View Full Version : capabilities of epub compared to mobi


comet
06-23-2011, 04:28 AM
Hallo,

I am interested to know, whether there are large differences in the capabilities of epub and mobi. Is is possible to do "more" using epub? Or doesn't matter, which format is used?

Adjust
06-23-2011, 04:47 AM
I would say you can do more in ePub that will NOT translate to .mobi

comet
06-23-2011, 05:20 AM
Do you have any examples?

Adjust
06-23-2011, 05:25 AM
Floats... from Memory

Jellby
06-23-2011, 05:29 AM
Possible in ePUB, not in Mobi: Embedded fonts, CSS, right margins, SVG, PNG.

Possible in Mobi, not in ePUB: Dictionaries.

comet
06-23-2011, 05:45 AM
Of course, the following question may be nonsense, but why not asking it: Is epub or mobi the "better" format? Or on which facts does that depend on?

I read at the moment a mathematical book with al lot of formulas I bought at Amazon. Of course it is a topaz. Other sellers only offer such books as pdf that is awfull for small e-reader-screens. So I wonder whether epub can do the same job as topaz.

Jellby
06-23-2011, 07:18 AM
Isn't topaz roughly equivalent to PDF formatted for a small screen?

comet
06-23-2011, 07:56 AM
I believe that's not true. The topaz-book I read at the moment has sometimes a very poor quality: Sometimes mathematical symbols or whole parts of formuals are absent (I compared it to the extract of the print edition on Amazon). The reason is that in topaz they probably use scanned parts of the book, and it seems that there are some problems concerning quality control for topaz-books, I believe. PDF is an exact copy of the layout of the original book with high quality - but poor for small e-reader screen. If I use pdf-reflow (e.g. on my sony) the mathematical formulas are completely destroyed. So I have to use topaz, if I want to use the ebooks.

DiapDealer
06-23-2011, 08:17 AM
Isn't topaz roughly equivalent to PDF formatted for a small screen?
Sorta... but not really. ;)

Topaz has a dictionary of glyphs (SVG fonts) that represent characters, groups of characters, or sometimes whole words that are used in the book. XML is then used to assemble those SVG glyphs and any images the book may contain into "pages" that are reflowable. The original OCR scan is included but only used for text search purposes—which due to uncorrected scan errors, often produces interesting results.

I've always felt that Topaz had the ability to produce some beautifully rendered books, but many times were hastily thrown together and oftentimes the glyphs were "fuzzy." But when a topaz book was good... it was damn good. ;)

Most of the hate for topaz comes from it's bloat—they're huge files (which probably contributed to sluggish page turns), the user's inability to be able to change the font (the SVG glyphs are hard-coded), and of course the largely unsatisfying results when trying to convert them to another format.

comet
06-23-2011, 10:46 AM
I've always felt that Topaz had the ability to produce some beautifully rendered books, but many times were hastily thrown together and oftentimes the glyphs were "fuzzy." But when a topaz book was good... it was damn good. ;)


So unreadable mathematical symbols are due to bad scans? The topaz-book, I read, was written in the eighties. If I look on the other side into a topaz-book that is published 2010 it looks really well!

DiapDealer
06-23-2011, 10:59 AM
So unreadable mathematical symbols are due to bad scans? The topaz-book, I read, was written in the eighties. If I look on the other side into a topaz-book that is published 2010 it looks really well!
Possibly. There's really no rhyme or reason to it. Other than shoddy vs competent work when creating the ebook.

AlexBell
06-24-2011, 05:15 AM
Possible in Mobi, not in ePUB: Dictionaries.

Err, I have my choice of dictionaries on my Sony PRS 650. I presume they are using ePub since I can check words in ePub files.

HarryT
06-24-2011, 05:23 AM
Err, I have my choice of dictionaries on my Sony PRS 650. I presume they are using ePub since I can check words in ePub files.

But those dictionaries are not ePub books - they are custom dictionaries created by the reader manufacturer and incorporated into the device's firmware. You can't buy a new dictionary for your Sony in an ePub bookstore; there's no specification of "how to write a dictionary" in the ePub standard.

With Mobi, dictionaries are a part of the format specification. If you want a Russian dictionary for your Kindle, you can buy one in the Kindle bookstore and install it on your Kindle.

HarryT
06-24-2011, 05:29 AM
I read at the moment a mathematical book with al lot of formulas I bought at Amazon. Of course it is a topaz. Other sellers only offer such books as pdf that is awfull for small e-reader-screens. So I wonder whether epub can do the same job as topaz.

