A problem with ADE?

[Tex2002ans]

Originally Posted by GrannyGrump

Wouldn't embedding a main body-text font make it difficult for users who want sans-serif? Or who just don't like the font you choose to inflict on them?

Not to mention those older/buggy devices which do not let you change to a user font once there is an embedded font.

Originally Posted by Jellby

That (the bold part) may be the significant difference. The rest of the book is probably justified, which apparently alters the default word spacing (or you are being unlucky and find too many shrunk lines)

Good thinking, I thought of that when posting my second test. I tested it in my older books by removing the justification, and still saw the problem. Just forgot to report about it/post images. Great minds think alike!

[AlexBell]

Thanks for all the responses. to me it seems more and more likely that it's an ADE desktop reader fault. I don't know whether to worry about it until I know what proportion of people use the ADE2 desktop reader; I wouldn't be seen dead in a ditch with it myself.

Jellby: Both the publisher and I very much prefer to see long line poetry indented when the line breaks, and I don't know of any other technique to do that. I've used the technique quite successfully (I think) for long titles also in this ebook and several others, and haven't seen the problem - but then I rarely use ADE2 desktop reader.

[Jellby]

Originally Posted by AlexBell

Jellby: Both the publisher and I very much prefer to see long line poetry indented when the line breaks, and I don't know of any other technique to do that. I've used the technique quite successfully (I think) for long titles also in this ebook and several others, and haven't seen the problem - but then I rarely use ADE2 desktop reader.

Nothing wrong with that, in fact, that's what any sensible ebook creator should do. I was merely commenting that the difference in word spacing between the poetry section and the rest might be due to the different text-align values.

[AlexBell]

Originally Posted by Jellby

Nothing wrong with that, in fact, that's what any sensible ebook creator should do. I was merely commenting that the difference in word spacing between the poetry section and the rest might be due to the different text-align values.

Sorry, I misunderstood you.

[AlexBell]

Update: GrannyGrump kindly beta read the whole book and suggested that I try a serif font to see if the problem still occurs. I changed the CSS to include

font-family: Amasis, serif;

rather than font-family: sans-serif;

and that fixed the problem. Please see attached images of the original screen shot of the sans font read with ADE and the new serif font read with ADE.

So far as I can see it's an ADE problem. Should I care? How many people use ADE to read ebooks?

[eping]

Originally Posted by AlexBell

So far as I can see it's an ADE problem. Should I care? How many people use ADE to read ebooks?

You can ignore it.
I think the portion of ADE is near zero.
ADE is only used by publisher for approximate preview.
And there're too many bugs on ADE(Shame on Adobe).

To be a serious professional publisher, you must test on real devices, such as Nook, Kobo, iPad.

[DaleDe]

Originally Posted by eping

You can ignore it.
I think the portion of ADE is near zero.
ADE is only used by publisher for approximate preview.
And there're too many bugs on ADE(Shame on Adobe).

To be a serious professional publisher, you must test on real devices, such as Nook, Kobo, iPad.

Actually ADE in the mobile version is used by a majority of devices including Nook and Sony as well as many others. Adobe is the leading supplier of code for mobile devices.

Dale

[eping]

Originally Posted by DaleDe

Actually ADE in the mobile version is used by a majority of devices including Nook and Sony as well as many others. Adobe is the leading supplier of code for mobile devices.

Dale

It sounds reasonable in theory. ADE and most readers are based on Adobe SDK.
But in reality, they are quite different on many respects. They share the same bugs of Adobe SDK, but each of them has particular bugs and advantages of their own.
For example, Nook has no landscape mode, so test on ADE can not reveal true user experience of your book on Nook.
Such as TOC, ADE has a side TOC tree, but such things doesn't exist on most real devices. Tree are not supported at all.
And Nook has several font choices which you can not test on ADE, so you must test on Nook to see real effect of those fonts.
Furthermore, if there's error in CSS, epubcheck doesn't report it, and ADE can ignore it, but on Nook, it crashes.
So only real device test is reliable.

[mrmikel]

You can check css through w3 validation service, though.

http://jigsaw.w3.org/css-validator/

[DaleDe]

Originally Posted by eping

It sounds reasonable in theory. ADE and most readers are based on Adobe SDK.
But in reality, they are quite different on many respects. They share the same bugs of Adobe SDK, but each of them has particular bugs and advantages of their own.
For example, Nook has no landscape mode, so test on ADE can not reveal true user experience of your book on Nook.
Such as TOC, ADE has a side TOC tree, but such things doesn't exist on most real devices. Tree are not supported at all.
And Nook has several font choices which you can not test on ADE, so you must test on Nook to see real effect of those fonts.
Furthermore, if there's error in CSS, epubcheck doesn't report it, and ADE can ignore it, but on Nook, it crashes.
So only real device test is reliable.

ADE is a good first check. ADE also doesn't support landscape mode although you can resize the window. (Resizing the window is a good way to check for several problems by the way.) ADE TOC in my experience shows identically to the external NCX TOC you get on reading devices and a tree is certainly supported by the SDK. Nook does have some changes and a unique DRM but it is still based on the SDK from Adobe and depending on which version of ADE you run the CSS errors will be dramatic in ADE in that it will ignore the whole CSS entries if there is an error. ePubcheck does not check the CSS at all. You need a special CSS checker. Certainly a real device is better but then you need all of the real devices since they are all different to fully test. ADE can give you a good first check that will catch most errors for most Adobe SDK devices.

Dale

[eping]

"only real device test is reliable" and
"ADE is a good first check" don't contradict.