Marvin page numbering and images

[Faterson]

Originally Posted by eschwartz

"logical and natural" is your subjective opinion.

Of course. And your differing opinions are likewise subjective opinions. What else is new? That's what discussion boards are for – they allow us to exchange our subjective views.

Originally Posted by cedhax

since that's not the case now with Marvin 3's newly invented page numbering scheme, I just want to know how images are counted as "pages" as conceived by kguil.

I get you, but if you permit me to give an answer as well, my answer to your question would be: "Does it matter at all?" The current "pages" in Marvin are pure fiction – the numbers literally mean nothing, and it infuriates me every day to see them in the footer. So, does it really matter how images are treated in this purely arbitrary scheme? I mean, when the page-counts are literally meaningless, what difference would it make if an image is counted as 0.25 of a (fake) "page", or 0.5 of a (fake) "page", or 2.0 of a "fake" page?

Originally Posted by JKenP

The only thing that seems a solution here for those with their shorts in a knot is to quit ebooks and buy paper books whose numbering is in stone/ink.

Those would be my opponents in this and the other thread. Because that is exactly my line of reasoning: e-book pages can never be set in stone/ink, but Marvin 3 is attempting to do that, absurdly pretending that the same e-book has the same number of pages on both a 4-inch iPhone and a 13-inch iPad. So, it's really Marvin 3 who has its "shorts in a knot".

Originally Posted by Turtle91

You'll never win an argument against someone who ignores everyone else's opinion as "stupid" or "nonsensical" and imposes their own opinion as the only logical way of doing something.....

That's right. Were you describing anyone in particular? You certainly couldn't have meant me because I keep tirelessly repeating that I do respect all differing views (I said so explicitly to Jon earlier in this thread); all I'm asking for is for Marvin to give us an alternative way to display page-counts – different from the current way which I do consider absurd and untenable from the point of view of consistency in software.

Marvin 3 currently insists that a page in a chapter is a screen flip (my preference), but at the same time, Marvin 3 insists that a page in a book is something entirely different – an arbitrarily invented (and ill-chosen), abstract concept of "250 words". How can Marvin 3 call two entirely different concepts the same thing – "page", even simultaneously inside the same footer? Such inconsistency should be anathema to any serious-minded software developer. I know that Kris is a seriously-minded software developer, and therefore I very much hope he will rectify the current page-count situation in Marvin 3. If Kris simply gives those of us who prefer it, the option to count pages in books in the same way that Marvin has always counted and still counts pages in chapters (screen flip = page turn), then I'll be perfectly satisfied. The current abstract "page" scheme can be kept (even as default) for those Marvin users who prefer it, and yet another "page" scheme, ADE, might be added as an option to oblige users like Jon.

The esteemed MobileRead poster Tex2002ans sent me a private message on the current subject, alerting me to an important MobileRead thread on cross-referencing in e-books, and I'm quoting a part of my reply to Tex here below.

*****

That post #129 is another classic from you. My position is – exactly for the reasons you detailed in that post – that trying to conceive "pages" in e-books as any sort of cross-referencing platform is pure illusion: it's chasing a chimera, and it will never work. That's why I argue that Marvin should simply employ "pages" in the common-sense manner: a screen flip = a page turn, and that's it. No attempt whatsoever at pretending that those page counts have any relation whatsoever to anywhere other than your current reading device and your current formatting settings. Unfortunately, it's pretty difficult to communicate it in these forums; but I do hope that Kris as the Marvin developer will recognize that we should at least get this option in Marvin. For folks who believe that "pages" in e-books are something other than simple screen flips on your current reading device... well, let them cling to that illusion. I don't really care, as long as Marvin also gives us that other, "natural/realistic" page-count option.

As to my own preference for cross-referencing in e-books, it would be "percentage inside e-book", using the 4-digit format: 72.54%, whereby the percentage is determined by word count (say, 5,000th word in a 10,000-word book would be 50.00% inside the book). It's not precise either, and there are potential pitfalls you detail in your post #129, but it seems to me to be the best option from among all available. But let's just forget about pages for cross-referencing purposes in e-books, please. As I believe you indicated in your post #129, e-books resemble webpages a lot more than they do physical books, and we have no page numbering for webpages, right? And, of course an e-book is basically built out of HTML code just like a webpage is. E-books and webpages are "brothers and sisters", so to say, whereas printed books are just distant relatives. What works for distant relatives (cross-referencing them via page numbers) will never work for webpages or e-books, because the latter are (thank heavens!) fully fluid formats.

Word count is an excellent option, in my opinion, to determine the "percentage read"/location inside an e-book; but word count is a terrible choice even for abstract page counts, which is why I believe ADE is right in defining a page (the "abstract/fake" page) as 1024 characters, but Marvin is completely wrong in defining it as 250 words. As I mentioned elsewhere, German (for example) has much longer words than (say) English – so does this mean now that all German e-books will have fewer pages, on average, in Marvin, than English e-books? That's just unacceptable to me, even for an abstract/arbitrary "page" count. If some of the Marvin users like this particular "page"-count illusion, OK – let them use it in Marvin (by default, for all I care). But please don't enforce this as the only page-count option for all Marvin users.

