Pandoc is written in Haskell (a language I’m trying to learn), and so I thought it would be fun an appropriate to give it a try. Happily, the Haskell environment is available on Windows and Linux (and I’d already installed it on both), so I installed it and got started.
As we all know, my editor of choice is Vim, so I fired it up and started a new document (let’s call it “test.markdown”). Initially I started with:
Brief Comments == Herein I describe my this document and the ...
Then I compiled this with the following pandoc command:
pandoc test.markdown -o test.pdf
Indeed this does produce a pdf! I was happy. It was, of course, formatted as if it had gone through a LaTeX stage; and indeed this is the case. Beautiful; “what more could I want!” I thought to myself.
Well, it turns out I shortly wanted an actual title for the pdf; the above produces only headings. So I soon found myself reading (probably with a little bit too much enthusiasm) the documentation on pandoc’s extended markdown format.
I noted with much happiness that the LaTeX-style math I included in the document was rendered appropriately, and you can do equation referencing in the typical way (just write the appropriate LaTeX command).
I was also very happy to learn that the correct way to start a document with a title is as follows:
% Title % Author % Brief Comments == Herein I describe my this document and the ...
(if you leave the line below Author blank, it will generate the date, otherwise you can simply write the appropriate data there). And I was also really overjoyed to find the section of bibtex support and bibliography style handling via CSL. In particular, the pandoc documentation will direct you to a github repository that contains an amazing number of bibliography styles.
From here I did as any good vim user would; I tried to find a vim plugin. I found one, called vim-pandoc. However, after a little bit of use on Windows, I noted that it was really slow in some aspects (mainly because it uses inline python in the plugin, instead of vim code). I noticed some other problems as well, so with the project being on github I decided to fork it: silky/vim-pandoc. My version will probably be in a little bit of flux over the next few days, but will hopefully stabilise shortly thereafter.
All-in-all using pandoc for the small report I had to write was successful, and I do hope to try it with future documents containing maths. In particular the markdown format also matches nicely with my other vim-based notes (so I could convert them if I decided it was appropriate) and it’s just plain easy to read and nice to use.
Another reason I was attracted to pandoc was because of the ability to output to slide formats (including beamer), so I’m really excited to give that a go.
And of course, this blog post itself was written in pandoc (ouput to .html)! So, maybe you will consider it the next time you need to write something!