Different ePub displays in iReader/Sigil and Bluefire

[dgatwood]

Originally Posted by odedta

@JSWolf: That's a CSS3 rule isn't it?
Correct me if i'm wrong, devices that don't support CSS3 will ignore the entire file if it "has errors" such as font-face rules or media queries. So, it's best practice to put them in a separate stylesheet.

@font-face is in CSS 2.0 and CSS 3.0, was removed from CSS 2.1 because no browsers supported it, and is AFAIK the only part of the EPUB 2 spec that explicitly references CSS 2.0 instead of 2.1.

With that said, even if it weren't, the CSS specification (CSS 2.1 section 4.2) explicitly defines the rules for parsing declarations, and how to handle declarations that you can't interpret. Those rules are there precisely to ensure forward compatibility in situations like this.

More to the point, any reader that ignores an entire file because it "has errors" is egregiously violating the CSS specification, and should be considered defective by design.

That means you, Adobe.

[odedta]

Originally Posted by dgatwood

@font-face is in CSS 2.0 and CSS 3.0, was removed from CSS 2.1 because no browsers supported it, and is AFAIK the only part of the EPUB 2 spec that explicitly references CSS 2.0 instead of 2.1.

With that said, even if it weren't, the CSS specification (CSS 2.1 section 4.2) explicitly defines the rules for parsing declarations, and how to handle declarations that you can't interpret. Those rules are there precisely to ensure forward compatibility in situations like this.

More to the point, any reader that ignores an entire file because it "has errors" is egregiously violating the CSS specification, and should be considered defective by design.

That means you, Adobe.

I understand dgatwood and will look it up on the specs, thank you very much!

[JSWolf]

Originally Posted by dgatwood

@font-face is in CSS 2.0 and CSS 3.0, was removed from CSS 2.1 because no browsers supported it, and is AFAIK the only part of the EPUB 2 spec that explicitly references CSS 2.0 instead of 2.1.

With that said, even if it weren't, the CSS specification (CSS 2.1 section 4.2) explicitly defines the rules for parsing declarations, and how to handle declarations that you can't interpret. Those rules are there precisely to ensure forward compatibility in situations like this.

More to the point, any reader that ignores an entire file because it "has errors" is egregiously violating the CSS specification, and should be considered defective by design.

That means you, Adobe.

Actually, I do like that ADE dumps the CSS if it finds an error. It means that it is very easy to know that the CSS has an error. It could be a simple very easy to overlook error and do you really want to sell an eBook with an error in the CSS?

[odedta]

Wolf, disregarding the entire file for a simple mistake is silly, the book will look just fine if the device would just disregard it, if you sell ebooks i think you should validate all your files, that for me would be a minimum requirement of every ebook sold

[JSWolf]

Originally Posted by odedta

Wolf, disregarding the entire file for a simple mistake is silly, the book will look just fine if the device would just disregard it, if you sell ebooks i think you should validate all your files, that for me would be a minimum requirement of every ebook sold

But sometimes an error in the CSS can go unreported by the error checker. With the way I format things, it's very easy to know when the CSS is being ignored.

[mrmikel]

There is a great deal of silliness in how epub renderers work and it is NOT consistent even within brands. Nook barfs with too many changes in the first line, early apple doesn't match later apple, Kindles are not kindles if they are early Kindles etc.

This is why Hitch's company owns so many devices. It is not because she wants to support the manufacturers, but because that way she can be sure it works adequately in devices.

Epubs are not web pages. They are rendered a page at a time, so it has to be structured in a particular way to keep from blowing up. A web page keeps on going one and on and on and has access to a powerful computer to guess what the dopey programmer meant when they left off all the paragraph close tags.

[dgatwood]

Originally Posted by mrmikel

Epubs are not web pages. They are rendered a page at a time, so it has to be structured in a particular way to keep from blowing up. A web page keeps on going one and on and on and has access to a powerful computer to guess what the dopey programmer meant when they left off all the paragraph close tags.

Actually, they are just web pages. EPUB readers, however, are not web browsers. They're XHTML-only paged media viewers, which means that should behave like a web browser would if you were printing to a PDF (with a very small paper size).

The fact that they're XHTML-only means that they explode if tags don't match, but so do web browsers if you provide XHTML content with unbalanced tags.