That’s a lot of quotes in the title, but essentially it’s describing functionality requested by Brian Croxall (via Twitter): he wanted to have an embedded Cirrus tool that wouldn’t make it quite so easy to recover the full text (when you click on a word in Cirrus, it opens up a new window with the default skin, which includes a full-text reader). I’m not sure if Brian wanted to demonstrate the frustration of non-consumptive analysis (where you can’t retrieve the original text) or if he wanted to use a corpus that shouldn’t really be circulating, but we cooked something up for him.
We’ve added a parameter called
noOtherExport that can be appended to the end of a URL and that makes it a bit more difficult to recover the full text – something like this:
Now, if someone clicks on a word, a growl-like message will appear to indicate that new windows are disabled.
This parameter also disables an option in the export dialog box (that appears when you click on the diskette icon in the upper right-hand corner), which normally allows the user to generate “a URL for a different tool/skin and current data.”
Of course, a resourceful person can still access the full corpus by manually generating a URL with the corpus ID, but the new noOtherExport parameter forces a few extra steps.
The new Voyant architecture, currently under development, will allow finer-grained control for accessing documents and corpora.
One annoyance (bug) was that a selected stopword list would not persist when you were working in the standalone tool skin and the tool spawned a new window with a new skin. This is more common than it may sound, as any embedded tool (like Cirrus) could have a stopword list set but then the new skin wouldn’t inherit the stopword list (like when clicking on a word in Cirrus). The issue hasn’t been entirely resolved, but at least now when you’re in a tool skin and you select a stopword list and choose to apply it globally, any new window spawned should also have the same stopword list set. Thanks to Susan Brown for encouraging us to fix this.
Voyant Tools is fairly flexible for ingesting XML. You can define an XPath expression that points to various bits of metadata (title, author, etc.) as well as the body of the text. You can even define an XPath expression to divide a single XML file into several sub-documents (as one might want to do with several items in an RSS feed, for instance).
We have recently added an initial version of a special adapter for TEI documents that pre-defines these XPath expressions. You can use this feature by selecting the TEI option from the loading texts screen.
Alternatively, you can use this parameter in a URL such as this: http://voyant-tools.org/?inputFormat=TEI&input=http://www.dh2012.uni-hamburg.de/wp-content/uploads/2012/07/ab-321.xml&stopList=stop.en.taporware.txt