observation about ADE page numbers

[Jellby]

Originally Posted by HarryT

There's a "1" in the margin, but a "5 of 24" on the status bar, so the "goto" is agreeing with the status bar numbers, not the margin numbers.

OK, that's a sensible solution.

[WillAdams]

FWIW, the old tradition of using roman numerals for frontmatter, then arabic for the balance was done so that new editions of a book could be issued w/o re-numbering the main text block if only a page or two was added to the frontmatter.

While the expense argument doesn't apply to ebooks, having page numbers be invariant between editions does have a certain appeal and could probably be handled better in the reader interface, say have a dialog which pops up when one begins to enter a page number:

FM ____ Text ____ Print edition ____

Where entering a folio in the first field would take one to the roman numeraled frontmatter sections, the middle to the text block, the last could be the pages as they are in the printed edition.

William

[frabjous]

Originally Posted by Jellby

That's similar to what Acrobat Reader does with PDFs: it will show "100 (107 of 234)".

Sort of. The difference is that the goto box in Acrobat Reader works the other way, which I find preferable. If I enter 107 into the Goto box, I'll end up at "107 (114 of 234)", which is far preferable than ending up at "100 (107 of 234)", which is how it works on my Sony with ePubs like this.

Edit: Ah, I see you've covered this.

Does anyone know offhand what Stanza does with these?

[HarryT]

Originally Posted by frabjous

Sort of. The difference is that the goto box in Acrobat Reader works the other way, which I find preferable. If I enter 107 into the Goto box, I'll end up at "107 (114 of 234)", which is far preferable than ending up at "100 (107 of 234)", which is how it works on my Sony with ePubs like this.

Does Acrobat Reader allow you to "goto" non-Arabic numerals, then? Eg, could you goto page "iv"?

[Jellby]

Originally Posted by HarryT

Does Acrobat Reader allow you to "goto" non-Arabic numerals, then? Eg, could you goto page "iv"?

I can't say for sure, but I seem to remember it does. You can also go to page "100 (107 of 234)" by writing the "(107)" inside the parentheses, so you can choose the page by absolute number or by label.

[Abecedary]

Originally Posted by HarryT

Does Acrobat Reader allow you to "goto" non-Arabic numerals, then? Eg, could you goto page "iv"?

Yes, Acrobat and the desktop ADE client both handle non-Arabic numerals in the same way. If you have the Heroic Measures PDF from the Oprah giveaway from last month, you can see how it works.

[DaleDe]

Thanks for the Pictures of Dorian Gray book. It is nicely formatted. As to the behavior of my EZ Reader. It displays each number on the upper right corner of the starting page. Since each number is a chapter and starts a new screen that is as it should be. It displays roman numerals and arabic numbers ok. the Cover screen does not display anything.

The page progress display shows x/24 where x is a number from 1 to 24. 1 is the cover page and 2 to 4 are the roman numerals (i to iii). The book proper starts on page 5. The goto command on this unit goes to the progress display numbers. That is, to go to the first page of the book you have to enter 5.

Dale

[HarryT]

Originally Posted by DaleDe

The page progress display shows x/24 where x is a number from 1 to 24. 1 is the cover page and 2 to 4 are the roman numerals (i to iii). The book proper starts on page 5. The goto command on this unit goes to the progress display numbers. That is, to go to the first page of the book you have to enter 5.

Dale

Same as the Opus - as one would expect, since both are using the same ADE reader.

[Jellby]

Originally Posted by DaleDe

Thanks for the Pictures of Dorian Gray book. It is nicely formatted.

I'm glad you liked it, but please download the one in ePUB Books forum, as it has a real title and not these silly pages

[DaleDe]

Originally Posted by Jellby

I'm glad you liked it, but please download the one in ePUB Books forum, as it has a real title and not these silly pages

Ah yes, but I appreciate the effort to make the one posted here to illustrate this discussion.

[frabjous]

Has anyone tried any of the following?

• Repeating page numbers/names?
• Having them go backwards?
• Using words rather than numerals of either sort?
• Linking to extra documents not in the manifest or toc?

(For the last couple, I imagine trying to create a kind of ADE ePub "Easter Egg" where the only way to access certain content in the book is to type "xanadu" in the "Go to" box or similar.)

Probably this is all asking for trouble.

