Is hand-writing recognition dead?

[Jason Dunn]

<a href="http://news.com.com/2100-1040-934637.html">http://news.com.com/2100-1040-934637.html</a><br /><br />That's what Handspring is betting on - they've found with their own products that people prefer using the small keyboard to writing on the screen. Ahem - Mr. Hawkins, what rational human being would want to learn Graffiti? I applaud those of you who are proficient with it, but I found it very daunting to learn as I'm sure many people do - it's not logical or rational. People are willing to accept learning curves with most software and hardware, but when it comes to input, I think people have less patience - they just want it to work. Keyboards, even tiny ones, are easy to understand - there's no learning curve, except perhaps to adjust to the key spacing, etc.<br /><br />I think this is why Transcriber is such an instant hit with people - we all know how to print, and Transcriber needs no training. People pick it up, and it works - Graffiti simply can't compete in this area. That said, I hope Microsoft/Paragraph seriously improves Transcriber in the future - it has a great technological base, but so much of the implementation sucks. The "dictionary" is a joke and the correction features are downright bizarre. But are keyboards the answer? Or smarter software? I'm betting on smarter software. What do you think?<br /><br />"Chief Operating Officer Ed Colligan, speaking at a CIBC World Markets investor conference here, told attendees that his company was surprised that consumers preferred entering information via keyboard, rather than using writing technology popularized by Palm's Graffiti technology..."

[ChrisD]

Hmm, it also could be the patent litigation with Xerox over Graffiti. Obviously a keyboard would not conflict with the patent.

[jpmierau]

I hated Graffiti until 8 months ago, when the flu laid me up with my new PPC2K2 device...lo and behold, 2 feverish days late, I no longer hated Graffiti, and now do 50 wpm :D :D

I hope new Pocket PC's don't follow this abandonment trend!

John Mierau
http://www.compu2go.com
~ultramobile computing & tech for people on the Move!

[tim254]

I would love to see a Pocket PC with a keyboard. I really want a Pocket PC that looks like a Treo: flip lid, keyboard, very small, vibrating silent alarm and very light (4 oz.).

Hopefully they'll make both a normal one and a keyboard version, but I really think a keyboard version is better. Why? Because with a keyboard you are using two hands/fingers to input your text which should be faster than handwriting or Fitaly (which I use now). Also, my hand gets tired of writing after a couple of sentences with Transcriber (It's just not the same as pen and paper).

However, I still would like to use Fitaly or possibly handwriting on occasions when I just want to jot down a small note.

[Inaki C]

I think character recognizing is essential to any handheld device.
Transcriber is a good example of what can be done.
External Keyboards existence is a proof of lack of maturity in the PDA industry.

Of course, both input system and applications must collaborate to offer te end user an easy way to work. However text input is fundamental so a resonable way to enter text is necessary. Keyboard is reasuring for the user because it mimics the traditional desktop keyboard but character recognition is the right way to go.

By the way, Paragraph has received a number of times recommendations on how to achieve better recognition rates but they seem to ignore external advice. It is a pity for final users because these methods do work. Anyone remembers Transcriptor?

[rubberdemon]

Although I've been firmly in PocketPC land ever since I sold my Visor and got an ipaq, I have to stand up and say a few kind words for grafitti. Although it puts a learning curve on users, it was a pretty elegant solution for the problem of text entry on a device (Palm) without the horsepower for full handwriting recognition (and in the wake of the Newton's widely mocked recognition software). I learned it pretty quickly, and still use it on my ipaq when I want to take a quick, short note. My friends with Palm devices all use it too, and it didn't take them long to get the hang of it (after I told them to play Giraffe for a while), and I've met people who can get real speed out of it.

Although keyboards are great (especially the folding ones), I think there will always be a place for using a stylus on a PDA, as if you're using a notepad. Maybe Transcriber will be in that place (it is impressive, when it's not mangling my words), but I'm sure a lot of people will continue to rely on Grafitti.

