[Plugin] ePub3-itizer - epub3 output plugin for Sigil

[Julie34]

Thank You !

[Julie34]

Hi,
A small thing on an epub a bit complex (132 books, several levels...) When epub2 it pass Epubcheck and Flight Crew. Result with Epub3itizer :

nav.xhtml 47 1792 Col: 47: WARNING(NAV-011): nav 'toc' doit être dans l'ordre de lecture; la cible du lien 'OPS/book_0018.xhtml#auto-toc1' est avant la cible du lien précédent dans l'ordre spine.
nav.xhtml 47 1792 Col: 47: INFO[INF-0O1]: The previous rule is under review and its severity may change in a future release. See the discussion at https://github.com/w3c/publ-epub-revision/issues/1283

Not important as explained...

Thanks anyway for this plug-in which works fine.
Julie

[KevinH]

Since the nav is being created for the first time, its complete ordering is taken from the existing NCX while may or may not follow the OPF spine order as that was not a requirement under epub2 although it tends to follow naturally from the playorder attributes typically used in the NCX.

So before running ePub3-itizer you may want to verify that the NCX playorder matches the spine order to be safest.

Hope this helps.

[Julie34]

Thanks for the explanation ! Helpful !
The problem is that there is the chapter's spine (level 1 titles) but another "stanza" creating books too (the text is issued from oral tradition's novels) and not part of the ncx spine (and concerning more books than the those issued from chapters). The first volume had no problem but it was luck, i suppose. I'll try to look the books order and the 18th book.

[Mister L]

Would it be possible to generate a new uuid when updating, if the existing epub2 uses an unsupported format? Example, a Project Gutenberg epub uses a url for the id. When I epub3itize the file, the id is not changed.

More details here:

https://www.mobileread.com/forums/sh...4&postcount=37

[DiapDealer]

Just leaving this here because it seems somewhat relevant:

http://idpf.org/forum/topic-1729

[Mister L]

Originally Posted by DiapDealer

Just leaving this here because it seems somewhat relevant:

http://idpf.org/forum/topic-1729

Thanks, very interesting.

The more I think about this the more it seems to me that a new uuid should be generated as a matter of course, because converting from epub2 to epub3 alone is a big enough change that, I think, it cannot be considered the "same" file anymore.

[DiapDealer]

Whereas I don't think the id should be changed. There's an argument for the existing id to be made to validate within the new format's rules, but that might not be possible given all the schemes that epub2 allows. However in my opinion, id-changing decisions should always be firmly in the content creator's camp and not an algorithm's. I certainly wouldn't recommend that an epub2 to epub3 converter automatically generate a new book id by default. Let alone attempt to alter the scheme used.

Something like this would be better in its own ID-related Edit plugin, I think.

[Mister L]

Originally Posted by DiapDealer

Whereas I don't think the id should be changed. There's an argument for the existing id to be made to validate within the new format's rules, but that might not be possible given all the schemes that epub2 allows. However in my opinion, id-changing decisions should always be firmly in the content creator's camp and not an algorithm's. I certainly wouldn't recommend that an epub2 to epub3 converter automatically generate a new book id by default. Let alone attempt to alter the scheme used.

Something like this would be better in its own ID-related Edit plugin, I think.

Oh, really? Why? To me, the unique ID serves to identify a particular file, meaning not only the book text but also all the meta-data and other info not visible to the reader. So I suppose the question is how many changes can you make to a file and still say it is the same file (identified by the same unique ID)?

Or maybe the question is, does the uuid serve to identify a file in a particular state (meaning, a specific text, with specific meta-data, specific format, specific structure...) like an ISBN identifies a specific edition of a book, or is it meant to be able to track the evolution of a file however many times it's modified? But then you could end up with 100 radically different files all with the same "unique" ID, that doesn't seem logical to me...

If I change the format from epub2 to epub3, at the very least some structural changes are made and the meta-data changes, but presumably other significant changes will be made as well (in my case for example, I will use new CSS including some elements not used in epub2, the file will be divided differently, the internal structure will change, some text may be added...)... If (hypothetically) I referenced a file by its ID expecting to get an epub2 with a certain number of characteristics, but the file in question turned out to be an epub3 with a different internal structure, styles, maybe some extra text, I would be surprised; if I were expecting the epub3 file with the extra text, and got the epub2 instead, I'd be disappointed. Plus there is theoretically the question of compatibility... for the moment backwards compatibility is fairly good but it will likely become more of a problem as the format evolves; if you have a file that started out as an epub2 but then it gets upgraded to an epub 3.1 and then later to an epub (say) 5, is it still logical to identify it as the same original file?

I may be overthinking this, in which case I apologise, but I have been turning it around in my head since yesterday and it is kind of an interesting philosophical question to be honest (like the story of the carpenter who says "I love my hammer, I've had it for 50 years and it will last 150 more; I replaced the handle 3 times and the head twice and it's never failed me, they don't make quality like that anymore.")

[DiapDealer]

For me, it's not a matter of what you, or I (or anyone) think should warrant an ID change. It's the simple fact that I don't want code making that decision.

I'm fine with a button, or plugin that allows the user to consciously say, "New id please." (or new id scheme/refine for that matter). I'm just not fine with it automatically happening by default with every conversion from epub2 to epub3. The decision to generate a new, unique book id should not be tied to other processes, in my opinion.

[Mister L]

Originally Posted by DiapDealer

For me, it's not a matter of what you, or I (or anyone) think should warrant an ID change. It's the simple fact that I don't want code making that decision.

I'm fine with a button, or plugin that allows the user to consciously say, "New id please." (or new id scheme/refine for that matter). I'm just not fine with it automatically happening by default with every conversion from epub2 to epub3. The decision to generate a new, unique book id should not be tied to other processes, in my opinion.

Ha okay, well if it's down to a question of code then you are certainly better placed than me to say if it's a good idea or not. Why is this case different to creating a new empty epub file (which comes with a fresh ID)? (Sincere question, I am really interested).

An option to explicitly request it would be a good compromise in my opinion.

[DiapDealer]

Originally Posted by Mister L

Why is this case different to creating a new empty epub file (which comes with a fresh ID)? (Sincere question, I am really.

Because an epub HAS to have an id. There's a big difference between automatically providing something that didn't exist (and must), and automatically changing something that already did. The former can do no harm. The latter could cause issues. Think of someone converting a book (their book. one that they're going to upload and sell) that already had an isbn number for an id. Would they even notice that their registered isbn had been wiped out in favor of a uuid?

[DNSB]

AFAIR, unless you de-obfuscate the fonts before changing the identifier, you can also run into issues with font obfuscation which uses the dc:identifier.

[DiapDealer]

Forgot about that. That's entirely true as well.

[Mister L]

Originally Posted by DiapDealer

Because an epub HAS to have an id. There's a big difference between automatically providing something that didn't exist (and must), and automatically changing something that already did. The former can do no harm. The latter could cause issues. Think of someone converting a book (their book. one that they're going to upload and sell) that already had an isbn number for an id. Would they even notice that their registered isbn had been wiped out in favor of a uuid?

Originally Posted by DNSB

AFAIR, unless you de-obfuscate the fonts before changing the identifier, you can also run into issues with font obfuscation which uses the dc:identifier.

Ha okay, yes I see what you (both) mean now. Thanks for explaining. I do think it would be good to include the option (even if it's opt-in rather than opt-out) because I also think there is a case to made that a uuid should not stay the same if significant changes have been made to the file.