Do you use predictive text? it can slow you down

read also, The government wants backdoor access to encrypted WhatsApp text. It can also be a backdoor for hackers

break it

It is interesting to consider the poor relationship between predictive text and typing performance. The idea makes sense: If the system can guess the word you want before you type it, it will save you time.

in my best a recent study On this topic, a colleague and I explored the conditions that determine whether predictive text is effective. We have combined some of these conditions, or parameters, to simulate a large number of different scenarios and therefore determine when predictive text is effective – and when it is not.

We built some fundamental parameters associated with predictive text performance into our simulations. The first is the average time it takes a user to hit a key on the keyboard (essentially a measure of their typing speed). We estimated it in 0.26 seconds based on previous research,

The second basic parameter is the average time it takes the user to view and select a predictive text suggestion. We fixed it at 0.45 sec, again with . on the basis of existing data,

Apart from these, there is a set of parameters which are less clear. These reflect the way the user engages with the predictive text – or their strategies, if you prefer. In our research, we looked at how different approaches to the two of these strategies affect the usefulness of predictive text.

The first is the minimum word length. This means that the user will only see predictions for words that exceed a certain length. You can only see predictions if you’re typing words longer than six letters apart – as these words require more effort to write and type. The horizontal axis in the visualization below shows the effect of changing the minimum length of a word before the user wants a word prediction from two letters to ten.

The second strategy, “type-then-look”, controls how many characters the user will type before seeing word predictions. For example, you can see suggestions only after you type the first three letters of a word. The intuition here is that the more letters you type, the more likely the prediction is correct. The vertical axis shows the effect of a user making changes to the type-then-look strategy, from seeing predictions of words before typing (zero) to seeing predictions after one letter, two letters, etc.

A final covert strategy, Persistence, captures how long the user will type and checks before discarding word predictions and only typing the word outright. While it would be practical to see how variation in persistence affects typing speed with predictive text, even with a computer model, there were limits to the amount of variable data points we could include.

So we set the persistence to five, which means that if there are no suitable suggestions after the user has typed five characters, they will complete the word without consulting the predictive text. Although we don’t have data on average persistence, it seems like a reasonable estimate.

What did we find?

Above the dashed line the net entry rate increases while below it, the predictive text slows the user down. Dark red indicates when predictive text is most effective; An improvement of two words per minute compared to not using predictive text. Blue is when it is least effective. Under certain conditions in our simulation, predictive text can slow down a user to eight words per minute.

The blue circle shows the optimal operating point, where you get the best results from the predictive text. This is when word predictions are only asked for words with at least six letters and the user looks at the word prediction after typing three letters.

Therefore, for the average user, predictive text performance is unlikely to improve. And even when that happens, it doesn’t save much time. The potential gain of a few words per minute is little compared to the potential time lost.

It would be interesting to study long-term predictive text usage and observe users’ strategies to verify that our assumptions from the model hold in practice. But our simulation reinforces the findings of previous human research: Predictive text might not be saving you time—and could be slowing you down.

per ola christensenProfessor of Interactive Systems Engineering Cambridge University

This article is republished from Conversation Under Creative Commons license. read the original article,

read also, WhatsApp introduces ‘missing messages’ that are automatically deleted after 7 days

subscribe our channel youtube And Wire