[Scott R]

I applaud those of you who are proficient with it, but I found it very daunting to learn as I'm sure many people do - it's not logical or rational. People are willing to accept learning curves with most software and hardware, but when it comes to input, I think people have less patience - they just want it to work. Keyboards, even tiny ones, are easy to understand - there's no learning curve, except perhaps to adjust to the key spacing, etc.
I disagree, the learning curve for Graffiti is quite small and, IMO, the long-term benefits far outweigh the short-term pain. Do you use Transcriber? I see a lot of PPC users tout Transcriber as the greatest input method ever, yet few of these people even use it. Why? Because it's inefficient, inaccurate, and a pain to make corrections. Even a thumbpad has a learning curve. First, a lot of people don't know how to touch type and so have to learn where the letters are. By this logic, the easiest most dumbed-down input method would be a thumbpad where the keys were in alphabetic order. The other problem with a QWERTY thumbpad is that the QWERTY layout was optimized for using 10 fingers to stretch to the keys with the least amount of stretching for the more commonly used keys (let's leave DVORAK out of the discussion for now). On a thumbpad, you're now using two fingers (thumbs actually) and the paradigm no longer works as well. FITALY has a learning curve but has a definite speed advantage once you do learn it (but has other faults which I won't get into).

There's probably no perfect input method. They each have strengths and weaknesses. One of the neat things about the PPC with it's virtual input area is that you can use the best input method for the specific task at hand. You could also use Transcriber for initial data input and FITALY for corrections, for example.

Finding the right "standard" input method for a device also needs to take into account the primary uses of the device. When the first Palm Pilot came out, it was designed primarily to be an efficient calendar, address book, to-do list, and note taker. For all but the note taking task, a stylus-based input method (and, IMO, Graffiti specifically) works best. Only in the case of inputting medium-length notes does the thumbpad excel (again, I'm talking about efficiency gains after any initial learning curves). In the case of the Treo communicator devices, the thumbpad's advantages strengthen because the device now adds email and Instant Messaging to the list of "commonly used" applications (though I have issue with that assessment as well, due to the fact that there's no unlimited usage plan offered - but that's an issue for another discussion). With email and IM, you're doing medium-length strings, which is why the thumbpad makes sense with these devices. But with the Treo 90, which is not wireless-enabled, the thumbpad, while a nice option for those who are afraid of learning Graffiti, is not the best input method for the device.

Scott

[Jason Dunn]

I disagree, the learning curve for Graffiti is quite small and, IMO, the long-term benefits far outweigh the short-term pain.


Scott, you forgot to add "in my opinion" to your statement. :wink: I think the learning curve for Graffiti is huge, you think it's small. Which one of us is right? Graffiti requires you to learn a new alphabet (or at least parts of it), and I consider that to be a significant learning curve, especially for people like myself who are highly uni-lingual (I have no talent for learning new languages).

[andydempsey]

In my opinion too, the learning curve for graffiti is small for the normal alpha numeric characters. The extended punctuation for characters like # and % take a little more learning. The only real issue is being willing to put the time into learning it! (I too am highly uni-lingualy, having given up on foriegn languages over 15 years ago).

However, the current crop of thumb boards are great too. My favourite solution is on the Sharp Zaraus - with the keyboard being hidden underneath a sliding panel. The major benefit with this solution is that you get the function keys / cursor control. Devices like the Treo loose the benefit of this.

The only problems I have had are with Pocket PCs' - But this is ENTIRELY down to not spending enough time one and being too used to graffiti!

[Sven Johannsen]

The other problem with a QWERTY thumbpad is that the QWERTY layout was optimized for using 10 fingers to stretch to the keys with the least amount of stretching for the more commonly used keys Scott

No it wasn't. It was designed to slow down the most proficient typists by not having the letters in a statistically appropriate layout. The issue was one of mechanics. The faster the typist, the more often the mechanism jambed. You are probably not old enough to have had to stop and clear jambs in a Royal upright :)

I actually find transcriber remarkably accurate when I am handwritting (not printing) and using real words (not acronyms). Unfortunately my handwritting is atrocious so I tend to print, and my environment supports and fosters acronyms.

