Sigil, the ePub editor

[Valloric]

Originally Posted by Hamlet53

One thing I've noticed about the use of TIDY; it often 'corrects' code in a way I don't want. My HTML editor of choice is HTML Kit and this includes TIDY as a plugin. However, it produces a side-by-side comparison of original code and corrected with changes highlighted.

When it works, Tidy usually doesn't mess up your code too much. It is also used internally in Sigil to keep the HTML source of the book in an XML format; this makes it much easier (and faster) to manipulate the source. You can just use an XML tool.

In 0.2.0, a lot of the performance gains have come from removing the regex manipulations and using an XML tool (like QDom). This along with the multi-threading work make opening and closing the TOC editor instantaneous in 0.2.0. It was very slow in 0.1.x. This could not have been achieved without something to keep the source as XML.

And since the user can always edit the source in the Code View, there has to some automated tool to correct any mistakes the user may have made that change the XML validity of the document. Sadly, the only error-correcting parser that can handle really ugly HTML and is in C or C++ is HTML Tidy.

So both of us will just have to live with it.

[Sabardeyn]

Originally Posted by Valloric

But really... everything in an epub is Unicode. There are things like & nbsp; that you want to see in the code, but other than those hard-to-distinguish characters, I really don't see the point of character references. So personally, I'm not that bothered by it.

Just a quick question... is their any negative effect from removing an &nbsp inserted into the text? I was trying to convert a PDF to epub and found several in my document. Generally near the use of inline italicized words and/or long sentences or paragraphs (cannot remember which). Removing them had no ill effect, but just thought I would ask after seeing you mention the character.

[Valloric]

Originally Posted by Sabardeyn

Just a quick question... is their any negative effect from removing an &nbsp inserted into the text? I was trying to convert a PDF to epub and found several in my document. Generally near the use of inline italicized words and/or long sentences or paragraphs (cannot remember which). Removing them had no ill effect, but just thought I would ask after seeing you mention the character.

A   is a named entity for a non-breaking space. The only reason why Sigil preserves this entity is because it's impossible to visually tell the difference between a normal space and a non-breaking one. There are other such problematic characters, like soft-hyphens etc.

If your book had lots of nbsps, they were used as "padding".

[Sabardeyn]

Padding seems to have been the situation. In a ~150 page book there were at least 150+ instances of &nbsp usage that I recall. It might have been as high as 200; re-importing into Sigil afterword showed no errors that I could see in the text / layout.

While I did not think their were any repercussions, I did want to take the opportunity to double check. Just in case.

[efithian]

I am trying to convert a book that I wrote into epub format for use with the iPad. I have the book on the web using html. There are 100 or so images in the book that I need in the epub format. I tried the first five pages with Sigil and the text came in fine. The images were the little blue boxes, indicating it didn't have a clue.

Part of my web code includes a base ref statement such as:

<BASE HREF="http://homepage.mac.com/efithian/Geometry/">

Each of the images has a reference such as:
<IMG SRC="P-05B.gif" ALIGN="BOTTOM">

I tried removed the HREF statement from the code, and removed the images from the page view, then inserted the images in the correct locations. That seems to work. Am I going through an unneccesary process?

[pdurrant]

Originally Posted by efithian

I am trying to convert a book that I wrote into epub format for use with the iPad. I have the book on the web using html. There are 100 or so images in the book that I need in the epub format. I tried the first five pages with Sigil and the text came in fine. The images were the little blue boxes, indicating it didn't have a clue.

Part of my web code includes a base ref statement such as:

<BASE HREF="http://homepage.mac.com/efithian/Geometry/">

Each of the images has a reference such as:
<IMG SRC="P-05B.gif" ALIGN="BOTTOM">

I tried removed the HREF statement from the code, and removed the images from the page view, then inserted the images in the correct locations. That seems to work. Am I going through an unneccesary process?

Copy all the images off your iDisk into the same folder as your HTML file on your hard disk. Remove the <BASE...> statement from your HTML. Open the HTML and Sigil, and all your pictures should be there.

I don't think Sigil will retrieve images from the internet for you, even on your own web sites.

[Valloric]

Originally Posted by pdurrant

I don't think Sigil will retrieve images from the internet for you, even on your own web sites.

It won't, for security reasons.

