Epub check tool: ePubs do not pass - Page 2

tompe · 08-19-2009, 03:29 PM

Quote:

Originally Posted by kovidgoyal

No, I'm saying that those files will still work in all current readers.

I still think it is wrong to call these files ePub files. We shoud have another term for "nearly ePub and happens to work for now" files.

kovidgoyal · 08-19-2009, 04:05 PM

Quote:

Originally Posted by tompe

I still think it is wrong to call these files ePub files. We shoud have another term for "nearly ePub and happens to work for now" files.

pepak · 08-20-2009, 03:43 AM

Quote:

Originally Posted by tompe

So what will happen when a new reader that is more restrictive in what is accepted is released?

Let me be brutally honest: Who cares? The fact is that strive as it may, EPUB as it is now is not an universal format and books in it must be modified for different implementations (e.g. Adobe DE as implemented on PRS-505 will not display international characters unless the EPUB is specifically modified, it will crash the Reader on certain CSS constructs [sadly, I can't remember what it was, but it "worked" quite dependably], etc.; other implementations have their own specific quirks that need to be addressed). Does it really matter if there is one more thing to fix?

Also, get back from the academic paradise to the real world. As Kovid said, most sources are not valid XHTML. The conversion process then must decide whether it will be strict or whether it will be usable (consider what would happen to a web browser which refused to render invalid pages!).

Jellby · 08-20-2009, 05:30 AM

Quote:

Originally Posted by kovidgoyal

Umm come up with an automated way to take arbitrarily bad HTML ad produce standards compliant XHTML from it while still preserving its meaning to human readers and I will be happy to implement it.

Hmm... If it can be displayed by a (X)HTML viewer, it can be converted to valid XHTML, can't it? I mean, if there are unclosed or invalid tags, the viewer already has to do something about them, so it can be done. There is the problem that different viewers might do different things, but that's the problem of non-compliant sources

In practice, I guess you could add a flag to "force ePUB compliance" or issue a warning if the generated ePUB is non-compliant.

I'm still an idealist, and think that we should press Adobe to make their reader standards-compliant and hope that other companies, groups or individuals develop independent standards-compliant ePUB readers. Then, when it comes to convincing someone to fix a bug in their readers, it's probably better to say "it does not work with this standards-compliant files" than "it does not work with this file that happens to work in ADE".

tompe · 08-20-2009, 06:10 AM

Quote:

Originally Posted by pepak

Also, get back from the academic paradise to the real world. As Kovid said, most sources are not valid XHTML. The conversion process then must decide whether it will be strict or whether it will be usable (consider what would happen to a web browser which refused to render invalid pages!).

But the principle on the web has always been to be strict and correct in what you generate but allow a lot of mistake when rendering. And that is exactly the principle I have argued for here. A tool that are used by a lot of people should not generate files that are not standard compliant and still claim that they are.

acidzebra · 08-20-2009, 06:18 AM

Quote:

Originally Posted by tompe

But the principle on the web has always been to be strict and correct in what you generate

You must be posting to a different internet than me. The principle of many HTML-generating tools and programs seems to be "churn out any old slop and let the browsers worry about it".

That said, I agree with you in principle but not in practice

*yes, I know the difference between internet and WWW. I just don't care*

kovidgoyal · 08-20-2009, 10:39 AM

@Jellby: There's a lot more to valid XHTML than just closed tags. Closed tags are needed for valid XML (and calibre does close all tags). In fact I'm pretty sure calibre will take arbitrarily bad input HTML and return valid XML, though not valid XHTML.

JSWolf · 08-20-2009, 10:49 AM

Quote:

Originally Posted by pepak

Let me be brutally honest: Who cares? The fact is that strive as it may, EPUB as it is now is not an universal format and books in it must be modified for different implementations (e.g. Adobe DE as implemented on PRS-505 will not display international characters unless the EPUB is specifically modified, it will crash the Reader on certain CSS constructs [sadly, I can't remember what it was, but it "worked" quite dependably], etc.; other implementations have their own specific quirks that need to be addressed). Does it really matter if there is one more thing to fix?

Also, get back from the academic paradise to the real world. As Kovid said, most sources are not valid XHTML. The conversion process then must decide whether it will be strict or whether it will be usable (consider what would happen to a web browser which refused to render invalid pages!).

I get accented characters with ePub just fine on my 505 using Fontin (font family).

pepak · 08-20-2009, 12:09 PM

Quote:

Originally Posted by JSWolf

I get accented characters with ePub just fine on my 505 using Fontin (font family).

Exactly my point.

Chang · 10-30-2009, 09:04 AM

Seems that the creator of this thread didn't have answer for his/her problem and I didn't find any other threads from the whole forums with this topic.

When I run my epub file through the epubchecker, I get this warning:

ERROR: book.epub/OEBPS/content.opf(9): unfinished element

I checked every element in opf file and every one of them is closed. I also compared with other epub books (e.g. Project Gutenberg) and didn't find any significant difference. I exported this e-book from InDesign book file. What could be the problem? My ebook works perfectly in Digital Editions and Sony PRS-505 but I'm still curious to know what causes the error.

