Do you think trace typing (as provided by apps like Swype and SwiftKey, and now integrated in many smartphones UI) can replace touch typing? Any data on typing speed with this new input method?

I was thinking that it would be interesting to have a flat PC keyboard on which you can combine both typing methods for even faster typing speeds?!!

  • 4
    If there are studies out there I wouldn't necessarily extrapolate them to full-size keyboard usage. A big desktop keyboard sized touch pad may not be suitable for such usage.
    – JonW
    Mar 9, 2015 at 14:02
  • True, but it doesn't have to be direct extrapolation. Maybe a new typing method can evolve in which you combine touch with tracing, saving the time of finger lifts between keystrokes.
    – jak123
    Mar 9, 2015 at 14:06
  • 1
    maybe, and it's certainly something worth considering. Keyboards have been around for so long without much innovation, unlike most other interaction interface systems. Whether swype style interactions can help move the technology and experiences onwards is an interesting idea.
    – JonW
    Mar 9, 2015 at 14:09

4 Answers 4


Since you asked for data, the following answer on skeptics stackexchange has good references to two studies which conclude that there is no net performance advantage to using swype over conventional touch keyboards:


enter image description here

The two studies are here:

There have been some attempts to create hybrid swipe/touch keyboards. One notable effort which grabbed headlines last year (circa Jan 2014) was hipjot:


I don't think a full size swype keyboard would be successful. You can see for yourself using a low-fidelity prototype: just print out a full size keyboard on a sheet of paper (eg this) and try swiping across it. You will see that the distance your fingers need to travel is much greater, which renders the keyboard much slower than the standard touch typing interaction.

Btw it's interesting that the studies show a bias in perception towards swype. It's an interesting UX point to ponder. I believe the swipe interface provides a greater illusion of speed for users because there is tactile feedback as the user slides between keys, even though performance yields similar results and air provides less resistance than glass when moving between keys.

  • 3
    Interesting. Subjectively, it seems much better.
    – user31143
    Mar 9, 2015 at 15:07
  • 1
    I used gesture typing almost exclusively on my Nexus 7. (Tapping out non-standard words.) I miss it now that I've got an iPhone. And yes, my perception was that it was faster. Maybe it was just easier. Mar 9, 2015 at 15:26
  • 1
    @tohster, there is a perception-reality gap, but there also can be a research-reality gap. So I'm not 100% convinced yet. The most recent version of Swype I had was terrible at guessing words. I downloaded Google keyboard and did much better with that. So, the technology could be beneficial, but not a particular implementation of it.
    – user31143
    Mar 9, 2015 at 15:29
  • 3
    Interesting read in the links: "However, our subjects all preferred using Swype over ScreenDoors [the on-screen keyboard]"
    – Jason A.
    Mar 9, 2015 at 15:35
  • 4
    I notice, though, that the typing rates for either are in the 10-14 WPM range, which is perhaps a third of what 'average' computer users do on a normal keyboard, and less than 20% of the rate a good typist can acheive.
    – jamesqf
    Mar 9, 2015 at 17:47

The wikipedia page on Swipe states:

The creators of Swype predict that users will achieve over 50 words per minute, with the chief technical officer (CTO) and founder Cliff Kushler claiming to have reached 55 words per minute.

This appears to be based on company claims rather than actual evidence. However, even if we take these claims at face value, this speed is at the low end of what a professional typist can do:

An average professional typist types usually in speeds of 50 to 80 wpm, while some positions can require 80 to 95 (usually the minimum required for dispatch positions and other time-sensitive typing jobs), and some advanced typists work at speeds above 120 wpm.

Several people have managed to type more than 200 WPM.

And this is comparing a "dumb" keyboard to technology that enhances accuracy with software. Keyboard entry can be even faster when combined with similar software technology. For example, some stenotypists can type over 300 WPM and >200 WPM is the norm.

It is possible that trace typing speeds will be competitive for non-experts. However, I doubt they could ever compete with keyboards for professional typing. The action of moving one's finger across the screen requires greater precision than typing with keys, and I think this is a physical limitation that won't be overcome.

The technology is a great step forward for text entry on a phone, though (update: maybe in my subjective opinion only).

  • 2
    Are you really comparing swiping on a phone with touch-typing on a keyboard?
    – Pieter B
    Mar 10, 2015 at 10:51
  • 1
    @PieterB, yes. That is how I interpreted the question...maybe wrongly. After re-reading it, I'm still not sure. It does refer to a computer keyboard, and this is traditionally what "touch typing" means.
    – user31143
    Mar 10, 2015 at 11:19
  • 1
    For clarity, pushing letters on a touch screen is not touch typing. Touch typing is what I'm doing right now. I'm typing on a keyboard without looking at the keys.
    – Bill Dagg
    Mar 20, 2015 at 20:59

Trace typing isn't new. It was invented over a decade ago and was known as IBM SHARK.

There are several crucial differences between current smartphone keyboards and the original concept:

  • SHARK used a non-standard keyboard layout called ATOMIK. The purpose of this layout was to greatly reduce the number of mis-recognitions by making the shapes of common words more distinct. For example, "is" and "its" were traced very differently, unlike when using the QWERTY layout.
  • SHARK was meant to be used with a stylus, which is arguably more dexterous to manipulate than a finger and lends itself to not being lifting.

The ultimate point was not to let the user connect letters without lifting a finger, but to let the user write in "shorthand." Ie, to not think about letters at all but let the word's shape be memorized and written as a sort of hieroglyph.

Shorthand (hieroglyph-like) handwriting is known to be a very fast alternative to standard handwriting and typing, allowing speeds in excess of 100 WPM, so the potential for a computerized version is large.

I believe that by embracing SHARK's two advantages, using powerful algorithms, and anticipating sufficient practice on a user's part (which, of course, is unthinkable these days) can lead to significant typing speed gains over a tap touchscreen keyboard. However, I'm not sure if it would be faster than a full physical keyboard, especially when used for non-lexical input.

  • 1
    Please share where you are getting the 100WPM from. The co-inventor of SHARK, Per-Ola Kristensson, estimated expert-level performance at 45WPM. This was arrived at with a good degree of both empirical and theoretical study in an academic research paper: citeseerx.ist.psu.edu/viewdoc/…
    – tohster
    Mar 9, 2015 at 23:20
  • 2
    @tohster >100 WPM is in reference to shorthand handwriting. Mar 9, 2015 at 23:42

Trace typing cannot replace key-based entry methods because it does not have the capabilities that key-based methods do.

In order to provide a smooth input experience, trace typing is generous in converting your swipes to words rather than precise. Thus it represents a "best guess" as to what you intended to type based on your action and on the words it thinks you may be trying to enter. This is usually beneficial but if I want to enter a word that is not in the dictionary it suddenly becomes almost entirely useless. To enter new words you need to switch to key-by-key entry.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.