Sage Kobo (still) can't count?

[davidfor]

Originally Posted by Uncle Robin

I mentioned that fact 2 different renderers offered very similar progress percentages was evidence (imo, ofc) that they were more reliable than the "% figure shown on the sleeping display (and in the book stats under "Activity")

True, but I was also making it clear what I was talking about.

I have been further intrigued to discover an unexpected trend - the gap is getting BIGGER the further I go into the book. I had thought that as I progressed through the book, the 2 divergent stats would move toward each other, but so far, the opposite is true.

When I took those screenshots, the progress bar said 14 % and the "book %" said 10%. Then, in one of my replies above, it was 21% (22% in the ADE epub) versus 16%. Now the progress bar says 27.5% while the "book %" says 21%. If this trend continues, I'll be halfway through my 2nd read before reaching 100% on the sleep cover

Nowhere in you screenshots do I see 14%. I can see the sleep screen says 10% and the page progress says "Page 224 of 1593". That does work out to about 14%. Are you saying that you had "Percent of book read" at the bottom of the screen displayed and it was 14% at this point? In the other posts are you calculating the percentage from the page numbers or this is what the device shows?

My point was that the percent read for kepubs is not based on the pages. Or screens or whatever you want to call it. It is based on something else. For a book with simple coding, it is probably fairly close. If one chapter had a lot more complex code in it, then it will be different. If there are images, or lots of vertical spacing, then it will also affect things. Or something stupid like a chapter with 1000 words but each in separate paragraph will be different than 1000 words in a single paragraph.

[Uncle Robin]

Originally Posted by davidfor

True, but I was also making it clear what I was talking about.


Nowhere in you screenshots do I see 14%. I can see the sleep screen says 10% and the page progress says "Page 224 of 1593". That does work out to about 14%. Are you saying that you had "Percent of book read" at the bottom of the screen displayed and it was 14% at this point? In the other posts are you calculating the percentage from the page numbers or this is what the device shows?

My point was that the percent read for kepubs is not based on the pages. Or screens or whatever you want to call it. It is based on something else. For a book with simple coding, it is probably fairly close. If one chapter had a lot more complex code in it, then it will be different. If there are images, or lots of vertical spacing, then it will also affect things. Or something stupid like a chapter with 1000 words but each in separate paragraph will be different than 1000 words in a single paragraph.

You're right - the percentages (other than the one on the front when sleeping) are calculated from the progress markers. The book has (so far) no images at all, no unusual vertical spacing. Every page looks very much like the sample in the attached screenshot.

I have just discovered that the ADE version has a clickable "page list" listing all 726 pages of the print edition. I'm currently at the start of chapter 31/55 which in the print edition is page 347/726 - 47.8% of the way through the book. Which makes a nonsense of ALL the other stats, I guess.

[Uncle Robin]

Having finished chspter 53, with 2 chapters left to read, when I put my Kobo to sleep I am apparently 44% of the way through the book - The page list for the print version says I'm at 82%.

This has been a fun and interesting exercise in trying to figure out the mystery of the maths, but when it comes to the Kobo stats I know neither whodunnit nor how. As shown above, it's not a complex book in terms of layout, all text, and lots of it. It will be interesting to try a de-drmd epub version in KOReader to see how its measurement of progress stats compare.

[Guego]

Originally Posted by Uncle Robin

Having finished chspter 53, with 2 chapters left to read, when I put my Kobo to sleep I am apparently 44% of the way through the book - The page list for the print version says I'm at 82%.

Iirc, the percentage is where you are in the file, not in the text. For example, if your last chapter is full of hyperlinks, formatting instructions, etc, it will take a large part of your epub file even if there is very little text in it. So you can be at 90% of the pages, but 50% of the file.

[milady133]

Originally Posted by Uncle Robin

As shown above, it's not a complex book in terms of layout, all text, and lots of it.

Just from your screenshot, I think it might be more complex than what you imply. Your page just shows a lot of hyperlinks (footnotes) in just one page. That possibly will break havoc in the calculations for the stats, depending on how they are internally stored:
- A file for each footnote?
- One dedicated file for all of them?
- The footnotes stored at the end of the chapter file?
...
Just my two cents

[davidfor]

Originally Posted by milady133

Just from your screenshot, I think it might be more complex than what you imply. Your page just shows a lot of hyperlinks (footnotes) in just one page. That possibly will break havoc in the calculations for the stats, depending on how they are internally stored:
- A file for each footnote?
- One dedicated file for all of them?
- The footnotes stored at the end of the chapter file?
...
Just my two cents

It is more complex that what @Uncle Robin implies. And he knows that. But, before you suggest some things, make sure you actually read the all the thread. Especially posts that explain something about how the calculation is done.

