Starting with EPuB, 4 questions (to start out with).

[eschwartz]

Nitpick: RAR/7Z are formats.
WinRAR and 7-Zip are programs that handle ZIP/RAR/7Z



Seriously, try out advzip -- I think you will be pleasantly surprised.
Zopfli is the new kid on the block, but it has quickly taken the crown for efficient ZIP compression. Beating out 7-Zip's DEFLATE algorithm as well as KZIP.

advzip.exe -z4 -i 500 ZIPFILES...
advzip.exe --recompress --shrink-insane --iter 500 ZIPFILES...

[BetterRed]

WinRAR can exclude specific files from being compressed e.g. mimetype. I've used it to 'reconstruct' EPUBs after pulling apart and editing xhtml file(s) with Notepad++.

BR

[ProDigit]

Thanks!
Kzip probably was the program I used years ago.
I'll try advzip.

[Toxaris]

Be aware that the gain is not as much as you might want. My program ePUBOptimizer takes a valid ePUB and tries to make it smaller by using better compression, reducing image size without altering the quality and subsets fonts (if available). Of the functions, the compression rate has the least impact.

[ProDigit]

Hi,
Me again,

I'm manually creating my ebook,
downloaded Sigil for the generation of a UUID number,

And now I'm going over the content.opf file, in search for where can I trim this file down?

I've read that even though a UUID is technically 'required' for EPuB 3, it can run without it.
Also, a lot of content, like writer, author, date of creation, and other stuff, I don't really want to display.

My question would be, if removing the following line would cause any harm to the ebook? (like make it unreadable in some readers?):

<dc:identifier id="BookID" opf:scheme="UUID">urn:uuid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx</dc:identifier>

[Toxaris]

Be aware that some metadata is mandatory. Omitting that will ensure the ePUB is not valid. It might still work, but no garantuee.
From the top of my mind, the following is mandatory: Author, Title, Language, Date and an identifier.

[ProDigit]

I've seen books without everything, but the identifier.

[Toxaris]

Originally Posted by ProDigit

I've seen books without everything, but the identifier.

Then they will not validate. The items I mentioned are mandatory, with the exception of the author and apparently. I was sure those where mandatory too.

From the specs:

At least one of each of Dublin Core title, identifier and language must be included in the metadata element. Dublin Core elements, like any other elements in the OPF Package Document, may have an id attribute specified. At least one Dublin Core identifier, which is referenced from the package unique-identifier attribute, must have an id specified.

Be aware that most readers and reading applications will expect the author metadata also present. They will use that small subset to fill the internal database and will make it easier to select and read the book.

[dgatwood]

Originally Posted by eschwartz

It is XML. The spaces don't matter.
whitespace in XML/HTML only matters inside an element, and then it is collapsed down to one space.

Not strictly true. In some elements, e.g. pre, whitespace is preserved. Also, space between inline elements is significant, as is whitespace around comments. (I know this because I was the guy who fixed the bug on that last one in libxml2's HTML parser many years ago.) And with a proper parser, you can add xml:space="preserve" to any arbitrary element to preserve whitespace within in.

But yes, that's roughly true in the general case.

[ProDigit]

If I change the 'content.opf' filename (eg: to c.opf) and change it in the container.xml file as well;
Will the ebook still be readable, or will it give errors?

Meaning, if any file linking to content.opf, has been modified to reflect the correct file name, will the ebook still work, or is the renaming of Content.opf a non-negotiable for EpuB 2 books?

[Jellby]

You can use whatever name you want, as long as it's properly specified in META-INF/container.xml (this one you can't change). If any reader has problems with this, it's the reader's fault (I haven't seen that so far).

[DaleDe]

Originally Posted by Toxaris

Be aware that some metadata is mandatory. Omitting that will ensure the ePUB is not valid. It might still work, but no garantuee.
From the top of my mind, the following is mandatory: Author, Title, Language, Date and an identifier.

Author is not mandatory since it may not be known. Should be present if known. Similarly the date is not required for the same reason. Title, Language, and unique identifier are required.

[Toxaris]

Originally Posted by DaleDe

Author is not mandatory since it may not be known. Should be present if known. Similarly the date is not required for the same reason. Title, Language, and unique identifier are required.

I had already corrected my answer in a later post. After checking for sure in the specs I was surprised to find this as well.

[Doitsu]

Originally Posted by DaleDe

Similarly the date is not required for the same reason.

BTW, epub3 books need to have a date modified metadata entry. For example:

    <meta property="dcterms:modified">2015-12-19T19:23:07Z</meta>

Otherwise you'll get the following ePubCheck error message:

Error while parsing file 'package dcterms:modified meta element must occur exactly once'.

[DaleDe]

Originally Posted by Doitsu

BTW, epub3 books need to have a date modified metadata entry. For example:

    <meta property="dcterms:modified">2015-12-19T19:23:07Z</meta>

Otherwise you'll get the following ePubCheck error message:

Error while parsing file 'package dcterms:modified meta element must occur exactly once'.

Interesting. The date property is ambiguous at best since it does not indicate the purpose of the date. The ePub 3 seems to have corrected this in requiring the current date as opposed to the many other dates such as copyright date or original publication date. But technically this property is not a date property since there is an actual date property defined.

Dale