Problem with inline image

[eggheadbooks1]

Quote:
Originally Posted by eggheadbooks1 View Post
Amazon still promote the idea that Kindle books can only read Latin-1 (ISO-8859-1) characters.
May I ask where you saw this?

It's on the KDP website -- https://kdp.amazon.com/help?topicId=A3G4LY8RGZ9SP6 -- which contains a lot of stupid stuff not found in the "official" guidelines. Specifically it says:

"Supported Characters:
Amazon Kindle Direct Publishing supports text in the 'Latin-1' (ISO-8859-1) format and all characters in that character set. The only characters from that set not currently supported are: spades, clubs, hearts, up-arrow, down-arrow, alpha, beta, gamma.

"That character set includes the most characters found in western European languages and many associated glyphs such as the 'copyright' mark, superscript (power of two) and German umlauts. You can see a list of all supported characters here [https://en.wikipedia.org/wiki/ISO/IE...-1#Code_table]. An image showing all those characters can also be seen here [https://images-na.ssl-images-amazon....p/Latin1.gif.]"

I know the KDP website is aimed at the lowest common denominator, but you would think at least the Guidelines would be more robust. Page 23 on text is really lame.

[Tex2002ans]

Originally Posted by eggheadbooks1

It's on the KDP website -- https://kdp.amazon.com/help?topicId=A3G4LY8RGZ9SP6 -- which contains a lot of stupid stuff not found in the "official" guidelines. [...]

That is just absurd. From a quick look through the pages there, it seems like nearly all of that information is geared towards Microsoft Word as an input format.

I bet this is just a case of REALLY dumbing information down to try to minimize the amount of Customer Support Amazon has to do.

[Hitch]

Originally Posted by Tex2002ans

That is just absurd. From a quick look through the pages there, it seems like nearly all of that information is geared towards Microsoft Word as an input format.

I bet this is just a case of REALLY dumbing information down to try to minimize the amount of Customer Support Amazon has to do.

Guys:

The online stuff IS oriented toward the DIY Word person. Stick with the new(er) Formatting Guidelines.

I haven't yet run into the unicode issue. Is it giving you agita on a specific item?

Hitch

[eggheadbooks1]

Originally Posted by Tex2002ans

That is just absurd. From a quick look through the pages there, it seems like nearly all of that information is geared towards Microsoft Word as an input format.

It is aimed at the DIY crowd, most of whom upload either a Word doc (in the earlier days) or an HTML file exported from Word (and possibly tweaked a bit by the publisher). Many are still uploading poor-quality mobi files made in programs like Scrivener. Or using auto-conversion software like Jutoh or Calibre to convert Word docs to mobi.

Originally Posted by Tex2002ans

I bet this is just a case of REALLY dumbing information down to try to minimize the amount of Customer Support Amazon has to do.

Yes, because Amazon still want to promote the concept of easy. You cannot even find the Publishing Guidelines PDF unless you first go to Getting Started then click on the link to Simplified Formatting Guide, then happen to notice there is a new left menu pane where there is a link to File Formatting Tips, click on that, then on HTML, then click on the link to the PDF. It's that convoluted. And the KDP website, like the Publishing Guidelines, are "updated" by software engineers instead of proper writers. The KDP website is an absolute mess.

Which brings me back to the PDF. If the Publishing Guidelines are the official coding guidelines, why not make mention of Unicode? This may be a question no one except Jeff Bezos -- who has demonstrated repeatedly that he follows a logic not often shared by others -- can answer.

But about the question regarding exporting Word to HTML for importing into Sigil: I vaguely recall a conversation I had with John Schember about Sigil mapping Windows-1252 characters to UTF-8 upon import, but I notice that if one exports to HTML using Windows defaults, and then imports into Sigil, the charset is still listed as Windows-1252 even though the header code in Sigil says the file itself is UTF-8.

So now I am wondering if there is any situation where that would cause problems on Kindle or ePub. Should one change the defaults in Word to export to HTML as UTF-8 when following the workflow of Word to HTML to Sigil to ePub? Or should one just change the header code in Sigil to say the charset is UTF-8 since Sigil has already automatically mapped the character set?

Or, to put it another way, does it make any difference, or cause any problems, if the charset (Windows-1252) differs from the file encoding (UTF-8)?

[Tex2002ans]

Originally Posted by Hitch

I haven't yet run into the unicode issue. Is it giving you agita on a specific item?

So I went back to reread the section yet again, and they have this sentence right after mentioning the two spaces:

Originally Posted by AmazonKindlePublishingGuidelines

Use of any other space can break the selection, dictionary lookup, and line-wrap algorithms.

I don't have a Kindle on hand, can anyone test Search/Selection (Copy/Paste)/Dictionary when using Thin Spaces? It could be like some devices with Soft Hyphens. Theoretically it shouldn't break anything (and might look ok on the surface), but in reality, many devices/programs don't handle those characters properly in some sub-surface cases.

