Why all the spans in rtf to epub conversions?

[Berzelius]

If I use Calibre to convert a rtf file (originally a docx file) to an epub one and then look at the code view of the epub file in Sigil I find it is full of what seem like superfluous span tags. A typical format is:

<span class="none1">text text text</span>

Quite often it will simply split a word, for example the word "to" is split here:

. . . it is beginning t</span><span class="none1">o mean . . .

I'd say that, without exception, every paragraph contains at least one example, often many more than one.

It doesn't seem to have any effect at all on the text as read, I'm just curious as to what is going on.

[kovidgoyal]

Those spans correspond to formatting instructions in the rtf file. Open your rtf in a text editor like notepad and you will see the same thing.

[Berzelius]

Thanks for the reply.That's another of life's little mysteries cleared up.

As you suggested, I had a look at an rtf file in Notepad and I was astonished by the sheer amount of formatting code. No wonder I'd always found rtf files so large!

Apart from increasing the size of the epub file and being visually annoying when looking at the code is there any disadvantage to all these 'spans'?

[kovidgoyal]

Some epub viewers may have trouble with lots of spans, though I dont know of any offhand.

[Berzelius]

I've just done a little experiment in Sigil and I was rather surprised to find that a bulk deletion of the opening span tags (they all seemed to be either class="none1" or class="none2") also removed the corresponding closing tags.

[FaceDeer]

Sigil automatically runs an HTML tidy process on save (and in a couple of other situations) that ensures that it never saves malformed HTML to disk. Cleaning up stray closing tags is a pretty safe thing for it to do.

I also frequently take advantage of it to add closing tags as well. For example, I sometimes do a search and replace to turn "<p>Chapter" into "<h2>Chapter", and then the tidy step turns the corresponding </p> tags into </h2> tags. A quick and dirty way to format chapter headers, though it's important to manually check to make sure it generates a nice table of contents to catch any glitches.

[Berzelius]

Thanks for the info. It's nice to know the reason why it works.