The current ePub standard can't replace PDF or Topaz for complex layouts, but the new ePub 3 standard should be able to do so. However, it's going to be some years before ePub 3 devices are widespread.

comet
06-24-2011, 05:53 AM
Where can I find detailed specifications of mobi and topaz?

HarryT
06-24-2011, 05:57 AM
You can't; the format specifications are proprietary and have never been published. They've been reverse engineered, however.

DaleDe
06-24-2011, 01:13 PM
Where can I find detailed specifications of mobi and topaz?

there is almost nothing out on topaz. mobi has be reversed engineered and there is a great deal of information available. The absolute best data is in the mobileread wiki.

jbcohen
06-24-2011, 01:31 PM
I think the main difference between mobi and epub is that epub is an open format that anyone can use while mobi is a proprietary format unique to Amazon and Mobi Pocket.

HarryT
06-24-2011, 04:00 PM
That's certainly "a" difference, but I'm not sure I'd agree that it's "the main" difference.

comet
06-25-2011, 04:49 AM
there is almost nothing out on topaz.

But calibre can convert topaz files. So there must exist informations about this format.

DaleDe
06-25-2011, 12:38 PM
But calibre can convert topaz files. So there must exist informations about this format.

I guess I should have said there is very little published. The best definition I have seen is that "Topaz is a collection of glyphs arrange on pages, along with an unproofed OCR text version. An Amazon proprietary format, used to make older books available quickly, since conversion is essentially automatic from scans of the pages of a book, but it reflows very well."

This would make it similar to DJVU but likely with a mobi like structure. Perhaps someone can read the Calibre source code.

Elfwreck
06-25-2011, 12:57 PM
But calibre can convert topaz files. So there must exist informations about this format.

There's a bit of information of how it's put together, based on taking it apart. But since Amazon doesn't allow any to use a Topaz converter, there's no info assembled on "how to make a Topaz ebook"--because you can't.

Info on how they're made & what's possible with Topaz has to be done by spot-check comparison: acquire several Topaz ebooks from Amazon & compare them. This will show what *is* possible with Topaz; it won't show what isn't, because all that can be said so far is "Amazon has never released a Topaz book that has [X feature]."

DiapDealer
06-25-2011, 01:04 PM
But calibre can convert topaz files. So there must exist informations about this format.
As far as I know... no... calibre can't convert Topaz. It can handle the metadata for Topaz, but it can't convert it. I suspect that you may be using a certain plugin that automatically converts the OCR'ed text of DRM'ed Topaz files to xhtml (technically an HTMLZ archive) when you import a Topaz file into calibre. Calibre can then easily convert HTMLZ to other formats.

Any clues to the Topaz format will likely be found in the source code of that plugin.

rjnagle
06-29-2011, 12:00 PM
for the OP:

Seriously, mobi is an inferior and doomed format. It is like Netscape 4. It is finicky, doesn't support many features which epub does and doesn't have good docs. (Josh Tallent's book on kindle formatting is good, but out-of-date). Its support for CSS is limited and its handling of images is terrible. Without floats or borders (supported by epub, not by Kindle), there's not a lot you can do.

The only reason it is still around is that Amazon has 50% of the ebook market now and we have to "downconvert" epub files so that kindlegen can convert them to Kindle.

DiapDealer
06-29-2011, 11:32 PM
Seriously, mobi is an inferior and doomed format.
Ah... a prognosticator!
Remember... the Beta VCR format was superior to VHS in almost every way. ;)

DaleDe
06-30-2011, 01:53 AM
Ah... a prognosticator!
Remember... the Beta VCR format was superior to VHS in almost every way. ;)

Except for one, Sony refused to license its manufacture to anyone else until after VHS was a done deal. Oops, that is exactly like Amazon.

Dale

DiapDealer
06-30-2011, 08:27 AM
Except for one, Sony refused to license its manufacture to anyone else until after VHS was a done deal. Oops, that is exactly like Amazon.
Hey, I'm not fanboy-ing for one or the other. I just think it's rather humorous to see laypersons predict the doom of specific technology. Time will tell, I guess... no skin off my back either way. ;)

rjnagle
07-01-2011, 01:16 PM
Who you calling a layperson? I sir am a priest :)

Let's just say that I have been writing about ebooks and working in ebooks for a long time.

I realize we're just jabbering in all good fun, but I usually experience several moments of disgust every time I try to downconvert an epub file to a kindle.

Again, I ask you, how many people are using Netscape 4 these days?