Abecedary · 10-30-2009, 09:13 AM

Quote:

Originally Posted by Chang

Seems that the creator of this thread didn't have answer for his/her problem and I didn't find any other threads from the whole forums with this topic.

When I run my epub file through the epubchecker, I get this warning:

ERROR: book.epub/OEBPS/content.opf(9): unfinished element

I checked every element in opf file and every one of them is closed. I also compared with other epub books (e.g. Project Gutenberg) and didn't find any significant difference. I exported this e-book from InDesign book file. What could be the problem? My ebook works perfectly in Digital Editions and Sony PRS-505 but I'm still curious to know what causes the error.

The (9) indicates that the problem is occurring in the 9th line of the content.opf file. Double-check that particular element. I'm sure you'll find some minor error with it (bad reference maybe).

Chang · 10-30-2009, 09:19 AM

What founds from my 9th line is:

</metadata>

and it's closing tag for line 3:

<metadata xmlns:dc="http://purl.org/dc/elements/1.1/">

Can't see what's the problem. Inside this element are those <dc:title>blaa</dc:title> and all of them are properly tagged as well. Could it be that I'm missing some important attributes in my metadata element?

Abecedary · 10-30-2009, 09:50 AM

Hmm. Try deleting that line and rekeying it. There might be a weird line break character or something that the validator doesn't like. Otherwise, without seeing the file I don't know. Is it something you can upload here?

pdurrant · 10-30-2009, 11:36 AM

Unfinished element doesn't indicate that the element is unclosed, but that it is missing required sub-elements or attributes.

You probably need to specify some more metadata to be valid.

Quote:

Originally Posted by Chang

Seems that the creator of this thread didn't have answer for his/her problem and I didn't find any other threads from the whole forums with this topic.

When I run my epub file through the epubchecker, I get this warning:

ERROR: book.epub/OEBPS/content.opf(9): unfinished element

I checked every element in opf file and every one of them is closed. I also compared with other epub books (e.g. Project Gutenberg) and didn't find any significant difference. I exported this e-book from InDesign book file. What could be the problem? My ebook works perfectly in Digital Editions and Sony PRS-505 but I'm still curious to know what causes the error.

GeoffC · 10-30-2009, 11:47 AM

Quote:

Originally Posted by Chang

What founds from my 9th line is:

</metadata>

and it's closing tag for line 3:

<metadata xmlns:dc="http://purl.org/dc/elements/1.1/">

Can't see what's the problem. Inside this element are those <dc:title>blaa</dc:title> and all of them are properly tagged as well. Could it be that I'm missing some important attributes in my metadata element?

I get similar errors on a cover image I have - like you there is nothing obviously wrong with the line in question.
Mind you, ADE will display it but only with the qualifier (this file may not display properly).
(Have same errors with books generated via Sigil - so one wonders about ePubcheck itself).

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
epubcheck, any software can pass its validiation check?	droople	ePub	96	08-16-2011 01:38 PM
epub - force a 2nd pass to improve structure detection ?	cybmole	Calibre	10	10-08-2010 01:00 AM
errors after epub check	ericp20	Sigil	2	06-13-2010 10:40 AM
epub check mentiones an error	NASCARaddicted	Calibre	13	02-15-2010 09:50 AM
ePub Corrections Check - Help.	nboshart	ePub	3	10-01-2009 04:25 PM

08-20-2009, 10:39 AM	#22
kovidgoyal creator of calibre Posts: 45,604 Karma: 28548974 Join Date: Oct 2006 Location: Mumbai, India Device: Various	@Jellby: There's a lot more to valid XHTML than just closed tags. Closed tags are needed for valid XML (and calibre does close all tags). In fact I'm pretty sure calibre will take arbitrarily bad input HTML and return valid XML, though not valid XHTML.

10-30-2009, 09:04 AM	#25
Chang Connoisseur Posts: 92 Karma: 50000 Join Date: Oct 2009 Device: none	Seems that the creator of this thread didn't have answer for his/her problem and I didn't find any other threads from the whole forums with this topic. When I run my epub file through the epubchecker, I get this warning: ERROR: book.epub/OEBPS/content.opf(9): unfinished element I checked every element in opf file and every one of them is closed. I also compared with other epub books (e.g. Project Gutenberg) and didn't find any significant difference. I exported this e-book from InDesign book file. What could be the problem? My ebook works perfectly in Digital Editions and Sony PRS-505 but I'm still curious to know what causes the error.

10-30-2009, 09:19 AM	#27
Chang Connoisseur Posts: 92 Karma: 50000 Join Date: Oct 2009 Device: none	What founds from my 9th line is: </metadata> and it's closing tag for line 3: <metadata xmlns:dc="http://purl.org/dc/elements/1.1/"> Can't see what's the problem. Inside this element are those <dc:title>blaa</dc:title> and all of them are properly tagged as well. Could it be that I'm missing some important attributes in my metadata element?

10-30-2009, 09:50 AM	#28
Abecedary Exwyzeeologist Posts: 535 Karma: 3261 Join Date: Jun 2009 Device: :PRS-505::iPod touch:	Hmm. Try deleting that line and rekeying it. There might be a weird line break character or something that the validator doesn't like. Otherwise, without seeing the file I don't know. Is it something you can upload here?