[efithian]

Originally Posted by pdurrant

Copy all the images off your iDisk into the same folder as your HTML file on your hard disk. Remove the <BASE...> statement from your HTML. Open the HTML and Sigil, and all your pictures should be there.

I don't think Sigil will retrieve images from the internet for you, even on your own web sites.

I think that the problem is in the html file which has the <BASE...> statement in it. All of the images are in the same folder, but the html file isn't looking there. So that should work. Thanks.

[efithian]

I created an epub file from html, added images, and saved to disk. I placed this in iTunes and transferred to my iPad. The book shows up, and has all of the chapters showing, pretty much correctly formatted with images.

I open Sigil and load in the same file and all I can see is the first page. I open the same file in ADE and the entire book loads and is visible. I can no longer edit the file in Sigil. What is the problem?

MBP 10.6.3 Sigil 0.2.0b3.

[Valloric]

Originally Posted by efithian

I open Sigil and load in the same file and all I can see is the first page.

There's a Book Browser to the left. It has a Text folder. Open the other HTML files if you want to edit them.

[efithian]

Originally Posted by Valloric

There's a Book Browser to the left. It has a Text folder. Open the other HTML files if you want to edit them.

The Book Browser used to be there but has disappeared. Is there a prefs file I can remove to get it back? I tried reinstalling Sigil but it was still missing.

Just deleted both prefs files and the Book Browser is back. And the rest of the document is back. Problem solved.

Thanks for this application. Nothing else comes close.

[Valloric]

Originally Posted by efithian

Thanks for this application. Nothing else comes close.

Try the new 0.2.0RC1 that was uploaded two hours ago. It has a truckload of new features and bug fixes, including a menu entry for opening/closing the Book Browser. That will come in handy if it ever disappears on you again.

[efithian]

I was cleaning up an ePub doc I created with Sigil and I came across a web based ePub validation site. It is free to use and provides some valuable info about the ePub doc.

http://threepress.org/tools/

I also found a site that will publish an original ePub book on the iBook store for use with the iPad. As long as you have a validated ePub file, they will assign an ISBN number. Apple gets 30% of the ePub price, the company gets 20% of the remaining 70% and the author gets 80% of the 70%.

http://www.lulu.com/apple-ipad-publishing

[Jedai]

Originally Posted by Sabardeyn

Padding seems to have been the situation. In a ~150 page book there were at least 150+ instances of &nbsp usage that I recall. It might have been as high as 200; re-importing into Sigil afterword showed no errors that I could see in the text / layout.

While I did not think their were any repercussions, I did want to take the opportunity to double check. Just in case.

Ok, that was posted a long time ago already but I just wanted to say that there are legitimate usages for non-breaking spaces : mainly you want to use them between two words/symbols that must be separated by a space but never by a newline. Good usage includes thing like "yada&nbsp;:" or before a !, ? or ..., they also should be used between the first and last name of an author or things like that. Of course visually there's no difference in most case but since you're outputting a reflowable format, you may get wrong results on some displays if you don't use them.

To be honest I'm not sure they're used much, even in professional epub (though given the quality of many of those that may not be a criteria) but as a result you get some really misplaced newlines from time to time, so using them is one of those invisible little thing that improve the comfort of your reader (and of which he'll never ever be aware, life as a typesetter sucks).

(of course I'm not advocating their use for padding, you should normally try to avoid such abuse since it's unlikely to degrade as nicely as a CSS directive)

--
Jedaï

[Sadhu]

I've been trying to convert a book displays properly in Word and Acrobat, but when converted with Calibre and displayed in ePub (on the iPad) spaces incorrectly. I'm wondering if this is a case where I can use Sigil to find and fix the problem.

I began by removing all footers an d page numbers from the original document, and then removed all hard page breaks except for the page before a new chapter (there are four chapters, and three hard page breaks). What surprised me was the when proofing the iBook (ePub), not only didn't the chapters always break properly, but there were line breaks where they shouldn't be (they are correct in Word and the PDF). The line would just seem to break at random, and sometimes a space would even be inserted in that broken line. At other times, the line would remain in tact but instead of being a space where there should be one, there was no space, as if there was no hard line return. Very disheartened by this, since Word and Acrobat have no code to make visible to find what is causing all this havoc.

Have others had these kinds of conversion problems, and if so, what to do about them? Do you recommend Sigil to try to find them and edit them out? Can the Calibre ePub be edited successfully with Sigil?