Thanks to the work of several volunteers and an anonymous helper at Nuance, the latest version of Dragon NaturallySpeaking (DPI 15) now works with NatLink and Dragonfly! I’ve been testing it for the last couple weeks and it works well with only a few minor issues to work around.
Although I wouldn’t call this a must-have upgrade from Dragon 14, it is a step up in a couple ways. It seems noticeably faster and more accurate than previous versions. Without an automated test setup this is difficult to measure, but I was able to achieve 99% accuracy without trying too hard on a sample passage with a brand-new profile that hadn’t been trained, which matches my previous best on a trained profile. The new profile also has the “Speed vs. Accuracy” slider at “Fastest Response”, which is the new default. Dragon 15 also has better support for formatting (e.g. “john smith at gmail dot com” –> “firstname.lastname@example.org”). You can read the full list of new features here. Note, however, that it still hasn’t fixed the rather annoying issue where natively-dictated text sometimes contains randomly repeated characters, nor does it bring back support for editing of recent utterances in unsupported applications, which has been absent since version 13.
Ready to upgrade? First, before upgrading, consider backing up your profile and exporting your vocabulary list. Dragon 15 will auto-upgrade profiles from 13 and 14 but not 12, so if you need to create a new profile you’ll be happy you exported your vocabulary.
After installing Dragon 15, I recommend upgrading to 15.6.1 before running it for the first time (I’ve had issues in the past when I did this later). It’s not a critical update but it fixes some minor usability issues so you might as well pick it up.
Next, you’ll need to install a recent version of NatLink (stable or experimental).
Finally, you may need to restructure your grammar slightly to get it working properly. Specifically, Dragon 15 has trouble with grammars with many different exported rules, causing some rules to behave as if disabled. This is easy enough to work around: just separate the rules into different grammars, and it’ll work fine.
Huge thanks to the volunteers that managed to get the latest version of Dragon working again with NatLink. If you’d like to donate to support their efforts, you can do so through PayPal.