[Tex2002ans]

Originally Posted by Hitch

The line-height and the vertical align won't really work for Amazon (yes, I know that this is the ePUB sub-forum). and I think v-align is dicey at best. We've tried various solutions like that, with varying degrees of success.

Yep, I was going to comment the same thing. That solution won't work across many devices. At best, it would degrade into a small font footnote on the same line, at worst, it would cause some wonky madness with the line. I just settle on hideous in-line bracketed footnotes. Those won't break anywhere and are easy to read/click.

Originally Posted by Hitch

Vis-a-vis the other question, which isn't a bad one (I've asked it myself)..."how do it know?" That's an interesting question, isn't it? How does the device (say, a PPW) distinguish between the links in the TOC, versus footnote lines? My answer is, PFM, which means Pure Freaking Magic. Whatever the distinction is, it certainly seems to work. TOC items don't "pop up" as footnotes do.

Probably taking into account the Metadata (nearly everyone marks the TOC as "Table of Contents", it can probably know that links in there ARE NOT footnotes).

Then lets say it is NOT marked as a "Table of Contents" via metadata, but the thing is about >90% text inside of <a> tags due to heuristics... you probably know it should be treated a little differently.

I don't have a device that has this "pop-up footnote" functionality, but it would be interesting what it does on some of these books with linked Indexes.

Then for the actual footnotes, I am assuming some sort of heuristics approach which takes into account the length of the text in the footnote + the look of where the link leads to.

For example:

<p><a href="../Text/Chapter01.xhtml#fn1">[1]</a></p>

RED is likely not going to be very long at all (maybe 6 characters MAX) (Perhaps they also take into account superscript within a link as HIGHLY LIKELY it is a footnote?)

Since BLUE is also pointing to some sort of id, they will probably make an ok assumption that it is a footnote? Perhaps they might discount if the id uses the word "page" in there (so maybe an index might be excluded if it is pointing to #page123").

On the receiving end, it might use heuristics to tell

<p><a href="../Text/Chapter01.xhtml#fn1">[1]</a> Text Text Text.</p>

<p><a href="../Text/Chapter01.xhtml#fn2">[2]</a> Text Text Text.</p>

Or:

<p><a href="../Text/Chapter01.xhtml#fn1">[1]</a></p>

<p>Text Text Text.</p>

<p><a href="../Text/Chapter01.xhtml#fn2">[2]</a></p>

<p>Text Text Text.</p>

They could look at where the links are located (towards the very bottom of the HTML file? In its own separate HTML file?)

Anyway, Amazon has a fracking MASSIVE amount of books to sample. So they have A TON more data on what is footnotes and what isn't footnotes. Above is just how I would tackle it if I was coding it, and it would probably work pretty decently from the books I have seen.