I think learning graffitti would be somewhat like learning touch typing or shorthand. Some have more of an apptitude for that than others. Some can learn a dozen languages and some barely speak their own well. I've go to think speech recognition is going to be the Holy Grail, but we are a long way from that. At least I think so from this side. When we get to look back, I'll bet it will seem like just yesterday we were forced to scrawl on our PDAs :)

[KH]

Calligrapher is the MAIN application on my IPAQ, and the major impetus for moving to a new IPAQ with a faster processor. I stopped taking notes on paper two years ago when the 206 MHz processor made realtime notetaking practical. I personally find the cursive/text conversion far more usable than any keyboard, soft or hard, that I have tried. Calligrapher is essentially an enhanced Transcriber. One very useful feature is Pen Commander, which allows definition of short forms for commonly used words. A couple of my short cuts are 'tl' (for the tilde, which is apparently difficult to recognize) and 'mgt' for 'management'. You write the shortcut and circle it, and the result replaces the characters input.

As an aside, Tablet PC handwriting recognition demonstrations I have seen were awesome, although the devices will be too big to take the place of my PocketPC. On the other hand, a replacement for my laptop that allows handwritten input is very appealing.

Kathryn

[Scott R]

Most of what I say is my opinion. I usually add the letters "IMO" pretty liberally to my posts. That said, learning graffiti is less like learning a language and more like memorizing a pattern. Learning where the letter "B" is on a QWERTY keyboard requires that same learning process. Also, I've played with thumbpads before and there is definitely a learning curve even for experienced touch-typists as myself. This is because once you've learned to touch-type, you no longer have to think about where a letter is, you're fingers know through repetition where to go. When an experienced touch-typist picks up a thumbpad, they have to look down to figure out exactly where to push down with their thumb. Eventually, they can learn this new thumb-typing behavior. But that's a learning curve.

Back to Graffiti..many of the letters in the Graffiti alphabet are identical to the roman alphabet ("N", for instance). Many are equivalent to how you would write a letter if you didn't write the second (or third) strokes that would normally be required ("A", for instance, is written the way you write an "A" except that you don't draw the crossbar). Some letters are definitely a bit obtuse and require more learning. The end result, though, is that the learning curve for most users who give it the college-try is going to be relatively small (a day or two of regular usage to get the vast majority of letters down) and, equally important, once learned is retained in memory very well.

Scott

[Scott R]

No it wasn't. It was designed to slow down the most proficient typists by not having the letters in a statistically appropriate layout. The issue was one of mechanics. The faster the typist, the more often the mechanism jambed. You are probably not old enough to have had to stop and clear jambs in a Royal upright :)
Thanks for that bit of trivia. I actually am old enough to have learned to touch-type on a typewriter. It was the only class in high school that I failed. I had signed up because I wanted to learn to touch type (since I was a computer geek), but the class also required learning to write letters and other documents in certain formats. I didn't pay much attention (and do my homework) in those areas, so I failed the class despite learning to touch-type fairly fast and accurately. That said, I think my point still stands that while the QWERTY layout combined with learning to touch-type makes for an efficient (if not the ideal - given your post) input method for a desktop/laptop computer, knowing how to touch-type on a QWERTY keyboard doesn't translate into knowing how to quickly type on a QWERTY thumbpad.

I've go to think speech recognition is going to be the Holy Grail, but we are a long way from that. At least I think so from this side. When we get to look back, I'll bet it will seem like just yesterday we were forced to scrawl on our PDAs :)
That's another one of those notions that I don't buy into. I think speech recognition will eventually mature and have a place, but I don't think it will ever be the right method to use all of the time. There are the added issues of security (I don't want everyone around me hearing me write my personal notes), noise factor (imagine 10 people in a meeting all taking notes at the same time by talking to their PDAs while the speaker is talking), and even efficiency in certain cases (sometimes it's just easier to tap an icon with the stylus rather than have to speak out lengthy commands). On the other hand, speech recognition is great when I'm driving my car and want to jot down a note, start an MP3, or initiate a phone call.

