Hello,

I often work on Pandoc filters, and I regularly look at Pandoc's native format.

Here is an example of the output in my terminal:

> pandoc -i myfile.md --to native

[Para [Image ("fig:myfigure",[],[]) [Str "Figure",Space,Str "1:",Space,Str "This",Space,Str "is",Space,Str "a",Space,Str "caption"] ("generated/other\\8842295943125512100.jpg","fig:")]

,Para [Str "As",Space,Str "you",Space,Str "can",Space,Str "see",Space,Str "in",Space,Str "fig.\160\&1,",Space,Str "pandoc-crossref",Space,Str "is",Space,Str "compatible",Space,Str "with",Space,Str "the",Space,Str "ouput",Space,Str "of",Space,Str "pandoc-pyplot."]]

It would be very useful to get a pretty-printing version of the above, e.g.:

[

    Para [

        Image 

            ("fig:myfigure",[],[]) 

            [Str "Figure",Space,Str "1:",Space,Str "This",Space,Str "is",Space,Str "a",Space,Str "caption"] 

            ("generated/other\\8842295943125512100.jpg","fig:")

         ],

    Para [

        Str "As",Space,Str "you",Space,Str "can",Space,Str "see",Space,Str "in",Space,Str "fig.\160\&1,",Space,Str "pandoc-crossref",Space,

        Str "is",Space,Str "compatible",Space,Str "with",Space,Str "the",Space,Str "ouput",Space,Str "of",Space, Str "pandoc-pyplot."

         ]

]

After a cursory look at the source in Text.Pandoc.Writers.Native, it would be possible to make a nice, indented, pretty version of native Haskell, given that WriterOptions type has the writerColumns attribute.

Is this something that the Pandoc community might be interested in? Should it be a different format than "native", called "pretty-native" or something?

I'm happy to put together a pull request.

Laurent