I just got my last few tweaks into my Asciidoctor Packt backend. It was enough to drop the venerable AsciiDoc and replace it with Asciidoctor.
Most notably, I fixed the images so they would be centered, and text would not wrap around. I also centered the [[Layout]] note that always is placed below. It looks perfect!
I can say this: working with an asciidoctor backend is gobs simpler than that asciidoc backend. It was brittle, hard to decode, and most of the time I was editing a monolithic file. With asciidoctor backends, each logical fragment is split off into a separate file.
In each file, I was using slim, a very terse way of writing templates. Given that the output is FODT, i.e. XML, slim is pretty neat. You get to express XML structures without the overload of reading angled brackets or having to write the closing tag.
But none of this would have been possible without Dan Allen’s direct involvement. I tried to get started on my own, but it was a no-go until Dan stepped in and submitted a pull request with an initial cut of a backend along with some critical steps to use it. To top it off, we were able to work back and forth, either through more pull requests, or through a discussion thread, to resolve other critical issues. At the end of the day, Dan is one rock solid community builder.
This already opens the door to using options unavailable before hand. The biggest one is how I want to extract fragments of source files into my manuscript. There is still a pending pull request to slightly alter the section format of the manuscripts, but this is structural and doesn’t interfere with my writing.