Scott

[Jason Dunn]

The only problems I have had are with Pocket PCs' - But this is ENTIRELY down to not spending enough time one and being too used to graffiti!

You don't find the Block Recognizer, which is a licensed version of Graffiti from Xerox, to work well for you? I had heard that it was quite good...

[dave]

graffiti is better than that spare keyboard on the treo. why can't all these cats that are making thumbboards just copy the RIM? i can type almost as fast with my RIM 952 as i can on my pc -- a little exaggeration here, but trust me i'm pretty quick with it.

i have toyed around with the palm i705 thumbboard, the treo, the Ericsson chatboard, and that new Sony clie that the palm die-hards are clinging to as a savior. I listed these in this order for a reason. their level of suckiness increases in that order. that sony keyboard is ridiculous, but after many years of using other sony devices, i can't say i'm surprised.

[Chris Spera]

Or at least learning is is like learning shorthand, IMHO.

Shorthand isn't even close to intuative. My mother used to be very proficient in shorthand, and I remember looking at some of her notes as a child and wondering what the heck she was thinking while taking shorthand...

While Graffiti isn't shortand, it does require a learning curve of some kind. Input methods like Transcriber or Calligrapher don't really require a learning curve, and that's where they add instand value, and the purpose of this whole counter-point.

Graffiti requires you to change the way you write/ enter information. Handwriting recognizers like Transcriber and Calligrapher don't.

Christopher Spera
-------
pocketnow Writer

pocketnow.com -- it's all about portability...
http://www.pocketnow.com

[jdhill]

OK, since we are listing favorite thumb keyboards, here is my list from best to worst (includes only devices I have actually touched, IMHO, YMMV, all of the standard disclaimers):

1. RIM Blackberry
2. Motorola T-900
3. Motorola Timeport P-935
4. Handspring Treo
5. Sharp Zaraus
6. Sony Clie NR70

The Sony Clie NR70 is a groundbreaking device, but that keyboard needs a LOT of work (IMHO, etc.) !!!

[madmaxmedia]

Personally, I thought that learning alphanumeric characters in Graffiti were a breeze (and I HATE learning languages too.) But after a few years of use, I still can't really remember the punctuation marks and special characters (with the sole exception of perhaps '@'.)

Personally, I would rather have a good transcriber and a Stowaway keyboard than a built-in thumb keyboard, at least until MS or Palm can really integrate the keyboard without having to constantly use the stylus as well (that's what I've heard about the Treo.) The RIM devices were designed for keyboard input only, and I think there is a significant difference in useablilty.

I don't think the variety in input methods is indicative of the immaturity of the market as it is indicative of the fact that our hands were not designed for data entry. Anything that fits our stubby fingers is going to be too big to pocket.

[Scott R]

While Graffiti isn't shortand, it does require a learning curve of some kind. Input methods like Transcriber or Calligrapher don't really require a learning curve, and that's where they add instand value, and the purpose of this whole counter-point.

Graffiti requires you to change the way you write/ enter information. Handwriting recognizers like Transcriber and Calligrapher don't.
Again, IMO, this is a flawed argument. You have to look at the entire picture. There's a learning curve to using a Palm device. There's a learning curve to using a PPC. Entering data is one piece of the learning curve. When developing the primary input method, you need to determine what applications the device is going to focus on and use an input method which is best geared toward those applications. You also need to balance immediate ease-of-use with long term efficiency (both make up usability - a common misconception is that usability=ease-of-use). Transcriber is better than Graffiti as far as immediate ease-of-use is concerned. But in terms of long-term efficiency (after the learning curve of each), when used by itself it is less efficient (writing the entire letter "A" takes longer than writing an upside-down "V") and more of a pain to make corrections with (this is because it interprets letters within the context of an entire word). For example, draw the number "1" by itself and it may interpret it as the letter "I".