It's true that people still buy their kindles and that Amazon carries 50% of the ebook market, but the kindle format hasn't been updated. And if you remember, the people who complained the most about Netscape 4 or IE 5-6 were the web designers --- they spent wicked amounts of time dumbing down their web designs so that things looked noncrappy on multiple platforms.

It's true you can switch out css on kindle ebooks, but Kindle also requires different semantics as well...

The biggest problem with Kindle is they're stuck with having to maintain a legacy format -- it's unlikely that many ebooks published over the last 3 years will ever be upconverted by publishers....

I'll just mention one thing: images. Images had to be jerryrigged to work in the Kindle e-ink devices. When people start using Kindle tablets to read their old purchases, those old purchases are going to look terrible....

HarryT
07-01-2011, 01:41 PM
I'll just mention one thing: images. Images had to be jerryrigged to work in the Kindle e-ink devices. When people start using Kindle tablets to read their old purchases, those old purchases are going to look terrible....

What do you mean by "jerryrigged"? I create numerous Kindle books and have had no issues at all with images.

DiapDealer
07-01-2011, 02:03 PM
I'll just mention one thing: images. Images had to be jerryrigged to work in the Kindle e-ink devices. When people start using Kindle tablets to read their old purchases, those old purchases are going to look terrible....
I sometimes manage to read for several years without coming across a book that contains images. ;)

hapax legomenon
07-02-2011, 12:54 PM
I am no fan of the Kindle either. It works fine for all text books, but not for books with any kinds of images.

rjnagle
07-02-2011, 12:58 PM
HarryT

1. You agree, don't you, that Kindle can't float images?
2. I found that images blow up unexpectedly, and that extra blank pages are created. What's more, there is no real documentation about when and why it does this. Also, I occasionally find discrepencies between how images are rendered in the Kindle Previewer and the Kindle itself.
3. Have you tried to make images on pages in a way that look suitable on both Kindle for ipad and an actual Kindle?

I can't recall offhand, but I seem to remember that there are problems with wrapping images inside div tags in Kindle. This breaks my workflow, which is Docbook to epub to Kindle (via Kindlegen). I imagine there are fewer problems when you are just starting with HTML.

charleski
07-03-2011, 06:36 PM
Is is possible to do "more" using epub?

Yes, simple as that. The Adobe ePub renderer in my PRS505 is several years old, but can still render in ways that the latest Kindle couldn't even dream of, and we're talking about basic elements such as margin control. This is because Amazon has failed abysmally in their stewardship of the .mobi format - it's still basically the same thing you could get over a decade ago.

there's no specification of "how to write a dictionary" in the ePub standard.

I have to say that, being a fluent native speaker, I don't quite understand this obsession with ebooks and dictionaries. But I'll admit that it's the sort of thing to file under 'this might be a good thing'. There are a lot of things that need to be fixed before worrying about dictionaries though.

The current ePub standard can't replace PDF or Topaz for complex layouts, but the new ePub 3 standard should be able to do so.

ePub 3 is a SCANDAL. They had a chance to address the current failings and point the way forward and they failed completely. Where is the page model? Where is the specification for concurrent text flows? (Try typesetting Embassytown in ePub - there's no option other than to rely on embedded jpegs.) Where is the specification for overlapping windows? You can't do something as basic as vertically center text on a page in the current standard - guess what, you can't do it in ePub 3 either. FAIL. ePub 3 fails to approach the rich standards of print layout in any shape or form.

virtual_ink
08-29-2011, 07:12 AM
Possible in ePUB, not in Mobi: Embedded fonts, CSS, right margins, SVG, PNG.

Possible in Mobi, not in ePUB: Dictionaries.

No CSS - argh! Jeepers.

HarryT
08-29-2011, 08:16 AM
No CSS - argh! Jeepers.

There's no CSS within a Mobi file, but Mobi creation tools such as Mobi Creator will convert CSS into the corresponding tag attributes, so it doesn't mean that you can't use CSS in your source files.

DaleDe
08-29-2011, 12:36 PM
There's no CSS within a Mobi file, but Mobi creation tools such as Mobi Creator will convert CSS into the corresponding tag attributes, so it doesn't mean that you can't use CSS in your source files.

It would be better to say, It converts some CSS to Mobi, since not all CSS is convertible and it some cases the results are not anything like what you want or expect. It depends on the available tags in the Mobi format.

Dale

HarryT
08-29-2011, 12:45 PM
It would be better to say, It converts some CSS to Mobi, since not all CSS is convertible and it some cases the results are not anything like what you want or expect. It depends on the available tags in the Mobi format.

Dale

Yes, that's very true. Thanks for the clarification.