The wonders of predictive text

I don’t think I’ve talked about this before, if I am repeating myself then do forgive me, but this really does deserve some space on the web. It just amazes me sometimes how brilliantly flawed predictive text is. Although the algorithms have improved over the years (my phone now does predictive word completion too), it still isn’t fool-proof. Here is an example:

“Hiya! sorry not go, b good room”

That should have read “Hiya! sorry not in, b home soon”. The right keys would have been pressed, but the writer was in too much of a hurry, or too oblivious to the misinterpretation, that the wrong words end up being sent. Another one I was sent recently said “go if thanks” – I’ll leave you to work out what that was supposed to mean…

I propose an improvement to be grammar and syntax to be taught to mobile phones, so that it can go back over words and see if they make sense. Hopefully that way such mistakes as above would stop happening quite so often. I doubt it would ever be completely fool-proof, because fools are so ingenious. But if phones can surf the internet and manage a diary, I’m sure we must be able to teach them the basics of the English language!

One thought on “The wonders of predictive text

  1. I think it would be relatively easy for phones to remember commonly-used words. For example, every time I type my name, I have to change it from “Shill” to “Phill”. Surely it could remember the 1,000,000,000 times that I had to switch and on the odd occasion where I had to say “Shill” (if any?!) then I’d be able to switch. Similar for “in” and “go”, “body” and “andy” etc. Actually I think writing each letter separately would probably work out faster in the long run, but somehow I think predictive text ought to be faster… ah well!

Leave a Reply

Your email address will not be published. Required fields are marked *