While I'm struggling with installing a decent Linux distribution on my Jornada 820, I'm using my uncle's iPAQ 5555 as a toy PDA (he lent it to me so I learn how to use it and can teach him). Here are a few thoughts on the usability of this beast...
As for form factor, the iPAQ makes it certainly easier to hold the machine with one hand so as to browse a file. On the other hand, the limited width of the screen is not as comfortable. What is worse, all the stylus methods are quite slow: the letter recognition is surprisingly good, but even assuming you are trained to write in such a way that the recognizer will never make mistakes, that's still a far cry from a keyboard. Plus all the wait-dependence of the stylus interface adds a lot of latency to your inputs. And unlike what happens with a keyboard, that latency is not something that you can mechanically wave away, because you must make time-dependent interaction with the computer, and not just send one-way mechanical orders to your hands as when you type on the keyboard. Thus, typing text with the iPAQ really disrupts your flow of thought, or at least mine. All in all, the iPAQ form factor is very bad at inputing any kind of text. I suppose it gets better with training, but I can't imagine it ever becoming remotely as good as a keyboard. And I don't imagine me or any developer getting as familiar with stylus input as with a keyboard - the moment we can have the real thing, we drop the gadget. Think of it: if stylus input was better than a keyboard, then desktop computers would be selling with a stylus instead of a keyboard.
We saw that text was not something you'd want to overdo. The same applies to drawings. If it were but for note-taking, paper-and-ink-pen technology works better. However, the iPAQ note taking engine has a nice feature of voice input (but no voice recognition). You can intermingle text, sound and drawings in a same set of notes. (The Jornada 820 WinCE note-taking software cannot do that; also, it has sound I/O but no audio jack; grumble.) There is a little problem with the presentation of sound bytes in the interface, namely that they all appear as one anonymous sound icon, so you have write text next to the icon if you want to sort sound bytes, which can disrupt the recording and is best done after the recording session. Default recording quality isn't of very good for space-efficiency reasons; makes you wonder whether real-time mp3/ogg compression would be possible on such a beast.
Interestingly, taking notes by recording your voice implies quite a different structure from taking notes by typing text. Whereas the slow typing of text with the stylus broke the stream of consciousness, speaking to the machine leaves you with an unedited stream of voiced consciousness (which is not the same as an unedited stream of consciousness -- when will we have mind-reading machines?). This is quite unusual to me; when inputting text with a keyboard, I focus more about the ideas, the concepts, and will readily go back and forth in the text to modify the way I explained my ideas, to write a few key ideas in advance of the rest of the text so as not to forget them, etc. There is no such possible thing with voice notes: it keeps recording while you are searching for the right formulation; it catches the first attempt; you can easily add new notes, but you can't easily fix part of a previous note; if you split your voice notes in independent sound bytes (which requires more administrative hassle), you may change but part of the notes, but this majorly disrupts the thought process, and there is no reason that the next recording will be "perfect" either. You find that you have to leave all perfection-seeking to a further session of note-transcription, which will take a lot of time that you might have initially expected note-taking to save. No, it won't save that time. On the other hand, it might be an opportunity to express ideas that you wouldn't have expressed without the ability of taking voice notes at the moment the idea comes.
Voice recording imposes upon you quite a different way of thinking from keyboard typing when taking notes: you'll try harder to get an intelligible form the first time, yet won't try as hard to tinker it to make it perfect; you don't have an actual visual cue upon which to incrementally structure your note-making; instead you must make a more complex model of your discourse in your head, so as to be able to discourse about it coherently. You both have to let yourself go and try to express your flow of consciousness as seamlessly as possible, yet grow ways to cristallize and vocalize it in a quick yet intelligible way. It's quite a different discipline. Inferior in many respects; superior in many other respects; an enriching experience anyway. And this experience can help you organize your thoughts in new and different ways, and hopefully be more creative. In any case, this experience makes me realize by contrast much of a man of written culture I am.