[davidfor]

Originally Posted by jackie_w

I'm late back to this thread but most of what I would have replied has already been said by @MGlitch and @DNSB.

I was very late to deciding to try sideloaded kepubs via KTE so I didn't follow the early discussions in its main thread. However, I'm pretty sure KTE's original aim was to try to create kepubs with internals exactly like what you'd get if you'd purchased from the Kobo store. I think it succeeded at this pretty well.

Yes, that was one of the original aims of the extended driver. As we did not really know what was happening, it was thought best to duplicate what came from the store. Especially if there were any problems, it could be because we did not follow the (unpublished) kepub standards.

Over time tinkerers like me investigate what value each of the several "kepubification" changes actually offer. Is it actually necessary? Can it be improved without breaking something else? For kobospans the answers are 'Yes' and a more cautious 'Yes for my books on my Kobos'. I edit all my books to some fairly strict standards and I don't use annotations so I'm not a typical user. Will users who just Buy-Sideload-Read (aka the majority) see enough improvement to make it worth the KTE devs' time to implement/support the change?

And that's were we stand. I have a copy of @jackie_w code and am thinking about it when I have time (@jackie_w: Sorry, I haven't gotten back to you.). I have to decide how practical it is to offer both versions of the kepubification. I will not be removing the current version that matches what Kobo does. If I do anything, it will be adding the single-KoboSpan per paragraph method as an option. Then it will be up to the actual owner of the driver to accept or not.