I'm not saying that Graffiti is better than all other input methods in all types of scenarios. For me, it offers the best balance of a small initial learning curve, long term decent speed, good accuracy, and ease of making corrections within the contexts of how I use my PDA (60%+ PIM usage). If my usage patterns were different (such as doing a lot of email writing), I might prefer a thumbpad. If I did a lot of extensive writing on it, I would want to use a stowaway keyboard in addition to Graffiti.

What I find interesting, though, is that the the supposed "power-user" crowd, bored by Palm's simplicity, is against the notion of short-term learning curve for long-term efficieny gains.

Scott

[Scott R]

graffiti is better than that spare keyboard on the treo. why can't all these cats that are making thumbboards just copy the RIM?See my previous comments on this. The RIM was designed from the ground up with the thumbpad in mind. There is no touch screen. All data input is performed via the thumbpad, other dedicated hardbuttons, and the scroll wheel. Furthermore, the thumbpad has a more comfortable design than the existing PDA thumbpads because the device is wider (though shorter - thus keeping it smaller overall - I'm talking about the original Blackberry here, not the larger one which I'm guessing didn't sell nearly as well).

With the Treo, Handspring added a decent quality thumbpad and a scroll wheel. They even added some dedicated buttons to get to certain apps. But it doesn't eliminate the need for a stylus to tap on on-screen buttons, etc. One possible way to improve upon the PDA/thumbpad concept for a touchscreen PDA is to add some sort of directional pad and action button (I believe the Zaurus does something like this), so that you can do everything you need to without taking your thumbs off of the thumbpad.

Interestingly, this reminds me a lot of the concept of the pointing stick on a laptop. I love my Thinkpad because of this and have always hated touchpads. With a pointing stick, I can type and move the pointer without moving my keys off of the keyboard. Yet, it seems that most people hate pointing sticks and prefer thumbpads. Efficiency doesn't always win out.

Scott

[4AMFriday]

I disagree, the learning curve for Graffiti is quite small and, IMO, the long-term benefits far outweigh the short-term pain.


Scott, you forgot to add "in my opinion" to your statement. :wink: I think the learning curve for Graffiti is huge, you think it's small. Which one of us is right? Graffiti requires you to learn a new alphabet (or at least parts of it), and I consider that to be a significant learning curve, especially for people like myself who are highly uni-lingual (I have no talent for learning new languages).

I'm going to have to side with Scott on this one. I have been using a PPC for well over a year now. I have tried time and time again to get a grip on Transcriber only to end up frustrated at the sheer amount of errors it makes in recognizing my handwriting - which really isn't that bad at all. I almost always resorted to using the on-screen keyboard to input even the smallest strings into a PIM program.

I recently picked up my first PALM device about a month ago - the Sony Clie' NR70V (who could resist). To my surprise I not only ended up using the Clie' as my "daily driver" for the most part, but in little less then a week, I picked up the Graffiti input method and have become rather efficient with it, more so then I ever was with Transcriber.

I now find myself being capable of entering in more characters a minute then I could even with the on-screen keyboard. I do agree there is a "small" learning curve (most of the standard characters really aren't written any differently then you would normaly write them in Transcriber), but feel that grafitti, at the very least in the form it exists on a Palm device, seems to me to be both a more efficient and reliable input method.

Funny enough, I even found myself writing quite a few pages of workout routines in QuickWord last night while using Graffiti as the text input method the whole time. I could not even begin to imagine how much longer it would have taken me if I had to go back and correct all the recognition mistakes Transcriber would have made.

It should be noted that I own both a HP/Compaq iPaq 3850 PPC and Sony Clie' NR70V Palm Device. I don't consider myself more bias towards either one of these devices. I just get more done with Graffiti and the Clie. Oddly enough, the Graffiti on my iPaq never seemed to be nearly as efficient and reliable. What gives?

[Scott R]

Oddly enough, the Graffiti on my iPaq never seemed to be nearly as efficient and reliable. What gives?
For whatever reason, even among Palm OS devices, Graffiti sometimes works more accurately on one device over another. I never got around to upgrading my iPaq to 2002, but I did play with 2002 on a HP 565 at a store for a few minutes and found the improvements to Block Recognizer to be much improved over the 2000 implementation. My Samsung i300 doesn't translate Graffiti as accurately as I'd like. :( I'd attribute much of its problem to the smaller physical input area, since the screen is a lot smaller.

