bun, nice to see your enthusiastic style hasn't waned. The tip on how to get HWR faster was very helpful for me (and, yup, so counterintuitive I'd done just the opposite when I first tried the Tablet's HWR).

I second Underscore's request that you measure the respective times for the typing and writing of a paragraph. I did this with the weird circular keyboard (can't remember it's name) and abandoned its use.

I think one big problem for me with the HWR is the stylus, so I'm going to get a cheap pen-shaped stylus just to see how much it helps.

Good point, I always believe in numbers, I will do one this evening and do one next weekend, to see what is the difference speed/accuracy between HRW and OSK after 10 days 'forced' training.

Thanks for the suggestion. I would also encourage some of you try it too. One, after you try it, hopefully you would be buy into it, two, there will be an objective finding rather than dictated by bunanson one person opinion.

I would pick anything around 10 min typing, has to be a piece of new passage..... thanks for the good suggestions, Underscore.

Another note: See thread #1, there is two ways to change the settings. Please keep checking on and off to make sure it is the setting you like. For some weird reason, I found my setting 'mystically' changed to the slowest 5+. Just keep checking the setting if it misbehave...

bun

No biggie. I'm just very *n*l about it.

@bunanson: is there a way to just emulate Palm Grafitti (1\2, both are better in my book than maemo's HWR)? I mean, just take everything from it. The gestures, the input pad design, the timing, everything.

That was one of the things that Palm did just right way back even from the first Palm Pilot.
Fast, accurate, simple..

Here is my game plan, if everybody is doing the same, then we can group together the results and compare:
Attention:

I relaxed the rule a bit, it takes way too long. Instead of doing 5 paragraphes 2,3,4,5,6. I settled with only 2 paragraph, just paragraph 2 and 3.

Go to Google, news page, no, you are not searhing any news, just go to the news page, pick the 2nd story, whatever it will be, that will open a new page, pick paragraph 2,3 and type with a) HWR b) OSK letter poking c) thumb keyboard. So, the 2nd news story, start on the 2nd paragraph for a total of 2 paragraphes.

We measure both speed and accuracy. Everytime you have a mistake, it will be counted as 1, keep on typing and dont stop and YOU DO NOT GO BACK TO CORRECT IT, again, correction does not count. You count it at the end. We are NOT using shortcuts as it is difficult to compare from each other. So, please no shortcuts, we can always assume inclusion of shortcuts will be in the favor of HWR. Again, straight typing, no shortcuts.

You will end up with speed, how many minutes and accuracy, the number of letters right over wrong, space is counted as one letter.


I will do one later this evening. I am really curious about what I "perceived" vs what it really is,



bun

OK, and the same thing can fairly apply to portrait mode (except that finger-board is broken...). The concern that raises: won't it be substantially faster the 5th run through the text, on account of familiarity? (Or maybe I'm overestimating my writing speed, and the familiarity benefit actually won't be significant...)

But I'll give this a shot with my pseudoGraffitized HIM HWR, both keyboards, and portrait mode HWR and stylus-board. Maybe, if you get lucky, you'll get some results from Vista HWR, too (I've had it for about a week...).

Oh, and I assume no use of completion allowed? Else late runs would be favored for sure.

That is a very good idea. I would like you or whoever interested in palm GVM jump right in, go thru the game plan I just lay out, couple threads up, and we can ALL compare, how good is HWR in palm and how good is HWR in maemo. I DO NOT think it is a good idea for the same person to do both, as the HWR is very different. I speak for myself, I would be slow down if I have to relearn palm HWR and, in fact, it counter-learn my maemo HWR and further slow me down.

I am really really excited about this. With ALL this cheap talks, we now came very close to be able to actually measure it. And this would test, how one perceives vs what it really is.




bun

I practiced writing "pack my box with five dozen liquor jugs", which is way way better than that quick brown fox nonsense. I found the hardest part was leaving enough room for spaces to be recognized.

Also if you want a cheap real pen stylus, try this trick... get ahold of some nylon "safety dart" tips. You can either crazy-clue them into a tip of a pen or wedge them into an empty refill tube. Sand them down a bit and you have a decent stylus for cheap!

Wow, thanks for the tips, Bun!

Now I just wish there was a smoother way to teach it my handwriting. It's really slow going through each letter of the alphabet manually, pressing Add and finally writing the letter... CellWriter does it well by just listing them all and letting me write a letter on top of each cell. I seem to recall reading that the handwriting recognition system will "learn" one's writing, but I don't see how that is possible given that I cannot explicitly tell it when it guesses my letters wrong. Have I missed something?

I think you are talking about Artificial Intelligence for learning your handwriting. There maybe, I dont know, in the early days of wince. I dont think we have AI in the HWR of the tablet. But there is something call "teach" under text input, basically you pick the way you want to write that particular letter. They have several patern, say, letter 't', after I pick the one I want, I have zero problem with letter 't'. I think that is what Nokia called 'learning', correct me if I am wrong....


bun