Translating Sigil

[User_Z]

Hello.

What about the translation into various languages of the Sigil User's Guide?

Now I'm trying to finish assembling the next book, and after that I'll start translating the Guide. Given the circumstances, there is a lot of work to be done. If someone needs her.

[User_Z]

Hello.

The meaning of this advice is not entirely clear:
"If you are really keen, search for and replace code like &nbsp;, <br />, empty paragraphs <p>&nbsp;</p>, empty <div></div> tags, empty attributes like src="" , href="", or ="", change – to &ndash;, change ' to ', etc. – unless of course you want to keep them."
from the "Tips" section of the User's Guide.

Explain please.

[KevinH]

Many times early epub creators used nbsp in p tags and br to physically create whitespace to force a layout when using css is the more correct way to handle this. Things like space above, line spacing, and space below an image or paragraph are better set by using css so that screen readers for visually impaired people are not confused. The rule is "Keep Layout (css) and Content (text) separate".

That paragraph in the user guide is pointing out how you could use find and replace to "fix" these epubs that look like they came from the typewriter age of layout.

[Doitsu]

It basically says when you're done with the book, you should:

• search for &nbsp; entities and delete them, if they were accidentally added
• delete redundant html tags such as <p>&nbsp;</p> and <div></div>
• delete redundant empty attributes such as src="", href="" and all other empty ="" attributes
• replace characters such as an en dash (–) in epub2 books with the corresponding &ndash; HTML entity so that you can easily tell it apart from a hyphen*
• replace straight quotes (' ") with curly quotes (‘ “ ” ’)**

*You can't use named entities in epub3 books (you'll need to use &#8211; or &#x2013;); entities to be preserved, need be added via Preferences > Preserve Entities
**Users can automatically replace them with DiapDealer's PunctuationSmarten Sigil plugin

[User_Z]

Doitsu,

Many thanks.
So much clearer.

In my translation, most likely, I will replace the advice from the manual with your answer.


Excellent!

It was good, practical advice.

[KevinH]

Would love to see how one of the interfaces to Google Translate would do on our user guide.

For example here is a python program to translate text of an epub using Google Translate, and that allows custom translate word dictionaries that a user could create that would include proper names like Sigil, BookBrowser, Preview, CodeView, url, etc.

https://github.com/quantrancse/epub-translator

Might make a good plugin for Sigil that would at least jump start the translation process and hopefully make it easily editable in Sigil to get things close.

That said all of the user guide images would have to be redone to have the correct user interface language.

[User_Z]

Regarding what has been said:
"...include proper names like Sigil, BookBrowser, Preview, CodeView, url, etc."

I agree that "CodeView" is a proper name and I try not to translate such appellation and names. But you yourself in the text of the manual use "CodeView" - 2 times and "Code View" - 104 times; "BookBrowser" - 1 time, "Book Browser" - 83 times. Even in the title of the program window you have "Book Browser".
Therefore, "Code View" and "Book Browser" are already other.

Originally Posted by KevinH

That said all of the user guide images would have to be redone to have the correct user interface language.

Of course. This is understandable without explanation.

Therefore, a complete translation of the Guide will take quite a lot of time and effort.

[User_Z]

Originally Posted by KevinH

https://github.com/quantrancse/epub-translator

This program is complete chaos in translation, watch and forget.

Google Translate translates not bad, in general, when you need to quickly understand the general meaning. But in each specific case, much more tools, knowledge and understanding of the topic of translation are needed.

[KevinH]

Just to be clear as I have no history with translation software, are you saying there is no value at all in starting an epub translation using Google Translate?

[User_Z]

Originally Posted by KevinH

Just to be clear as I have no history with translation software, are you saying there is no value at all in starting an epub translation using Google Translate?

No. I say that I ran the Sigil User's Guide through the epub-translator program you suggested and got deplorable results.

[DiapDealer]

So bad you had to start from complete scratch rather than fixing what the program provided?

[User_Z]

Originally Posted by DiapDealer

So bad you had to start from complete scratch rather than fixing what the program provided?

I immediately started from scratch. It’s more convenient and familiar to me, because I don’t correct someone’s translation, but make my own. Of course, with the help of various tools, dictionaries, reference books and other things.

It's just that today I tried to translate the Manual with the program that KevinH recommended.
I have already written about the result of this experiment.
Translation is simply bad, from the word whatsoever.
Styles and formatting are broken in many places.

Let someone else fix this.

Yes, and one more thing.
To quickly check the results, I use the AZARDI reader.
So, this program could not open the automatically translated Manual due to errors.

If you're interested, here's what happened:

[DiapDealer]

I'm sure any structural errors that prevent Azardi from opening them could be corrected. Unless it's Azardi that's at fault, of course. There was a time I had to notify Azardi that it would not properly render Sigil created epubs that were spec-compliant in every way.

EDIT: Yes, the mimetype file and the META-INF folder are simply one-level too deep in the archive produced. That could be easily corrected (and then Azardi would--or should-- open it).

[KevinH]

Sigil had no trouble with it but that translation program unfortunately translates any style tags in head which really messes things up as it tries to translate css.

Given this translator is a python program that uses bs4, it could be made to skip over all style and script tags in head, preventing this error quite easily.

That python program does use Google Translate as its engine. It may just need a few tweaks.

The real question is if these minor nits can be fixed: would it create a good starting point?

I tried with German but my German is so bad I could not tell how "poor" the translated text was.

Unfortunately I do not know another language well enough to really gauge things.

[BeckyEbook]

Out of curiosity, I started translating the guide into Polish in the "epub-translator" program.
The result was tragic. Not only because the Polish language is specific by its very nature, but also because of the context-free thoughtlessness of Google Translator.

Translating simple sentences that weren't technical or didn't contain interface or e-publication vocabulary was fine, but the rest (which is most of the text) is a nightmare.
Many words and phrases, due to their ambiguity, were translated exactly the opposite of the meaning in Sigil (e.g. save as economize, case (as box, chest etc.). Some sentences were so weird that they were funny.

The current guide in English does contain some inconsistencies in the notation (e.g. filepath – translated as firepat(!), though there is no such word), but for someone who is at least partially familiar with the Sigil interface and even has a cursory knowledge of English – it is understandable. And even if you don't, you'll figure out what it's like to work at Sigil through trial and error.

If I were to translate the guide (unfortunately – due to lack of time, I can't do it), I would certainly prefer to do it from scratch, because correcting the automatic translation would take more time.

On the other hand, if such an automatic translation (even if it was incorrect) was displayed on an additional monitor, it could still speed up the work on the proper translation.