Scott

[fmcpherson]

I have a slightly different slant on this argument. For me my problem isn't so much that Graffiti is hard to learn. Rather, I have a problem with the concept that I have to change something that I already know how to do in order to work with the device. Instead of the device conforming to me, I have to conform to the device. (Yes, I have to write neater, but writing neater isn't learning how to write.)

It's like if I bought a computer and they only came with Dvorak keyboards and I already learned how to type on typewriters with QWERTY keyboards.

I am also a bit of a purist in that I think styluses are the best form of entry for PDAs, and I don't find any thumb style keyboard to be easy to use or of value. If I want to do a lot of typing I will get a handheld PC or a keyboard add-on.

[Ed Hansberry]

graffiti is better than that spare keyboard on the treo. why can't all these cats that are making thumbboards just copy the RIM?
RIM has a patent on that design, that's why. :)

[Scott R]

I have a slightly different slant on this argument. For me my problem isn't so much that Graffiti is hard to learn. Rather, I have a problem with the concept that I have to change something that I already know how to do in order to work with the device. Instead of the device conforming to me, I have to conform to the device. (Yes, I have to write neater, but writing neater isn't learning how to write.)

It's like if I bought a computer and they only came with Dvorak keyboards and I already learned how to type on typewriters with QWERTY keyboards.
IMO, this isn't much different than if you bought a computer with a QWERTY keyboard and didn't know how to touch-type at all.

I know I've beaten this horse past death, but the PDA was a new type of device. With the original Palm Pilot, the Palm designers went with a new type of input method that worked well for the limitations of the device and to maximize speed of data input. The DVORAK keyboard and a mouse are great input methods for a desktop computer. Voice recognition combined with a hard numeric keypad work well on a cell phone. While the input method of the PPC may not require much of a learning curve, assuming you have certain beliefs that it will be like your Windows desktop experience, you still need to learn that the Start button is now at the top, there's no close button, etc. Change for change's sake isn't good. But if there's an acceptable learning curve which pays dividends medium-term, I think it's worth it.

IMO, had Hawkins used the Unistrokes character set for Graffiti, it would have been a huge mistake. The learning curve would be much higher due to the fact that most of the characters look nothing like their Roman counterparts. One could argue that the Unistrokes characterset is more unique and, therefore, would provide greater accuracy (and perhaps speed). But as I said earlier, you have to balance long-term efficiency with short-term learning curve. Another example is that other input method (Quickstrokes?) which involved keeping the stylus centered and moving it to various quadrants and then back to the center position. The idea was that, once learned, it could offer very fast input speed. The problem was that the learning curve was too high for me and I became easily frustrated with it. I'm willing to concede that even Graffiti's learning curve may be too high for some.

Scott

[Scott R]

graffiti is better than that spare keyboard on the treo. why can't all these cats that are making thumbboards just copy the RIM?
RIM has a patent on that design, that's why. :)
I read that before. I believe their patent is on the V-shaped layout of the keys or something. This is important as it makes using the keyboard faster and more comfortable. I find it a bit difficult to believe that such a patent could be enforceable, though.

Scott

[dave]

patent, shmatent.

ok, if they can't copy it they can at least 'imitate' it.

[greenmozart]

I started with a Visor and found grafitti to be much more accurate and efficient than transcriber. Perhaps the best illustration of this is the fact that I reached a point where I didn't have to look at what I was writing - perhaps we can call it "touch-grafitti" - whereas with transcriber I have to hold my PPC upright and position my stylus in one of three areas (capital, lowercase or numeric) vs just 2 on the Visor. The best part about not having to look at the device is that holding up a PDA and having to flex my wrist backwards to write with my stylus is VERY uncomfortable compared to being able to hold it at a more natural downward slant, especially considering how much the stylus slips on a PDA screen.

