MobileRead Forums - View Single Post

Jaws · 03-13-2020, 10:33 PM

Milenski, it's also not a good idea to do this for four reasons related to publishing peculiarities:

(1) Not all publishers restart footnotes with each chapter, especially for multivolume works. (And in legal references, that still leads to footnote numbers in the thousands when they do!) So by default not doing this is the best policy.

(2) As Kovid notes, the concept of "chapter" is rather... undefined. For example, I recently struggled with a book from a continental European publisher (has a chesspiece logo) that splits its chapters into between one and eleven .xhtml files in the epub. Worse, it is inconsistent in whether the files and chapters even begin with an <hx> tag; some are <span>, some are even just a used-once-only <p class=y>! I can guarantee that converting this book successfully would require eyeball intervention anyway.

(3) A patch is almost certainly the wrong way to do it, because there are so many different ways to do the notes. For example, some older University of California Press (freely available so no issues with unauthorized copying) titles use an <ol></ol> list for notes, but others — even in the same subject area published in the same year — hard-code numbers, while yet others use a really kludgy JavaScript application. A patch could not handle all of those variants.

(4) Last, and far from least, is the dagger footnote. That is, not all footnotes are numbered in the first place! In particular, books that contain previously-published journal articles often have from one to four unnumbered footnotes at the beginning of each chapter, and as soon as one assumes it'll be the same number throughout the book Murphy will strike with a fatal exception error that screws up the rest of the conversion.

In sum, there are just some features that have been developed in books over the last few hundred years that are just going to require manual intervention. Modern footnotes (and even endnotes) is one of them; and be glad you're not trying to convert Swift's "The Battle of the Books" while maintaining any fidelity to the way it was published!

03-13-2020, 10:33 PM	#5
Jaws JCL Punch-Card Collector Posts: 80 Karma: 34468 Join Date: Jun 2014 Location: Antarctica Device: Aggressively Device Independent	Milenski, it's also not a good idea to do this for four reasons related to publishing peculiarities: (1) Not all publishers restart footnotes with each chapter, especially for multivolume works. (And in legal references, that still leads to footnote numbers in the thousands when they do!) So by default not doing this is the best policy. (2) As Kovid notes, the concept of "chapter" is rather... undefined. For example, I recently struggled with a book from a continental European publisher (has a chesspiece logo) that splits its chapters into between one and eleven .xhtml files in the epub. Worse, it is inconsistent in whether the files and chapters even begin with an <hx> tag; some are <span>, some are even just a used-once-only <p class=y>! I can guarantee that converting this book successfully would require eyeball intervention anyway. (3) A patch is almost certainly the wrong way to do it, because there are so many different ways to do the notes. For example, some older University of California Press (freely available so no issues with unauthorized copying) titles use an <ol></ol> list for notes, but others — even in the same subject area published in the same year — hard-code numbers, while yet others use a really kludgy JavaScript application. A patch could not handle all of those variants. (4) Last, and far from least, is the dagger footnote. That is, not all footnotes are numbered in the first place! In particular, books that contain previously-published journal articles often have from one to four unnumbered footnotes at the beginning of each chapter, and as soon as one assumes it'll be the same number throughout the book Murphy will strike with a fatal exception error that screws up the rest of the conversion. In sum, there are just some features that have been developed in books over the last few hundred years that are just going to require manual intervention. Modern footnotes (and even endnotes) is one of them; and be glad you're not trying to convert Swift's "The Battle of the Books" while maintaining any fidelity to the way it was published!