Quote:
Originally Posted by KevinH
My response was to your vehement argument that this is a bug. It is not. It is a change. A change induced by moving away from Tidy. If that bothers you, then please stay with the Tidy version of Sigil 0.8.6. Tidy will do lots of cleaning for you (up to and including losing large pieces of code). The entire point of the move away from Tidy was to give the ebook author more control over how the finished ebook looks and to greatly reduce the potential for the loss of text. In addition, Tidy is not html5 ready which makes it unsuitable for modern html5 and epub3.
|
We've stayed with 0.8.x for some time now. I've only in the last few days installed 0.9.x, as part of my interest in helping out on the docs. For commercial bookbuilding, I think that the end of the 0.8 series is a better choice, if you don't have significant XHTML/CSS and device experience.
Quote:
There are many issues that prevent an ebook from validating and Sigil never handled them all (you will still need to properly set the minimal metadata, etc). So all versions of Sigil expect the user to understand the product they are producing. If validation is important to you, then simply run epubcheck or flightcrew and fix the errors you have, just as before. If you want stronger auto-fixing, then use the Tidy version of Sigil. I don't know of any ebook reader who will fail to render the page if no title is present.
If there is demand, and if the new html5 Tidy is buildable/working, I may create a Tidy plugin.
That is all I can offer. Frankly comparing a missing title if and only if it was missing in the input in Sigil-0.9.x to the potential for lost text, stripped mathml, broken svg, and wholesale changes of Sigil-0.8.x's Tidy, I would go with Sigil-0.9.x. And fwiw, user-none agreed with me on the need to remove Tidy and Xerces.
Thanks for your feedback.
KevinH
|
I can tell you without blinking that we've had issues, more than once, with something stupid like "en" getting nuked from the metadata, and failing intake. And, of course, we find out AFTER we'd given it to the client. It happens all the time, particularly if you get slammed, and a bookmaker forgets to validate the ePUB. Unless a non-commercial (or, arguably, commercial; there are plenty of pretend-pikers out there claiming to know what they are doing, using those types of tools) bookmaker wants to use a tool like Jutoh or Scrivener's ePUB-output, there's always going to be the opportunity to fall flat on your face. {shrug}. ePUB simply isn't a Word file, and Sigil doesn't have all the "catch you if you fall" mechanisms built in that the more-end-user oriented programs do. Hell, it's how we all learn. Having seen half a book eaten by Tidy, (which, fortunately, we'd already saved--once burned, twice shy), I'm with Kevin on that one.
On a bright note, at least there's an ePUBcheck plugin now. I'm quite appreciate of that. Makes it easier and faster to find what you've screwed up. That's a happy thing right there.
Hitch