View Single Post
Old 04-06-2016, 09:26 PM   #201
JSWolf
Resident Curmudgeon
JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.JSWolf ought to be getting tired of karma fortunes by now.
 
JSWolf's Avatar
 
Posts: 79,970
Karma: 147448039
Join Date: Nov 2006
Location: Roslindale, Massachusetts
Device: Kobo Libra 2, Kobo Aura H2O, PRS-650, PRS-T1, nook STR, PW3
Quote:
Originally Posted by Tex2002ans View Post
I thought I was clearer with a few examples at the very end of that last post under "Pitfall of Byte Counts", but perhaps not.

Let me add one more potential case of the actual backend code of the book changing:

A lot of the ebooks that I work on is cleaning up crappily converted EPUBs. I will go through and fix OCR errors, correct footnotes, change hideous JPGs of Greek letters to actual Unicode Greek characters, digitize formulas, change images of Tables -> HTML versions, clean the code itself in the backend, [...].

Then we (as publishers) rerelease an updated "version 2.0" of the ebook.

Depending on how extensive the code fixes are, you can imagine that this could drastically change the size of the HTML files (and would throw off the Byte Method page numbering).

As a real life example, back in 2013 I worked on The Great Austrian Economists by Randall Holcolmbe (Before/After pictures + some discussion posted here):

https://www.mobileread.com/forums/sho...06#post2672206

Here is a single set of example images if you are too lazy to click on the link:

Original PDF Scan: https://www.mobileread.com/forums/att...0&d=1383241259
Old EPUB: https://www.mobileread.com/forums/att...1&d=1383241259
New EPUB: https://www.mobileread.com/forums/att...5&d=1383241264

I just took a look at one of the largest EPUB redo projects I ever handled, and the ADE "pages" went from 3627 "pages" (Before) -> 3614 "pages" (After), just from the sheer amount of code cleanup + corrections. I also ran it through KindleGen and it went from 73190 "locations" (Before) -> 71266 "locations" (After).

I also tested the first version of A Dance With Dragons that I purchased from B&N, and it was 1100 ADE "pages", the later version was 1101 ADE "pages". Most of what changed was minor typo corrections + different code for lists (went from <li> to using <div>).

You can probably extrapolate the hideous InDesign/Quark code cleanup (I don't feel like hunting down one of those books right now), but you can see from my Example #1-#3 above how the Kindle "locations" could easily be thrown off (and ADE's algorithm too, although not as wildly).
The original ePub is still the same. It hasn't changed. ADE's page numbers are still the same. The reformatted edition is a different version/edition. It's not the same file. The original ePub still has not changed anything. The new ePub is a different file. So yes, the ADE page number is going to be different. But then you have two different files, not one that's just suddenly changed. If I change an ePub, then yes, ADE page number will change. But after I finish, it's no longer the same as it was. The original file still uses the same ADE page numbers. So no, byte count page numbering works just fine.

Taking your example, if you had a pBook that was much reformatted as the ePub was, I would expect the page number to change in the newer edition.



Quote:
In the future, maybe a lot of other ebooks will be changing images of formulas to MathML... that is also going to throw off the Byte Method of page numbers.
Not until more Readers are using a version of RMDSK that can handle MathML.
JSWolf is offline   Reply With Quote