problem with italics

[bobcdy]

I'm working on a 16th century English book with literally thousands of changes from normal to italic, and back, text. Although I realize that the <i>/</i> are depreciated tags all the browsers and ADE accept them, I think because <i> is a LOT SHORTER than <span class="italic"> and the more correct tags would be much more difficult to incorporate into the epub.

Sigil generally accepts the <i> tags but occasionally even v.4 rejects it with Tidy turned off, and Sigil insists on introducing at the start of the offending file a new sgc-1 tag and inserting sgc-1 with other changes in place of <i>, sometimes everywhere in the file. ARGH! I thought that v.4 was done with such annoying inconveniences! At that point I absolutely must replace the tags and other Sigil changes with the css correct <span class="italic"> and </span> within either the entire file or just the offending paragraph, even though in the other files of the epub the <i> tags are accepted by Sigil without any objection. Any attempt to reintroduce the <i> </i> tags at the specific location causing the problem is met with continuing resistance by Sigil. Of course in a typical epub I might only need a few italics and I prefer using the css approved tags, but when there are many, many italic locations, then the css tags clutter the document.

It's odd but encouraging in that often only a single location within the entire epub will trigger Sigil's response.

Is there any way to insure that Sigil will accept the <i> tags? If not is it possible for the new versions to make that change?

[JSWolf]

Replace <i></i> with <em></em>. That should solve your problem.

[bobcdy]

JSWolf - thanks! I'll remember to use <em> in the future.

It seems to me though that the <em> tag, which I suppose stands for 'emphasis', is a poor choice of tags for 'italics'. Italics are not always used for emphasis, especially in older books, and other styles are also used for emphasis, such as all capitals, bold-face, or even smallcaps. In the book I'm working on italics are used for Latin and other foreign language texts, and for English poems, in addition to what I suppose is emphasis from context of the italics. I guess I don't understand precisely a well used tag such as <i> should be 'depreciated'; if <i> isn't acceptable, why not use <it> rather than <em>? After all the font-style is still 'italics'.

Bob

[charleski]

It would be helpful to see an example, because this shouldn't be happening at all with Tidy clicked off. If it's not something you can post publicly on the issue tracker, contact John or me and we can mark it private.

[bobcdy]

As I mentioned in my original post, it was distinctly odd that only a single location where <i> triggered Sigil's reaction; the text at that location certainly seemed no different than in several hundred other locations that Sigil ignored. With Tidy clicked off, however, there is still some tidy-like responses that create xhtml corrections. I never have tried to note these responses, or to experiment to see if deleting the entire section around the offending <i> would take care of the problem.

For the most part, though, I really appreciate Sigil's little notice of an error that usually can be manually corrected easily.

[Toxaris]

The only times I say that behavior is when the complete paragraphs is in italics. So, for example:
<p><i>This line in italic.</i></p>

That example will be converted by Sigil to a separate class in the internal stylesheet.

[Jellby]

By the way, <i> is deprecated (or disencouraged) because rather than specifying the style in the HTML you are supposed to specify why it needs a different style, and then specify the style in CSS. Something like this:

And then I found <em>him</em> on board the <span class="ship">Furious</span>, during the <span xml:lang="fr">hors-d'œuvres</span>.

even if all of them will be shown in italics:

em, span.ship, :lang(fr) { font-style: italics; }

they might be different in text-to-speech, for instance.

P.S. One might need more nesting levels:

em em, em span.ship, em :lang(fr) { font-style: normal; }
:lang(fr) em, :lang(fr) span.ship { font-style: normal; }

[bobcdy]

Jelby,
I appreciate your very well explained significance of the CSS tags - I've developed enough skill based on work on epubs pretty well (I hope) from working with html3 (and earlier versions), from study using Sigil and ADE of many examples given on the web including those of many MR epubs, and much trial and error experimentation, but I really don't understand much about the fundamental bases for css/xhtml or many details necessary for more complex css work.

Again, though, I'd like to mention that starting with an ocr scan of a very complex Old English book with thousands of italic-normal transitions of various types, trying to assign css tags in the proper way to describe each of the many different uses of italics would be extraordinarily complicated and difficult, and the result would be a multiplicity of tags to reflect Latin and French text, English poems, Latin poems, book titles, names of countries and towns, authors, etc. etc., and of course each tag would need to be pretty lengthy. Also each tag designating the use of italic text would probably be different if another epub creator was working with the same text. Why not then just use <i> as a style designation because the only other almost as simple tag <em> would give a false impression of representing emphasized text?

Bob

[Jellby]

In those cases, I "misuse" <em> (and <strong>) all the time, probably because I've never really seen or used a text-to-speech app (which could make a difference between emphasis and just italics). But yes, <i> (and <b>) would be more neutral, although apparently Tidy likes to fiddle with those.

[charleski]

Originally Posted by Toxaris

The only times I say that behavior is when the complete paragraphs is in italics. So, for example:
<p><i>This line in italic.</i></p>

That example will be converted by Sigil to a separate class in the internal stylesheet.

I've tested several possible cases, and there's no style conversion at all as long as Tidy is turned off all the time. See the attached ePub: you can open it, edit it, switch view and save it without any of the italics tags being converted. As soon as Tidy is turned on, though, any change of view will trigger style reorganisation.

So if anyone sees italic tags being converted and is sure they had Tidy off the whole time, please post an example.

[Toxaris]

I should have mentioned that it is a Tidy-effect. Sigil will not do it when Tidy is off.

With Tidy on, strike-through and underline (with <s></s> and <u></u>) will always be converted to an internal stylesheet.

[Jellby]

Originally Posted by Toxaris

With Tidy on, strike-through and underline (with <s></s> and <u></u>) will always be converted to an internal stylesheet.

Which is good, since they are not in the list of tags supported in ePUB.

[Toxaris]

According to the specs both underline and line-through should be supported according to the CSS2 specs. So, perhaps not the corresponding tags, but I have not found any reference to the tags.