My experiments suggest that it's a bad idea to skip certain subdocs from the page-map... I tried to avoid the roman numerals problem just by skipping those pages in the page-map, but on my Sony, it just made the front-matter inaccessible entirely.

[wallcraft]

Originally Posted by Jellby

Well, here is a test book (The Picture of Dorian Gray) modified with a simple page-map. I just make each chapter a page, and named them: (empty) the cover, i-iii the title, contents and preface, and 1-20 the chapters. So there are 24 pages; in ADE the Preface is "iii (4/24)" and Chapter 10 is "10 (14/24)" (you can edit the "iii" and the "10").

This basic approach, making each chapter (each item in the spine?) a "page" seems to me the best approach for those who want to get rid of the annoying page numbers over text. This is because a page number at the start of a chapter will rarely, if ever, cover up anything important.

If you don't want numbers over text at all, then exactly the same approach with all the names blank does this (see that attached version). On a mobile ADE device, this still allows you to use "goto page 14" to get chapter 10, but under Desktop ADE page numbers are then inactive (the page is always 1/24).

Like frabjous, I tried leaving some entries out of the page-map but this seems to be a bad idea in practice.

In a program like Calibre, it would be possible to produce an explicit synthetic page map that would be more consistent than the one used by Adobe. Adobe bases its synthetic page count on the compressed file size, so pages are not equal size. FBReader, for example, uses a similar scheme but based on uncompressed letter counts. In any case, if Calibre (say) produced a page-map of any kind it could have the option of either including the page number or a blank. In the latter case, the count would still be available to mobile ADE but there would be no page numbers overlaying the text on the screen.

[Jellby]

Originally Posted by frabjous

• Linking to extra documents not in the manifest or toc?

(For the last couple, I imagine trying to create a kind of ADE ePub "Easter Egg" where the only way to access certain content in the book is to type "xanadu" in the "Go to" box or similar.)

Hmm... The ePUB spec says every file should be in the manifest, and all files on an appropriate type (.xhtml) should be in the spine. Your suggestion, however, could be used if a reader supports linear="no" in the spine.

[charleski]

I think I got possessed by a malignant demon last night, because today I decided to play around with Adobe's page-map extension and discover its undocumented quirks.

I finally found solutions to my problems through intensive experimentation (aka trial-and-error) on version 1.7.2 and am writing them up here since I can find no mention of them elsewhere (probably because people don't use this extension that much).

Failure to obey these rules will lead to premature hair loss and a large dental bill.

1) If a page entry in the page-map file points to the first page in a flow, then the href MUST point to the file itself and NOT to a fragment within it.

Let's say you've been tempted into automating the creation of the page map and have the following entries in your page-map file:

<page name="1" href="content1.xhtml#page1"/>
<page name="2" href="content1.xhtml#page2"/>
<page name="3" href="content1.xhtml#page3"/>
<page name="4" href="content1.xhtml#page4"/>
<page name="5" href="content2.xhtml#page5"/>
<page name="6" href="content2.xhtml#page6"/>

Then ADE will display bizarre behaviour with page numbers appearing out of sequence or doubled-up at the same position even though all the anchors are properly formed and in the right places.

The page-map file should instead look like this:

<page name="1" href="content1.xhtml"/>
<page name="2" href="content1.xhtml#page2"/>
<page name="3" href="content1.xhtml#page3"/>
<page name="4" href="content1.xhtml#page4"/>
<page name="5" href="content2.xhtml"/>
<page name="6" href="content2.xhtml#page6"/>

2) The anchor identifying a page MUST occur in a block that also contains text.

<p class="graphic"><a id="page013"/><img alt="" src="[path]"/></p>

Will result in the page-number not showing in the margin, although it is still reported in the status-bar. Instead, the anchor should be moved to the closest block that contains text (i.e. either the end of the previous paragraph of text or the start of the following one).

[pdurrant]

Originally Posted by charleski

2) The anchor identifying a page MUST occur in a block that also contains text.

<p class="graphic"><a id="page013"/><img alt="" src="[path]"/></p>

Will result in the page-number not showing in the margin, although it is still reported in the status-bar. Instead, the anchor should be moved to the closest block that contains text (i.e. either the end of the previous paragraph of text or the start of the following one).

Umm... so by careful use of the page map, we could stop ADE displaying those blasted page numbers?