As this year’s iPhone software update nears I’ve become a bit disappointed with Apple.
Despite bringing an innovative, comprehensive implementation of speech input technology to iOS13, the company is falling into a well-worn speech input trap around command wordings. There’s a straightforward fix that I was hoping would show up in a beta before the release, but this is the 11th hour and I’m not seeing it yet.
The trap is that speech input companies are all using some form of Joel Gould’s first attempt at speech input vocabulary. Joel Gould was the first architect of the Dragon speech input software. Although the command wordings he worked out made for a good start 25 years ago, users have come a long way since then in constructing much more efficient and intuitive command vocabularies that work well with today’s more accurate speech recognition software.
The key to tapping some 25 years of user speech input command wording expertise is to allow users to adjust, save and share wordings. Users also need to be able to turn off commands they don’t use, including synonymous wordings.
It’s important to allow the folks who use speech input every day to develop and share vocabularies. This is the way speech traditionally develops. It’s crowd sourced, and it grows, adjusting to circumstances including how accurately the speech is heard. It gains efficiencies through use.
Use speech input to do real work on your computer or device for a day and you’ll tire of saying Gould-era commands like “move to the end of the line”, or “move left 3 words”. It will become obvious that you need shorter phrases, like “Go End”, or “3 Before”. Short commands are more memorable and can be more easily combined into command phrases, making speech faster and easier on your voice than having to pause between separate commands.
I’ve thought about speech input in great detail – and have used it for all the work I do on a computer for many years. Others have too. We need one more thing from Apple – the ability to tap the same efficient command sets we use every day on our computers for our phones.
For more about my thinking around efficient speech input commands see Human-Machine Grammar.
Please join me in pestering Apple to add this important feature – the ability to adjust, save and share speech commands – to their otherwise excellent implementation of speech input on the iPhone.