I used an Apple Newton MessagePad 2100 pretty much every day for about seven years (see my).
What was so great about it? What kept me using it?
Well, I used the Newton to keep track of the good things about the Newton--and the bad things too. So to write this article, I just powered up the unit (which still works fine, along with a Farallon Ethernet PC Card, a 24MB linear flash card, and two of the three nickel metal hydride battery packs, plus another pack for AA alkaline batteries) and reviewed the "Newton Notes" entry.
Updated 2007-08-02, 1500 PDT-- That's it in the photo. It isn't as pretty as an official Apple product shot, but that's how it looks today. If you click here, you can see a higher-resolution version of the photo. The stylus in the picture is a custom model I turned from a solid bar of titanium. It sits nicely in the stylus rest (shown extended) and locks properly into the internal storage slot.
The note shown is the last one I entered, when I took the machine to the Worldwide Newton Conference on January 14, 2006. Note that there are 2,777 notes in the Newton. There are a few hundred more that were created in the months before the Ricochet wireless modem problem I described in that earlier blog entry, and these were recovered from the device's memory to a disk file. So all told, I averaged well over one note per day for about seven years.
Here are some of the high points from my Newton experience:
Never having to save files. Files are always saved. In fact, there were a few times over the years when the Newton rebooted or the battery died while I was in the middle of writing a note. After restarting, the note was usually only a word or two behind what I'd been writing.
Ultrahigh reliability. Other than the Ricochet problem, the Newton rarely crashed (just a few times per year, not counting those caused by software testing, which I did a lot of), and almost never lost data. In fact, it was so reliable that I became quite blase about backups, and I never really worried about installing third-party software.
Extraordinary battery life. My Newton would run for upward of 20 hours with the backlight off, even with the Ethernet and flash PC cards installed. I think the Newton was designed to keep the Ethernet card pretty much entirely turned off when it wasn't in use; I couldn't see a difference in battery life with it in or out. The flash card was used constantly, since it held a lot of software and data. With the backlight on, battery life was still in excess of eight hours.
This kind of battery life makes a huge difference in usability. I could sit in conferences and leave the Newton on for hours, continuously ready to take more notes. Even though the Newton turns on and off in about a second, I could set the idle timer to wait half an hour before turning the unit off automatically; battery life was just no big deal. I could also take weekend trips without taking along a power adapter. For trips up to about a week long, I'd bring along a spare battery pack and still not bother with the power adapter.
I notice that the electroluminescent backlight on my Newton is really dim now. It probably hasn't gotten any worse since 2004, but I probably just got used to it as it dimmed over time.
The case had a nice high-friction coating. The slightly rubbery surface made it really easy to hang onto the Newton. Some people had a lot of trouble with this coating rubbing off, but it held up fine on mine. I'm sure that it helped a lot that I kept the Newton in a zippered slipcase made from a large travel wallet.
The user interface for the notes application was basically like a continuous scroll of paper. This seemed a lot more convenient than treating each note like a separate document. On the other hand, it put a lot of pressure on the scrolling controls. By the time I had hundreds of notes in the machine, I needed something more than the standard scroll arrows. Fortunately, Newton developers noticed this need, and there were good software solutions available.
The recognizer was an excellent match for my fairly neat printing. When I (or the recognizer) made mistakes, they were usually easy to correct by double-tapping the word to get a list of likely substitutes. I did notice that the list of suggested fixes seemed to be based exclusively on likely recognition errors; it would have been useful to get spelling corrections too.
Direct overwrite also worked well. If I could see a single letter that was misrecognized, it was usually faster to just write the correct letter over it. Today's Tablet PCs don't respond well to this approach, and I think that's a major omission.
The default mode of operation was that text was always recognized and converted to a font-based display, as opposed to the default in Tablet PC today, which is to recognize the text but leave the ink on the display. This creates the potential for the user to be distracted by the recognition process (and errors) but leads to higher-quality recognized text. This trade-off favors experienced users at the expense of initial impressions, but I think that's the right trade-off to make.
And there's a subtle point here: the default Newton font looks like very neat handwriting--a lot like the lettering in comic books. This means that characters with similar handwritten shapes have similar font shapes too. So if the Newton misrecognized a letter here or there, the resulting word was easier to read correctly by virtue of the same skill that helps people read bad handwriting.
In reviewing this "Newton Notes" list, I realize that there were a lot of things that bothered me at one time or another about how the Newton worked, and I spent a lot of time thinking up ways to improve the software and hardware. I'll blog about these notes later.
[to read part 2 of this piece, "What wasn't so great about the Newton?"]