I must admit, I never looked too closely at actually using Thin Spaces heavily in ebooks (I only worked on one French Canadian book a few years back, where I first came across all that Thin Space intel).

Side Note: IF you want to read about the French punctuation spacing, here is some info:

http://unicode.org/udhr/n/notes_fra.html

https://wiki.openoffice.org/wiki/Non....28U.2B202F.29

https://www.cs.tut.fi/~jkorpela/html/french.html

and here was some Canadian French spacing rules (slightly different from normal French):

http://www.btb.termiumplus.gc.ca/tcd...ect17&info0=17

Originally Posted by eggheadbooks1

And the KDP website, like the Publishing Guidelines, are "updated" by software engineers instead of proper writers.

LOL I wouldn't say that. If it was written by engineers it would REALLY get down into the nitty gritty (like none of my posts).

But try to explain something technical (like encoding) to your grandma. There is just no way!

Once you get a few paragraphs of technical words, their eyes glaze over.

Easier to just say: "Save it as ASCII/ANSI, if it doesn't work... not supported." Anything outside of (English letters + a few accented characters + most common symbols) turns to mush: "Well, if you keep having trouble, you can pay Createspace to create the file for you. *wink wink*"

Side Note: It reminds me of this topic on the LaTeX Stack Exchange, "How can I explain the meaning of LaTeX to my grandma?":

https://tex.stackexchange.com/questi...-to-my-grandma

Originally Posted by eggheadbooks1

Should one change the defaults in Word to export to HTML as UTF-8 when following the workflow of Word to HTML to Sigil to ePub?

I am not familiar with the intricacies of Microsoft Word, but by default mine is set to save as UTF-8 (don't know if that was just a setting I set a very long time ago or what, I have 2010 installed if that means anything).

There is just absolutely no reason I can see why any sort of new documents should be non-Unicode in this day and age.

[Doitsu]

@eggheadbooks1: You can ignore all confusing Amazon statements about Unicode support for the following reasons:

1. As far as Kindle books are concerned, only KindleGen Unicode support matters, and KindleGen and its precursor MobiGen have supported Unicode for a long time.

2. The epub2 specs require all books to be encoded as either utf-8 or utf-16. (utf-16 was presumably only added because it can theoretically reduce the size of CJK HTML files by 20-30%.)

3. KindleGen explicitly supports valid epub2 and epub3 books as input files. (KindleGen doesn't support all epub3/CSS3 features, though.)

Usage : kindlegen [filename.opf/.htm/.html/.epub/.zip or directory] [-c0 or -c1 or c2] [-verbose] [-western] [-o <file name>]

For all these reasons, utf-8 is the best format for both epubs and Amazon KDP books.

[Notjohn]

Originally Posted by Doitsu

@eggheadbooks1:

3. KindleGen explicitly supports valid epub2 and epub3 books as input files. (KindleGen doesn't support all epub3/CSS3 features, though.)

So I should not be eager to make the transition to ePub3 if I upload epubs to the KDP?

[eggheadbooks1]

Originally Posted by Tex2002ans

LOL I wouldn't say that. If it was written by engineers it would REALLY get down into the nitty gritty (like none of my posts).

Actually, my experience with stuff written by software engineers is that they usually skip over what they think are obvious steps in the process. So they'll tell you to look in X folder on your computer, but it's actually found in a sub-folder of a sub-folder in X. Drives me nuts.

Originally Posted by Tex2002ans

But try to explain something technical (like encoding) to your grandma. There is just no way!

I don't even bother. Last night my elderly mom (whom I take care of) asked me what I was doing. "Asking about Unicode characters in ebooks." "What's that?" "Never mind. Eat your dinner."

Originally Posted by Tex2002ans

Easier to just say: "Save it as ASCII/ANSI, if it doesn't work... not supported." Anything outside of (English letters + a few accented characters + most common symbols) turns to mush: "Well, if you keep having trouble, you can pay Createspace to create the file for you. *wink wink*"

Yes, I think that is often Amazon's strategy. I notice it elsewhere, too.

Originally Posted by Tex2002ans

I am not familiar with the intricacies of Microsoft Word, but by default mine is set to save as UTF-8 (don't know if that was just a setting I set a very long time ago or what, I have 2010 installed if that means anything).

The default encoding in all versions of Word shipped in North America is Western European (Windows), aka Windows-1252. You have to deliberately change it to UTF-8 when saving if you want UTF-8. The default encoding in Notepad is ANSI.

[eggheadbooks1]

Originally Posted by Doitsu

@eggheadbooks1: You can ignore all confusing Amazon statements about Unicode support for the following reasons:

1. As far as Kindle books are concerned, only KindleGen Unicode support matters, and KindleGen and its precursor MobiGen have supported Unicode for a long time.

2. The epub2 specs require all books to be encoded as either utf-8 or utf-16. (utf-16 was presumably only added because it can theoretically reduce the size of CJK HTML files by 20-30%.)