[kguil]

@cedhax None.

Footnotes:
The original poster asked a simple, direct question to which there is an equally direct and simple answer. Please don't hijack his/her post.

[eschwartz]

Originally Posted by kguil

Footnotes:
The original poster asked a simple, direct question to which there is an equally direct and simple answer. Please don't hijack his/her post.

I don't think that will happen until you give in and agree to obey @Faterson's every command, wish, and feature request.

Because as long as Marvin continues to be "flawed" in all these myriad ways, the most important thing in this forum is to take every opportunity to interject and describe how Marvin desperately needs to be "fixed".

[JKenP]

Well, I guess if you only read cookbooks then you need page numbers to get to the chocolate cake section. Faterson must be a baker of note. Hence not having page numbers leaves everything half-bake in his view. Somebody stick a toothpick in him.

[eschwartz]

In all seriousness, the (misguided ) demands Faterson made do not do very well at all for that purpose.

Now you are opening a huge can of worms, ask the gurus @Hitch and @Tex2002ans if you like...
"see page xx" type instructions essentially require either being replaced with hyperlinks, or Amazon's "Real Page Numbers" (or Page-map/PageList which few books seem to actually contain).

Option #1 requires shenanigans to create separate paper/digital master copies. And that's very fun when digitizing old books, too.
Option #2 requires a lot of confidence in the vendor, and breaks hard on many many reader apps no matter what.

Also, Faterson already vehemently rejected option #2.

[ShellShock]

Originally Posted by eschwartz

Also, Faterson already vehemently rejected option #2.

That settles it then.

I am starting to find his contributions oddly entertaining in a masochistic fashion.

[Tex2002ans]

Originally Posted by eschwartz

Now you are opening a huge can of worms, ask the gurus @Hitch and @Tex2002ans if you like...
"see page xx" type instructions essentially require either being replaced with hyperlinks, or Amazon's "Real Page Numbers" (or Page-map/PageList which few books seem to actually contain)

For those of you who missed the topic, there was a topic called "Sick of Amazon Kindle books without Page Numbers".

I went into quite a bit of the nitty gritty in Posts #129 and beyond (although I would recommend reading the entire topic):

https://www.mobileread.com/forums/sho...85#post3290585

Side Note: In the above topic, I also brought in the (in my mind) heavily related topic of Indexes + Referencing in ebooks as well.

Originally Posted by ShellShock

I am starting to find his contributions oddly entertaining in a masochistic fashion.

I had a chat with Faterson over PM. Page Numbering is a topic he is extremely passionate about (as am I).

There are quite a few alternate methods of "Page Numbering" in ebooks. In Post #166 I tried to categorize what I believe are the different methods:

https://www.mobileread.com/forums/sho...97#post3293297

Faterson is a proponent of what I categorized as the "Screen Numbering" method, while I argue strongly in favor for others (and try my best to lay out the Pros/Cons of the different methods).

I have been meaning to make a summarized post as a proposal to kguil, as I think Marvin (or any ereading program) would benefit from being able to choose between alternative numbering schemes.

Plus I wanted to brainstorm + come up with good ideas for the details/variables/options (for example, Percentage displaying the tenths/hundredths place). I didn't want to just reiterate what was stated elsewhere.

I will probably be making that post in one of the other topics (or in a topic of its own).

[cedhax]

Originally Posted by cedhax

In Marvin 3, a page is defined to be 250 words of text. But how many pages or fraction of a page is considered occupied by an image element?

Originally Posted by kguil

@cedhax None.

Footnotes:
The original poster asked a simple, direct question to which there is an equally direct and simple answer. Please don't hijack his/her post.

@kguil - Thank you for your straightforward answer to my simple question and your footnote.

But do you think it's better to have an image element taking up some number of page(s) other than zero? Otherwise, the result (as it is defined now in Marvin 3) may be very strange if not bad in some cases.

For example, consider some comic book in EPUB with hundreds of images all referenced from the same HTML file, where there is no interleaving whitespace between adjacent image elements. By your current definition of a page, this book will then have only zero page! I'm really not comfortable with this implication. What do you think?

[kguil]

@cedhax In the ideal case, yes, you're right. However, determining the "size" of an image requires prerendering the whole EPUB and it slows everything down. This prerendering would need to happen each time the text size, page margins, indents, etc... change.

[Faterson]

Please definitely make that pre-rendering a user option. If you can pre-render chapters, you can certainly pre-render entire EPUBs – at least as a user option. iBooks can prerender EPUBs, even showing nicely the number of real pages (screens) in Table of Contents – therefore, Marvin should definitely accomplish the same, given that it runs on the same devices as iBooks.

Really, calling two entirely different concepts the same name, "page", even simultaneously in the same footer (!), is the most glaring, in-your-face inconsistency in the history of Marvin. If most users don't care about that inconsistency, OK. But for those of us who do care, please give us the option to treat "pages in book" in the same way as "pages in chapter". Then, cedhax's concern would be "automatically" taken care of as well.

