Update: The latest release offers the option to save references directly to MLZ. To use the option, you may need to have the latest (very new) MLZ release installed, so it is worth updating that as well. Mainstream Zotero may not work — give it a try, let me know how you get on. The Ferret will avoid throwing duplicates when items are created from scraped cites, and parallel references in scraped data will be reflected in the Related tab of the parallel items in MLZ. Linked Data from unstructured plain text!
It’s been a busy few days since the Free Law Ferret plugin peeked out of its burrow. Short release cycles (sometimes several versions per day) became a habit with me while coding the citeproc-js citation processor: when thousands of users are affected by a known bug, it’s hard to avoid the temptation to squash it. I suspect that the Free Law Ferret has so far a much smaller audience, but hope springs eternal, and old habits die hard.
The flurry of ideas in the conclusion of the last post has continued to nag at my imagination, and I’ve racked up a string if plans in the project tracker. The most interesting bits in there so far are the direct item save option for MLZ, and proper resolver support. Feel free to add to the list.
Enough has changed in the past few days to justify a short list of improvements. If you took the Ferret out for a walk and found it less thrilling than expected, here are some morsels that might tempt you to install again and have another go:
- Display citation above search listing
- One of the things that most bothered me during early testing was the need to dig back through the original document to check whether cases shown in a search listing (from CourtListener) corresponded to an actual citation in the document. This is no longer necessary: citations applying to the case are listed in a fixed header above the display window of the search listing for immediate reference.
- Recognition of year-before-cite styles
- Digging round in Google Scholar, I found a surprising number of Federal judgments with citations in the form “Smith v. Jones, 9th Cir., 1978, 123 F.2d 456″. The Ferret now does a tolerably good job of picking up both the year and the court name from cites in this form, with correct identification of the party names.
- Citation samples wiki
- The repository on BitBucket now has an open wiki page with a short list of sample citations known to work, and another of citations known to fail in some way. If you run across cites that the parser handles poorly, feel free to add to the page.
- Improved citation parsing
- There have been quite a few small tweaks to the parser, resulting in significant improvements in overall recognition and accuracy. The parser will continue improving gradually over time.
- Please-wait widget
- Since large documents can take awhile to process, I’ve added a “progress meter” to show that it is actually doing something. The widget does not show actual progress, but if the parser breaks in some way, it will throw an error—if it says it’s working, you can trust that it is working on the document, and will eventually return.
That’s the story so far. If you try the Ferret and like it (or think you might, if this-or-that were fixed in it), send me a postcard. Seriously. My address is:
Faculty of Law
I’ve put in quite a bit of time over the past several years on Multilingual Zotero, citeproc-js and now the Free Law Ferret, and it’s nice to have a bit of physical evidence that the work is connecting with fellow tillers of the legal field. If you are moved to buy a copy of the MLZ book, Citations, Out of the Box (2013), I will of course be particularly chuffed!