A canonical export format with complete meta-data

I am currently trying to figure out what combination of rtf, html, mmd, mmd-html, mmd-latex, … I can make work for me. I am a newbie, very excited about using Scrivener and shaping my mental book model in this great tool, but also stumped with how much of that mental book model – specifically, meta-data at all levels of the document content and structure – is lost in the output formats that I can readily get. Even something that should be simple, like getting an assigned keyword for a (sub)section of a document into a style-able

, seems out of reach.

From the point of view of easily connecting it to xhtml, latex, powerpoint, … It would be really nice if Scrivener would provide a single canonical export format that captured all the structure, content, and meta-information that Scrivener knows about a project, in a standard form such as XML.

<document>
  <metadata>
     <settings>various scrivener settings for this doc</settings>
     <title>the title that appears in the binder</title>
     <keyword>k1</keyword>   <keyword>k2</keyword>
     <label>L1</label>   <status>S2</status>     <notes> ... </notes>
  </metadata>
  <text>content, in standard xhtml subset with <p>, <table>, etc ... </text>
  <binder_structure>
      <document>  <metadata>...<text>...</document>
      <document>  <metadata>...<text>...</document>
  </binder_structure>
</document>

I think this would be very useful for people to re-use or build many other useful bridges out of (and in to) Scrivener e.g. in the form of XSLT, perl, or Ruby scripts.

Thoughts?

Sophie

Just curious about others’ opinions either way on this one.

I know I would love to work entirely in Scrivener. Through the end of the project, and into the second edition :slight_smile: I don’t like the prospect of switching modes and making disconnected edits in a separate medium. To me, this means I have abandoned Scrivener past that point. If I have all of my scrivening-related information available in a clean exported form, I can automate the steps from there to polished output like Word or PDF. And never leave Scrivener.

Thoughts? — Sophie

Hi,

My thoughts are that I don’t really relish the prospect of doing an enormous amount of work to create my own XML format that would be of no use to anyone except those dabbling with converting it to other formats… Getting all of the text out would require lots of tags for all the formatting and would be equivalent to - or perhaps more difficult than - writing an XML format such as the .pages one. Ouch. So I’m afraid there are no plans for this. Thanks for the suggestion, though!

All the best,
Keith

You already generate xhtml, I think. This would add binder structure (either containing the xhtml, or cross-referenced to it), and the meta-data the tool knows about.

Actually, if I turn on subversion, is there any binary data or is it all rtfd + non-binary xml? That would be open enough, provided there was a way to work the rtfd into some xhtml-like form.

Thanks for listening (and responding!) — Sophie

If you turn on Subversion, the binder.scrivproj file becomes plain XML. However, I doubt it will be any use to you, as it is for internal uses and so it will appear gibberish - you won’t be able to get any meaningful information out of it. I don’t generate any XHTML. I use various built-in exporters to generate the various formats. What I was saying was that creating my own XML format with binder structure and meta-data information would be an incredible amount of work for very little gain (it would be of use to only a handful of people, and I wouldn’t be one of them :slight_smile: ), so I 'm afraid there are no plans for this.
All the best,
Keith

Word?? PDF?? Ugh!

Isn’t the polished output supposed to look something like this:

bokhög.jpg
And isn’t that out of most hands and a thing for the publisher/printer?

I produced a print-ready PDF (for lulu) directly from Scrivener, via MMD/LaTeX. Took a bit of messing about with the XSLT file to get the look I wanted, but I didn’t make any edits outside of Scrivener.

I guess I’m not understanding exactly what you’re looking for, since you asked for an export format with meta-data but now you seem to be asking for polished finished product, and polished finished product wouldn’t have meta-data, would it?

I am embedding quite a bit of semantic information about each part of my document into scrivener e.g. this here is an “example”, that there is a “warning” (e.g. using keywords). I would like this info, along with the text content, to be available via export. So that I can automate the “polishing” phase to either PDF (for print) or Word-2008 (for editors who want that).

Thanks – Sophie