| What I take away from this is: communicating heuristic-based behaviour to users is damn hard. The irony here is that the beautiful name, "Hemmingway", leads people to paste bloody _literature_ into a style linter! You wouldn't even be able to hire an unqualified human to give you meaningful style feedback. Nobody writing a novel should want this sort of thing. That's stupid. It's not for writing a novel. Verbal art is always going to break all the rules and invert expectations. This would be like checking visual design heuristics against paintings or art photographs. Surprise surprise, they break your design rules! Calling the app "Hemmingway" brands the app beautifully, and gets people engaged...unfortunately the engagement is jumping all over it for something it was never supposed to do. I think there's good potential for style linting, it's a really under-explored area. And I think probably the app's rules, as implemented in this alpha, aren't that much up to scratch. You'd at least want to run a POS tagger, and probably a parser, to give better feedback. So long as the heuristics are _correlated_ with common style problems, you can get some use out of the app. But apparently that's a difficult story to tell. |
On the passive voice, see: http://languagelog.ldc.upenn.edu/nll/?p=2922 and many, many more articles on that blog.
On adverbs:
http://chronicle.com/blogs/linguafranca/2013/02/20/being-an-...
and on adverb hunting via software:
http://itre.cis.upenn.edu/~myl/languagelog/archives/004271.h...
and on this app:
http://languagelog.ldc.upenn.edu/nll/?p=10416