Some Statistics on the First Two Weeks

Posted by Jean Ichbiah , Mon, Dec 02, 2002, 18:57:09 Reply   Forum

Here are some statistics on these first two weeks. In the table below, "Alternative" refers to methods that all try to minimize the stylus travel: Atomik, Fitaly and FitalyStamp, MessageEaseST. Qwerty includes various implementations of onscreen Qwerty keyboards. Handwriting includes Graffiti and Transcriber.

Ranking By Methods

Method Time WPM Entries
Thumboards 48.25s 58.10 6
Alternative PPC 50.61s 55.6 7
Alternative Palm 51.93s 49.56 19
Qwerty 83.55s 29.86 3
Handwriting 85.32s 23.31 3

Of course this is not yet much data, but one trend already emerges, namely a dominance of thumboard and alternative methods over handwriting and qwerty. I will venture the following explanation: Tapping methods are faster than writing methods because less movement is required for each letter.

Yet Qwerty manages to do no better than handwriting because it forces too much travel between keys. In other words, there is more to do to produce a letter with handwriting but Qwerty forces more work between letters, which explains the draw. Thumboards appear to overcome the drawback of Qwerty by using two fingers, with the result that finger travel is cut by two.

Another interesting aspect of the first two weeks is that a number of people have submitted entries with more than one method:

One Person Multiple Methods

Method Time WPM Participant
Fitaly Palm 46.07s 52.09 Will Lau
Treo Thumboard 46.75s 51.33 Will Lau
Fitaly PPC 42.43s 56.56 Richard Owens
SPB Qwerty PPC 81.69s 29.38 Richard Owens
Transcriber PPC 156.46s 15.34 Richard Owens
FitalyStamp 40.67s 59.01 Patrick Yam
Atomik Palm 45.29s 52.99 Patrick Yam
MessagEaseST 57.67s 41.61 Patrick Yam
Fitaly PPC 45.29s 52.99 Palmsolo
Sidekick Thumboard 57.67s 41.61 Palmsolo

It is definitely too early to derive any conclusion on comparing method from these few examples. The only clear conclusion at this stage is that using a PDA for text entry is a very realistic proposal!

Jean Ichbiah


| Edit | Reply   Current page