And:

• Links do not have an affect. Except that one part of the calculation uses the file size. And if there are links with hrefs, that increases the file size without increasing the displayed text.
• The calculation takes into account each file and anything in it, whether it is a footnote or note.
• If the footnotes are at the end of the chapter, then is just normal text for reading and progress calculations. To page to the next chapter, you must go through these footnotes. If you navigate to it, the page position and percent read should update. If, as is common, the footnotes are in a smaller font, it will reduce the total pages but not the percent read as you go through it.

And two things I didn't think of mentioning in my explanation above:

• The internal files are referenced in the manifest, spine and ToC. The latter two have a subset of the first. I don't know what happens if there is a file that does not have a ToC entry or for a file in the spine with attribute 'spine="no"'.
• There is an option to ignore end matter from the reading progress. I don't know how that works. I don't know if it affect the page count, the percent read or the time to read. Or how the end matter is determined. Though the latter probably requires the book to be ePub 3 as that does have markup for sections of the book.

[Uncle Robin]

Allegations to the contrary nothwithstanding, I was not being maliciously or deliberately obtuse. I know nothing at all about how ebooks are structured as files.

What I DO know is that the stats presented said "44% READ" (e.a.) when that was patently not true. As an indicator of the percentage of the book actually READ (the information it was allegedly presenting), the figure was meaningless to the point of absurdity. THAT is simple fact, regardless of the mechanics of file size, internal structure etc.
'
I've just checked a de-drmd epub in KOReader and its stats match those of the kepubs internal progress bar and the ADE epub - wrong, but not AS wrong as the "% read" figure shown on the 'cover' when sleeping or in the statistics under "Activity". The end matter is 15% of the entire book, not 50%.

[davidfor]

Originally Posted by Uncle Robin

Allegations to the contrary nothwithstanding, I was not being maliciously or deliberately obtuse. I know nothing at all about how ebooks are structured as files.

Sorry, I wasn't meaning that. I probably should have said "originally suggested".

What I DO know is that the stats presented said "44% READ" (e.a.) when that was patently not true. As an indicator of the percentage of the book actually READ (the information it was allegedly presenting), the figure was meaningless to the point of absurdity. THAT is simple fact, regardless of the mechanics of file size, internal structure etc.
'
I've just checked a de-drmd epub in KOReader and its stats match those of the kepubs internal progress bar and the ADE epub - wrong, but not AS wrong as the "% read" figure shown on the 'cover' when sleeping or in the statistics under "Activity". The end matter is 15% of the entire book, not 50%.

That probably means that KOReader has implemented the Adobe method.

The problem is that it isn't hard to construct a book that screws with any of these progress calculations. I've always found a certain persons insistence on the Adobe page counting algorithm as it is extremely easy to mess with (just use a lot of repetition as it is based on the compressed size or change the compression algorithm for one file, but that's cheating). For the kepub method, the percent read is a mix of file size and word count, so messing with that is just as easy. And if there are images, they are not counted unless it is the only thing in the internal file. Throwing in some CSS to put vertical whitespace in some places will mess with it as well.

Personally, I do not remember seeing the percent read out by anywhere near as much as you are reporting. If I calculate it from the pages, then I'm usually within a percent of what the device shows. But, I am reading Any chance of seeing the book that is doing this?

[Barty]

I believe for epub, the number of "pages" is simply the total # of bytes of the (x)html files divided by an arbitrary number, like 1024 or 2048 bytes per page

[davidfor]

Originally Posted by Barty

I believe for epub, the number of "pages" is simply the total # of bytes of the (x)html files divided by an arbitrary number, like 1024 or 2048 bytes per page

Epubs page counting is the Adobe algorithm. It is based on the compressed file size. See https://wiki.mobileread.com/wiki/ADE#Page_numbers.

[John F]

Originally Posted by Barty

I believe for epub, the number of "pages" is simply the total # of bytes of the (x)html files divided by an arbitrary number, like 1024 or 2048 bytes per page

What about images? *

* I don't feel like reading through David's link.

[DNSB]

Originally Posted by John F

What about images? *

* I don't feel like reading through David's link.

For Adobe synthetic page number algorithm, images that are not stored in the (x)html file do not add to much to the page count. This leads to some oddities in the page number shown since the image size is not taken into account.

Something I've seen rarely is images that were stored in the (x)html files using base64 encoding which lead to odd page counts. I've attached a sample epub using base64 which shows an Adobe synthetic page count of 109.