Wikipedia-style citations in web writing

Scott Rosenberg has just made the suggestion of implementing Wikipedia-style versioning for news stories. Publicly accessible version histories, he reckons, would render post-publication edits transparent and increase accountability. This sounds sensible enough.

It also reminds me of something else that Wikipedia does very well and that, as far as I’m aware, isn’t supported in today’s crop of web writing applications: citation. I was going to research the matter a bit more thoroughly before saying anything about it, but Scott has asked me to elaborate on a late-at-night, spur-of-the-moment remark I made, so here goes.

Wikipedia’s editors are required to cite reliable sources for everything they put into an article. This practice is supported by citation templates that are placed into the body of a Wikipedia article’s source code and that pass through a rendering engine when an edit is saved, resulting in nicely formatted citations, complete with a references list at the bottom of the article. What’s especially attractive about Wikipedia’s handling of citations: the finished HTML page will have hyperlinks that go back and forth between the in-line citation and the bibliographic data in the references list. In a word processor, neither “Save as HTML” nor “Save as PDF” will do that.

On this site, I’ve emulated Wikipedia-style citation in a few pieces about the history of blogging:

The gory details

I do academic research. As things stand, the “primary” form of my textual output is expected to be in word-processed documents, and I do most of my academic writing and note taking in OOo Writer on a Linux machine, saving it in Open Document format (indexed by Beagle), exporting it to Word or PDF format as required. I also use the Zotero citation manager, which maintains a database of works I might cite. A plugin allows me to access this database with a few key-strokes whenever I want to insert a citation, and Zotero will yield up the required data and insert it into my document, correctly formatted in accordance with a chosen citation style, which currently happens to be APA more often than not.

Fine, that’s word processing sorted. But what if I want to share a paper or a shorter piece with citations on this site? I cold export to PDF, sure, but PDF’s real merits lie in pre-print. As a publishing format, PDF can’t compete with plain HTML. Save as HTML, then? This has resulted in atrocious code for as long as word processors exported to HTML. Neither method results in hyperlinks between the in-text citation and the bibliographic data in the references list.

So, just to embarrass myself about its awkwardness, here’s the workflow I’ve been using to create Wikipedia-style citations on this site. In each case, I saved the underlying word-processed document as plain text, then exported the document’s citations from Zotero as Wikipedia citation templates, inserted the templates manually into the text file, then put the resulting file through the Wikipedia page renderer; edited the output somewhat, then copied and pasted it into my content management system‘s posting textarea.

Yes, I know.

To get this process automated, I think I would have to talk to the citeproc-js folks, whose citation processor will ship with Zotero’s next big release. They could create Wikipedia Source as an export format, with citation templates correctly in-lined in the running text. Then the Wikipedia source renderer would need to be available as a WordPress or Drupal plugin, so the Wikipedia Source output could be piped to it directly…

filed under: Uncategorised