Text input battle royale yields surprising results
Thinker and obvious gadget enthusiast Phil Gyford decided to pit four of the handheld devices he’s collected over the years against one another in a test of their text input capabilities. As his controls, he also tested his time on his Macbook keyboard and plain old pen and paper, all with a 221 word passage he’d memorized.
The handhelds he tested were the Apple Newton MessagePad (1997), Palm Vx (1999), Palm Treo 650 (2004), and iPhone 3G (2009). Both the Newton and Palm Vx used handwritten input, while the other two used their built-in keyboard and virtual keyboard, respectively. His results, while obviously specific to him and his personal skills, were somewhat surprising.
No shock here: the handwritten methods filled out the bottom half of his test. Interestingly, the Newton and Palm Vx took considerably longer than pen on paper, demonstrating the road block that is handwriting recognition.
While some might take that as a mark against pen input in computing, it’s worth noting that his straight handwriting came in very close behind his keyboarding. I think Phil would be an ideal candidate for taking notes in ink on a Tablet PC without text conversion. Remember, the handwriting recognition engine in Windows XP/Vista/7 can search ink, so handwritten digital notes are more useful than paper even when you don’t convert to text.
Typing on a full keyboard was the winner by a clear margin, even though Phil admits to not being an accomplished touch typist. His iPhone use slightly edged his Treo use (two thumbs on his Treo, presumably the same on his iPhone), but both were close and within a minute of his pen and paper use. I might have to run a similar test and see how I do.
01/22/2010 at 11:54 am
In the future?
An electronic notebook that fits in your pocket, instant on,
no frills, and under $50.
01/22/2010 at 12:56 pm
Testing, like conducting opinion surveys, can be a tricky thing. In both cases, one needs to be careful to look beyond the “obvious,” and be prepared to include subtle factors that may have very significant effects. The Newton and Palm results may be (much) less about the “road block of handwriting recognition” but about the square area available to write in. I can’t speak for anyone else, but having owned a Palm Vx, I remember clearly that writing on its cramped little screen was a chore, whereas writing on a conventional (large) pad of paper was easy and “natural.” Such effects have an influence on speed.
Regarding the content to be recorded, one can easily argue that keyboards are optimized for the task of text recording, so why wouldn’t one expect it to be very fast? However, what if Gyford’s content had included a sketch or a simple figure? The keyboard wouldn’t have even finished.
I could go on, but you get the point. Most amateurs end up proving whatever they want; that’s why testing is best left to professionals!
01/22/2010 at 1:41 pm
I don’t think it fair to take a 10 year old technology and call it’s results a ‘road block.’ I mean, using a Newton to evaluate handwriting recognition? You might as well say an Xbox 360 runs Call of Duty better than Apples’s Lisa computer.
01/22/2010 at 1:56 pm
Steve: Agree to disagree. Phil emphasized (and I reiterated) the results are specific to him, so no one’s pushing them as definitive, nor did he seem to go into with a point to prove. The focus was narrowed to text input, not note-taking. And amateur tester or not, I think his familiarity with all the given devices lends weight to his findings.
Scott: 10 years old or present day, handwriting recognition is an obstacle. Unless you don’t care about errors, converting ink to text as you write adds a step vs. writing without text conversion. I do both on the same Tablet PC all the time, as well as convert ink to text after writing. I could never honestly say active recognition doesn’t require more time.
01/22/2010 at 1:59 pm
I still use handwriting on my Treo 650.. Did he not realise that it has handwriting capabilities in it still!!!!
I use “Graffiti Anywhere”
01/23/2010 at 3:53 pm
Did they give a reason for not including at least one actual tablet pc? not to mention one of the widely acclaimed models?
01/24/2010 at 12:49 pm
Wait, the Palms are stylus input, but not handwriting recognition.
That means you have to learn the specific strokes to create each letter, as opposed to the handwriting recognition in XP Tablet, Vista, Windows 7, RitePen, and Evernote.
That means in terms of speed, a keyboard would win for someone who can type, but handwriting recognition should be faster than stylus input or thumb typing. How portable the device should be a size and weight calculation, but each device’s usefulness in different contexts should vary with the individual and the UI of opening a noted, was well as what you actually need to port (looking at a spreadsheet on a Palm screen with DocumentsToGo might be fine, or it might be a nightmare).
I always thought Graffiti was much easier to use than Graffiti 2, or Jot, or whatever early Windows Mobile called their stylus input. I kept a journal, wrote course papers, took notes in classes on Palm devices from 1999 to 2002, and used one or another while teaching until I bought the Pre. I used to print directly from my Palms using the IR port and software the name of which slips my mind (BadBoy? PrintBoy?)
I’m faster with a keyboard, can use many more capabilities with my Tablet PCs, and use a Samsung Q1UP with Win7 everyday in classes I teach, throwing up on the projectors and marking up documents, web pages, PowerPoints, or whatever.