[JKenP]

So, you want actual counts for consistency? Say you have a 4.6" phone and a 11.6" tablet. How consistent is the count here? And the variance will exceed Marvin's standard page which is at least consistent over a variety of screens to find a position.

[Faterson]

It's all perfectly natural and 100% consistent on the same reading device, using the current settings.

For finding your current reading position, you have Marvin's automatic location sync. Why would you wish to use pages instead? These are e-books, not printed books. Just forget about pages for that purpose.

Why would you, in fact, in a discussion about pages, mention "consistency" across different layout settings, or across different reading devices, or even across different e-reader apps and platforms? That is not at all what I'm talking about. No such consistency can ever exist, in my opinion – it's a pipe-dream. If you happen to be fond of that particular pipe-dream, good for you! Then you can happily continue using Marvin's current "250 words per so-called-page" scheme (and that's no standard, just an arbitrary setting), or use ADE's "pages" scheme (equally arbitrary), or use Kindle's (arbitrary) locations, or use Kindle's cross-references to pages of an (arbitrarily chosen) printed edition – whichever arbitrary scheme suits your personal preference best.

As to me, I discard all of those arbitrary schemes. Just give me natural page-counts instead, please – I flip a page, and the page-count goes up/down by 1 exactly, regardless of your current layout settings, regardless of your current reading device, app, or platform.

I'm simply devoid of those "consistent cross-device page-count" illusions. I do not wish to see the same number of pages for the same book across all reading devices (tiny and large) and even platforms. I don't wish to see it because I know it's a pipe-dream.

For cross-device and cross-platform references to text locations, I propose using the percentage-into-text metric instead, with two decimals, calculating it based on word counts. If you're 75.29% inside a book on one device or platform, and then look up the 75.29% location on a different device or platform, you should land pretty much reliably in the same location. No page references are necessary for this.

When you buy a printed paperback edition of the Bible, JKenP, and then a large-format hardcover edition of the Bible, you don't expect the two editions to have the exact same number of pages, right? Why should they? It's perfectly natural for the small-format Bible to have a higher number of pages than the large-format Bible.

I expect the same, 100% logical behavior from Marvin: the same e-book, when opened in Marvin on an iPhone, should by necessity have a much higher number of pages than when it's opened in Marvin on the 13-inch iPad.

If you don't care for that logic and prefer the fiction of "pages" instead, fine. Kris can keep that as Marvin's default, the way it is now. But for those of us who find it nonsensical, please give us the alternative – what I call the "natural page-counts". Marvin currently counts pages in chapters naturally, while digressing into fictional "pages" for page-counts in books. That is what's currently inconsistent within Marvin itself, on a single reading-device, and I will appreciate if Kris gives us the opportunity to avoid that inconsistency.

iBooks does give us this opportunity (in fact, it's the only page-count option in iBooks), so why should Marvin withhold it from us? Yes, it's more resource-intensive, but the calculation typically only needs to be performed once for every e-book, ideally when it's opened for the first time. iBooks calculates all of this without any noticeable delays; and I for one would be perfectly willing to wait even for a minute just for the privilege to be able to view natural page counts both for chapters and books as I'm reading my books in Marvin.

[eschwartz]

And I will go further and say you should deprecate page numbers entirely.

Call it what it is -- an arbitrary location indicator.

[Faterson]

They are not at all arbitrary – they are 100% precise, and correspond to reality exactly. You flip a page = the page-count goes up or down by exactly 1. That's exactly how it works in printed books, and that's exactly how it should work in e-books, too. Regardless of the size of the printed book, and regardless of the size of your electronic reading device. That's exactly what pages are.

There's no need to discard the concept or word "pages". It's an extremely useful and time-honored concept. The only confusion arises when you guys, for some reason unfathomable to me, expect pages to perform what they never can perform: to be consistent across various layout settings, or reading devices, or apps, or platforms. Why should they be consistent, for heaven's sake? Whoever thought of any such requirement? We never expected page-counts to be consistent in the world of printed books – so why should they, all of a sudden, be consistent in the world of electronic books?

Once again, for cross-referencing text locations across different layout settings, reading devices, apps, or platforms, I propose to forget about pages once and for all (but only for that particular cross-referencing purpose!), and use the percentage-into-text (with 2 decimals) metric (based on word counts) instead.

[eschwartz]

Generally speaking, a paper book does not change its own page numbering!
And multiple copies of the same edition of the paper book will always have the same page numbers.

It's an extremely useful and time-honored concept.

That is precisely why I don't like it! It has a tremendous amount of legacy baggage associated with the term.

Using the @Faterson definition of a page does not make it any less subjective than using the @eschwartz definition.

...

Having seen more than enough cases of people actually expecting ebook pseudorandom-calculated "pages" to be useful reference indicators (and the amusing intersection with acedemia), I no longer hold the delusion that holding onto that word will be helpful to the masses.