Kindlegen vs Calibre: Which one compiles faster and is more efficient? - Page 2

pdurrant · 02-05-2010, 07:15 AM

Quote:

Originally Posted by zdavatz

Ok, it takes about 4.5 Minutes to open the EPUB file in http://www.epubread.com/de/ do you think that is because of the compression, i.e. the file first has to be decompressed? Once the file is open in Firefox there is no problem and it is really fast! Any Feedback is highly welcome.

If your source file for the ePUB is a single 170MB HTML file, then the ePub will not work on many readers. The problme is that it will get stored as one large compressed chunk, and need to be completely decompressed to read it.

Mobipocket format automatically splits the file into chunks to be compressed, and only the current bit needs decompressing, so reading is quite quick.

You could try splitting the file up into lots of smaller HTML files (under 300KB I think is the reocmmended size for greatest compatibility). This should greatly reduce the time taken to open and display the epub.

user_none · 02-05-2010, 07:26 AM

Quote:

Originally Posted by pdurrant

If your source file for the ePUB is a single 170MB HTML file, then the ePub will not work on many readers. The problme is that it will get stored as one large compressed chunk, and need to be completely decompressed to read it.

Mobipocket format automatically splits the file into chunks to be compressed, and only the current bit needs decompressing, so reading is quite quick.

You could try splitting the file up into lots of smaller HTML files (under 300KB I think is the reocmmended size for greatest compatibility). This should greatly reduce the time taken to open and display the epub.

calibre automatically splits the document into smaller chucks when generating EPUBs.

zdavatz · 02-12-2010, 12:57 AM

Well we have about 3500 HTML files under 300KB and they are all linked in the .opf file. That obviosly takes time to open and parse.

pdurrant · 02-12-2010, 04:49 AM

Quote:

Originally Posted by zdavatz

Well we have about 3500 HTML files under 300KB and they are all linked in the .opf file. That obviosly takes time to open and parse.

I'd said it was a an ePub reader with some sub-optimal design decisions. There should be no need for it to decompress and parse everything before it starts displaying the book.

zdavatz · 02-19-2010, 02:30 AM

Well you can check the file here: http://admin.zenogentoo.zdavatz.oper...g.firefox.epub - let me know where the wrong design decisions are!

pdurrant · 02-19-2010, 05:52 PM

Quote:

Originally Posted by zdavatz

Well you can check the file here: http://admin.zenogentoo.zdavatz.oper...g.firefox.epub - let me know where the wrong design decisions are!

I'm sorry for the misunderstanding. It sounds like you've done sensible things with your ePub. I meant the software that was taking much too long to open your ePub had a wrong design decision.

Looking at your epub with epubcheck, I get over 8000 error and warnings reported. Checking them, here are things you ought to change:

(1) the id attribute if items in the opf is not allowed to start with a digit. (or several other charcaters). an id attribute value is an XML Name, and most conform to the Name specification: http://www.w3.org/TR/REC-xml/#NT-Name See especially

[4] NameStartChar ::= ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]
[4a] NameChar ::= NameStartChar | "-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]

Your id attribute values are mostly hexadecimal values, which often start with a digit.

Since id values are referenced from idref attributes, you have errors on idrefs as well. So fixing the id attributes will get rid of about 4,400 errors. You have a similar problem *unsurprisingly) with idref.

(2) You have some (two?) non-unique id attributes.

(3) The other 3600 errors are all of the form:
ERROR: /Users/pdurrant/Downloads/compendium_ch.oddb.org.firefox.epub/OPS/3500_Volmax.html(1): elements from namespace "" are not allowed
But I'm not sure exactly what's causing that. If might be the way your files are mixing xhtml and html, but I don't really have time to track it down.

Otherwise it looks to be in fairly good shape. It migh be that fixing the id attribute value problem with speed things up, I don't know.

zdavatz · 02-26-2010, 02:57 AM

Thanks for the hint! I did not know of epubcheck! We will try your ID trick. We will also give your other suggestions some time and thought.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Anyone using Kindlegen?	simonp	Kindle Formats	11	09-16-2011 05:17 AM
KindleGen 1.1	pdurrant	Kindle Formats	0	08-08-2010 06:18 AM
Lots of files combined into one - most efficient way?	Rasczak	Amazon Kindle	7	01-04-2009 02:03 PM
HELP -- Any efficient way to synchronize folders?	mr.giroro	iRex	27	03-27-2008 05:58 AM
3 Steps to Highly Efficient News Reading	Bob Russell	Lounge	0	03-14-2006 05:12 PM

02-12-2010, 12:57 AM	#18
zdavatz Connoisseur Posts: 64 Karma: 496222 Join Date: Dec 2009 Location: Zürich, Switzerland Device: Kindle, iPhone, HTC D2 not Sony.	Well we have about 3500 HTML files under 300KB and they are all linked in the .opf file. That obviosly takes time to open and parse.

02-19-2010, 02:30 AM	#20
zdavatz Connoisseur Posts: 64 Karma: 496222 Join Date: Dec 2009 Location: Zürich, Switzerland Device: Kindle, iPhone, HTC D2 not Sony.	Well you can check the file here: http://admin.zenogentoo.zdavatz.oper...g.firefox.epub - let me know where the wrong design decisions are!

02-26-2010, 02:57 AM	#22
zdavatz Connoisseur Posts: 64 Karma: 496222 Join Date: Dec 2009 Location: Zürich, Switzerland Device: Kindle, iPhone, HTC D2 not Sony.	Thanks for the hint! I did not know of epubcheck! We will try your ID trick. We will also give your other suggestions some time and thought.

Advert

Advert