3. KindleGen explicitly supports valid epub2 and epub3 books as input files. (KindleGen doesn't support all epub3/CSS3 features, though.)

Usage : kindlegen [filename.opf/.htm/.html/.epub/.zip or directory] [-c0 or -c1 or c2] [-verbose] [-western] [-o <file name>]

For all these reasons, utf-8 is the best format for both epubs and Amazon KDP books.

Exactly. I've been doing the Sigil-convert-to-mobi-in-Kindlegen for my ebooks for the last several years, and this has always created the best results. So I find Amazon's Guideline's strange. And as far as fonts are concerned, the guide mentions Monospace font as a secondary option that does not need to be embedded; why not also mention Unicode? It's a mystery. But then that's Amazon.

Which is why I come here for help, and not KDP's alleged technical help department, which is manned by people for whom English is a foreign language and who often don't understand complex questions. I gave up on them a long time ago. When Amazon/Adobe first developed the InDesign Kindle plugin, there was a feedback button that sent one's email directly to the Kindle programming team; THEN you got proper answers. But now that feedback button links (or is forwarded) to the KDP help department, which as stated is of no help at all.

[Hitch]

Originally Posted by eggheadbooks1

Exactly. I've been doing the Sigil-convert-to-mobi-in-Kindlegen for my ebooks for the last several years, and this has always created the best results. So I find Amazon's Guideline's strange. And as far as fonts are concerned, the guide mentions Monospace font as a secondary option that does not need to be embedded; why not also mention Unicode? It's a mystery. But then that's Amazon.

Which is why I come here for help, and not KDP's alleged technical help department, which is manned by people for whom English is a foreign language and who often don't understand complex questions. I gave up on them a long time ago. When Amazon/Adobe first developed the InDesign Kindle plugin, there was a feedback button that sent one's email directly to the Kindle programming team; THEN you got proper answers. But now that feedback button links (or is forwarded) to the KDP help department, which as stated is of no help at all.

FWIW--and I say this with NO hubris or misplaced pride or anything along those lines--I have found that often, we (my shop) have a better grip on what works, and what doesn't, and where bugs exist, etc., than most of the guys/gals in Seattle. It's not about their programming-fu; it's that they don't run into the issues that we do, because what they work on isn't necessarily based in doing real books. And we (I) do deal with the heads of departments, the hardcore code guys, and the like. For example, we were the ones that told them that the Voyage had a display bug, around images larger than 50% of the screen width and smaller than 100%. Can you imagine? All those resources devoted to creating a new reader, and WE are the ones who find the first large bug? It happens.

Lord knows, I've nearly pulled my hair out on some discussions. I had a lengthy, ongoing discussion with someone there, about the SRL (Start Reading Location). He persistently INSISTED that if we used the HTML "start" anchor, (uh...), that the SRL would work. (We were primarily using the "text" meta...). No matter how many times we demo'ed that that didn't work, he sort of refused to "see" it.

Trust me, I have a litany of those types of discussions. Like everything in coding, you don't find the glitches and bugs, really, until some power-user uses it, and breaks it.

So, I don't hold that against them. The people that you speak to, eggheadbooks1, don't really work in the Tech end. They know the basics, and can usually shepherd a DIYer, but they're not at the upper ends, vis-a-vis something as "sophisticated" as uploading an ePUB instead of a MOBI. Those folks are designed to work with the Word DIYers.

Hitch

[Notjohn]

Originally Posted by eggheadbooks1

Last night my elderly mom (whom I take care of) asked me what I was doing. "Asking about Unicode characters in ebooks." "What's that?" "Never mind. Eat your dinner."

I'm sure that made you feel a bit superior, but do you have any idea how it made HER feel?

[eggheadbooks1]

Originally Posted by Notjohn

I'm sure that made you feel a bit superior, but do you have any idea how it made HER feel?

She laughed. Don't pretend to know the dynamics of anyone's relationships unless you're there.

And don't comment on how YOU'RE sure I feel when you don't know me from Eve.

[eggheadbooks1]

Hitch,

I know of someone that needs a complex ebook done: he has portions of text that are on curves and such that he wants replicated. The only way I can see to do it is either as a regular image file (GIF or JPEG, maybe PNG) or, likely better, in an SVG. I'll send him to your website.

[Hitch]

Originally Posted by eggheadbooks1

Hitch,

I know of someone that needs a complex ebook done: he has portions of text that are on curves and such that he wants replicated. The only way I can see to do it is either as a regular image file (GIF or JPEG, maybe PNG) or, likely better, in an SVG. I'll send him to your website.

Well, we are always appreciative of folks being referred. I would most likely demur from trying to replicate curved text, but I can't know until I see the file. Thanks much!

Best,
Hitch

[Doitsu]

Originally Posted by eggheadbooks1

I know of someone that needs a complex ebook done: he has portions of text that are on curves and such that he wants replicated.

BTW, the first file of the epub3 test suite, EPUBTEST 0100, contains an SVG example of text on a curve.