I don't know what the ultimate input method is at this point. Everything that is currently available has its pros and cons - and I'd have to agree it depends on the application. I do like the idea of the Zaurus approach, however, which at least gives me OPTIONS. Perhaps that is the holy grail.

[Ravenswing]

I was looking forward to PPC 2002 for many reasons (most of which were incorrect it turns out, but that's another story), and Block Recognizer was one of those reasons. I liked Letter Recognizer, but I believed that Graffitti had more single stroke letters than Letter Recognizer.

It does, but the punctuation is a nightmare, so you win some, you lose some.

Nowadays, I use Block Recognizer, Keyboard, and my Stowaway, depending on the application. None of them are as accurate as typing on a real keyboard, and they never will be.

As for Transcriber and Calligrapher... No learning curve? Get stuffed! The only way I could get either of them to be even vaguely accurate was to adapt my handwriting to the way they expect me to write. Then you have to learn the dumb gestures to activate the correction mode (and you WILL have to use correction mode). Maybe it works well for some people. Then again, I'm sure speech recognition works well for some people.

My other issue with Transcriber (more so with Calligrapher) is the size. They are huge programs, both in storage and runtime memory. For that vast consumption of resource I expect returns and neither program has ever given me those returns.

[SassKwatch]

Transcriber 'just works'???

Not for me......even with significant 'training', it is still the least reliable input method I've tried on a pda. Don't *think* it's my handwriting as most people find mine very legible....and even attractive.

And though I wouldn't claim to be an overly big fan of graffiti, at least it gets the job done for me.

Thumboards....never had 'em, never will.

[Chris Spera]

[quote="Scott R Transcriber is better than Graffiti as far as immediate ease-of-use is concerned. But in terms of long-term efficiency (after the learning curve of each), when used by itself it is less efficient (writing the entire letter "A" takes longer than writing an upside-down "V") and more of a pain to make corrections with (this is because it interprets letters within the context of an entire word). For example, draw the number "1" by itself and it may interpret it as the letter "I".Scott[/quote]

Scott,

I agree that Transcriber isn't the best at what it does. Calligrapher is much better, but unlike Graffiti, requires only a small change in writing style. Instead of typing a "l" to get a number one, you may actually have to draw a "1".

My point and opinion, however, revolve around the fact that with Transcriber and Calligrapher, you don't have to learn any unnatural motions. Writing an upside down "V" for the letter "A" isn't more intuitive at all, IMHO. Its down right confusing.

Regardless of Palm or PPC, I shouldn't have to strain my brain to get data into the device. The best input methods are supposed to be a fast way of entering data. Graffiti isn't all that great because it requires you to learn a new alphabet.

Anyway, this is my opinion; and its why I don't use Graffiti on my iPAQ.

Christopher Spera
-------
pocketnow Writer

pocketnow.com -- it's all about portability...
http://www.pocketnow.com

[Ekkie Tepsupornchai]

Horses for courses, I always say. I found the Graffiti method on Palm to be relatively quick to learn, though you did have to spend time learning it. I remember a co-worker of mine ordered her Palm V the same time I did and she never got the hang of it, mostly because she never exercised the patience needed to learn it. She always ended up popping up that digital keyboard to enter data.

As Scott has mentioned, there's the whole balance between long-term benefit and short-term usability (I'm paraphrasing here), but as someone who switched to PPC a year and a half ago, I've never found Transcriber to be fit for my needs.

While you could immediately type in letters without re-learning them, I had already learned and understood the benefits of Graffiti in the past. I found Transcriber to be way to slow and error-prone for me.

That having been said, I do use my PPC for extended note-taking sessions or document creation tasks. My primary methods of input now are FITALY and a Stowaway. The Stowaway is obviously great for touch-typing and the FITALY has a tremendous learning curve relative to other input devices, but is pretty darn fast.

In summary, I guess for some people, speed of input is not all that important. I've seen a number of people that detest "learning a new handwriting method". For others, the long term benefit of speed is worth the learning curve associated.

Ekkie