on pins, or daemons who take time outs to lunch.

I maybe the only recent poster to this thread who has ever had a valid reason to care about word count accuracy. Aeons ago I was a freelance tech-writer who was a paid a pittance per 100 words for writing in 2B pencil on yellow foolscap. The copy typists (as opposed to the MS typists) did the word counting on the galleys, they omitted articles, prepositions and conjunctions... pronouns maybe. Yes, typists who could count, and knew the basics of English grammar - some could even conjugate and decline Latin verbs!

Not only are there no more copy typists, today's equivalents think conjugation is something that one does in bed and declination is when you fall down the stairs in the morning.

Interestingly some of the copy typists were male, but all the manuscript typists were women.

BR

[Katsunami]

Originally Posted by Terisa de morgan

Congrats! Glad to know the cover works.

Thanks

[davidfor]

Originally Posted by Katsunami

Does stuff like this really matter?

Does it matter if an algorithm is off by +/- 1000 words or so in a 100.000 word novel? Does it even matter if the page approximation is +/- 10% off as compared to a paper version?

If you read back, that's what I've been saying. My original change to use the ICU method was as much my curiosity as anything. But, using it also makes it consistent with the calibre editor. At least for how it counts if not for what it counts. And should handle non-English books better. But, if someone can demonstrate that one of the algorithms is completely wrong, I'll remove it.

The alternate reason for the count to be out is a bug in how the book is processed. If a file was missed or partly counted or counted twice, then that is something I need to look at and fix.

[davidfor]

It's been pointed out to me elsewhere that the Count Pages plugin is calculating something called the "Gunning Fox Index". Is that how many times a fox shoots a hunter in the book? Or maybe it's supposed to be the "Cunning Fox Index". Which might be counting the same thing.

[theducks]

Originally Posted by Katsunami

Does stuff like this really matter?



Personally, I like the count pages plugin fine as it is .

#1 Consistency (within my usage) is what is really important. it is just a scale to judge other books against.

I use it to replace paper book heft

All the nay Sayers.
Paper was rift with heft cheats

• bigger margins
• larger line spacing
• blank pages (so chapter start on even pages only )

[JSWolf]

Originally Posted by davidfor

So, now we have three different counts from three different algorithms or applications. Is there any reason why you think the Notepad++ TextFX plugin is the correct one? Can you point to where it defines what a word is? I can't find anything and I have no reason to think that it is more correct than the two built into calibre.

If you want an example, than look at the post YOU made last year that started all this. Your complaint was about the following being counted wrong:



You stated they should be six words, not three. I have just tested with all three algorithms; ICU, older count pages and TextFX. Only the ICU counts them as separate words.

So, are you going back on your original claim and you think these should be considered as one word?

And I can't work with "a book I have has a problem". I need to see the book and know what the problem is. Or a chunk of text and a description of how it is being counted wrong.

Ignore the count from Notepad++. The Word Count doesn't work properly. It takes word—word as one word and it takes word — word as three words. But given that the file I used had no spaces around the em dashes, that means the count for NotePad++ is off on the low side. If this isn't a problem, just Let It Be.

[davidfor]

Originally Posted by JSWolf

Ignore the count from Notepad++. The Word Count doesn't work properly. It takes word—word as one word and it takes word — word as three words. But given that the file I used had no spaces around the em dashes, that means the count for NotePad++ is off on the low side. If this isn't a problem, just Let It Be.

Jon: YOU are the one making the fuss. YOU are the ONLY one reporting problems with the word counts. And YOU are reporting them as MAJOR problems. And YOU are the one who brought Notepad++ into it without even checking that its count satisfied YOUR conditions. All this time I have been asking YOU to demonstrate HOW the counts are wrong so that can be dealt with. As I have already stated, I consider the statistics to all be approximations so I have been willing to Let It Be for a long time.

[BetterRed]

Moderator Notice
Put up or Shut up Jon

Better yet, waste your time by hacking the PI to implement 5 word count algo's for each of 10 languages, and post a spreadsheet of 1,000 books for each language showing the 50,000 counts.

You get to chose the algos, lingos, and books

[PeterT]

Originally Posted by davidfor

It's been pointed out to me elsewhere that the Count Pages plugin is calculating something called the "Gunning Fox Index". Is that how many times a fox shoots a hunter in the book? Or maybe it's supposed to be the "Cunning Fox Index". Which might be counting the same thing.

Gunning Fog Index
https://en.wikipedia.org/wiki/Gunning_fog_index

[BetterRed]

Fog in Gunning early to tomorrow morning, according to this.

BR

[Dirk-71]

Here is the latest Beta incl. German translation.

[JimmXinu]

Originally Posted by Dirk-71

Here is the latest Beta incl. German translation.

I've tested this a bit and it appears to work fine for me on Win10. (Although in truth, I only compared numbers for word count because I don't keep page counts or readability indices. Other counts didn't crash.) ICU word count is ~2% less than previous, but I consider that within tolerance.

My German isn't good enough to confirm the translations, but I can confirm they do show up.

The label still says Fox--not surprising, davidfor hasn't had another version since mentioning it.

This particular zip file, while functional, shouldn't be the new released version due to the presence of the Mac specific __MACOSX dir.