(During the summer of 1997, I brought an Apple eMate 300 with me to Japan. I wrote this evaluation to express my frustration with the product quality, or lack of it thereof. Not surprisingly, the product failed to achieve commercial success, and the entire development team lost their jobs.)


I never would have imagined that I would purchased an eMate 300. Now granted, I certainly wasn't one of the target customers for the product, but I couldn't imagine why anyone --- especially appearance-conscious kids and teenagers --- would even have the inclination to use the minimalistic, black-and-white Newton user interface, when given the choices of the more ornate, "cooler-looking" Windows 95 or Macintosh interfaces?

But, I needed a cheap, indestructable Internet-enabled laptop for spending a summer as an exchange student in Japan, and it seemed like the perfect choice, so I bought it, in all of its translucent greenness.

So, I set it up, and a week later, I'm still a bit surprised to find how much the eMate has grown on me --- even if I quite frequently have growing urges to test its alledged indestructability by slamming it into a wall, whenever I encounter the atrocious number of bugs and user interface inconsistencies littered throughout the system. But, let's start with the eMate's many good points...

First, the Newton OS on the eMate has, overall, been far more stable and maintenance-free for me than the Mac OS (or Windows 95) ever has been; I frequently spend 15% of my Macintosh time trying to resolve extension conflicts, deal with crashed applications, and other unpleasantries that should be unnecessary on a modern operating system.

I have only seen a single application crash under the Newton (Net Hopper 3.0), and have only seen only seen a full system crash twice, out of 75+ hours of usage. There's something to be said for not having to run an OS with 4 million lines of patched, bug-riddled code with over a decade of legacy issues; neither Macintosh or Windows affords this luxury, whereas it's a given on the eMate.

Likewise, I also have yet to see my eMate require more than a second to turn on or boot, presumably also as a result of not having a 150+ megabyte pig of an operating system to slaughter upon booting.

The most profound aspect of the eMate, of course, is its industrial design and luggability; the complete package simply evokes what could almost be described as an intimately personal computer.

Much like the obelisk in the movie 2001, the eMate's unexpected shape and identity emits its own mystique, and almost begs people to touch it, to feel it, and to interact with it. Never since the original Newton MessagePad have I seen a computer product evoke such a reaction from users or bystanders --- and unlike the original MessagePad, the eMate also offers functionality that's otherwise not available from any competitors' products within its price range.

Several other aspects of the user experience seem to compound this sense of intimacy when one actually uses the eMate. First is the pen integration in the interface; an eMate user not only sees and manipulates his or her documents and data with a keyboard, but during operations like text editing, will regularly touch the data with the pen, making eMate usage a tactile experience, rather than one performed through a mouse or keyboard at a distance.

Providing users with a feeling that they can't break their computer by pressing the "wrong button" has always been an essential characteristic of good interface design, and certainly one of the original tenets of the original Macintosh interface design; however, the eMate takes this principle a step further --- into the physical form, rather than just in the software. It doesn't matter how friendly a laptop's user interface might be if the user is terrified of dropping it or knocking it, and possibly losing their $7,000 investment; the user will probably never feel that the computer is truly an extension of their intellect, but merely an expensive tool.

I am sincerely amazed at how much more user friendly the eMate feels to me, relative to all the other laptops I've used over the years, just by virtue of the fact that it's as rugged as its user, and could probably quite comfortable be dropped, smashed, and knocked without problem. (Of course, I haven't actually done any of these, so it may be just as fragile as other laptops; but it's irrelevant. It's the user perception that dictates the user experience.) The eMate's (relatively) low price compounds this sensation of comfort with the realization that a computer damaged is not a life savings lost.

I'm sure there are other aspects of the eMate's design that (intentionally or unintentionally) end up providing a consistently comfortable user experience --- e.g. the "Casual" font, or the direct document editing metaphor (don't know the real name for this), in which the user doesn't ever have to worry about "saving" his or her work, since it's already pre-saved, but these are the core reasons.

Granted, the eMate, even in its released form, is clearly still a work in progress; I would not be the least bit surprised if it were rushed on the market by management. The bundled software is littered with glaringly obvious minor bugs as well as annoying user interface inconsistencies and inadequacies between the bundled applications; on average, I'd say I actually encounter a new bug (or at least a new manifestation of a bug) roughly during every hour of usage, which is --- by any measure --- wholly unacceptable for a supposedly finished product.

Some of the more grating examples that I've encountered:

I. The eMate's keyboard integration in the OS and bundled apps in both incomplete and inconsistent.

1. In the Notes application, the user cannot simply scroll to text below the boundaries of the text window with the arrow buttons (or with any on-screen control), but must actually press a special "scroll" key on the keyboard --- unlike the Works application, which allows cursor keys freely to be used. Same for the useful keyboard shortcuts that allow a user to reach the top/bottom/etc of their document. This is presumably a throwback to the original MessagePad 100 implementation of the notepad, which used hard-coded buttons on the touch-screen for scrolling, and it's very disappointing that this piece of ROM-burned software wasn't even updated to fully support the eMate's keyboard.

What's worse is that the InOut box's mail editor is even inconsistent with both of those interfaces, providing three inconsistent scroll models in the OS: in the InOut box, one must use on-screen scroll arrows, because the InBox uses the keyboard scroll buttons to scroll through your list of messages, rather than through the actual message itself. Accidentally pressing the keyboard scroll button in the InOut box results in a nearly 20 second wait to return back to the message you were reading --- and puts you at the beginning of it, requiring you to page through your message to get back to where you were --- easily another 20-30 seconds for larger messages.

In the end, this renders it impossible for the user to generate a motor-level eonnection with the scrolling-related component of the user interface, since the user must consciously think about what part of the modal Newton OS that they're in prior to pushing a button or tapping the screen.

2. Aspects of the user interface appear to still assume that the user would be primarily entering text through the handwriting recognition engine, rather than the keyboard, despite the fact that it would be physically awkward, if not downright impossible.

For example, double-tapping a word in Works still invokes a list of identically spelled words (for handwriting misrecognition correction), when the gesture could be made a lot more useful in a keyboard-centric environment by providing functionality equivalent to the Mac/Windows double-click of highlighting the word.

II. Some of the more memorable bugs in the bundled software (fewer than half of the ones I've encountered in under two weeks)

1. "Slips", or dialog boxes, frequently appear twice; first with only a single character or two, and are then redrawn again after a half-second with the full message text. Maybe this is meant as a feature by animating the slip's appearance in just two frames; who knows. But, it looks rather sloppy to me.

2. The Works application will frequently allow the user to place the editing 'caret' in a location where it is completely invisible. (a Macintosh wouldn't place the mouse pointer off the screen, neither should the eMate place its pointer.) The caret also appears at a vertically ambiguous location with its associated line of text is blank; namely, the caret appears to be one line higher than it actually is located.

3. Sometimes the cursor keys won't even work at all in the Works and Notes applications, even when the user is in the middle of a document and has an on-screen editing caret; deleting some text with the backspace key normally results in the cursor keys once again working properly, but not always. Sometimes the cursor keys will work, up to one line before the first line of a document, at which point one must scroll to the top line using the scroll bar to reach the top line. (This presumably happens in Notes since it deals with text as discontiguous blocks if the user points to a new location in the document by using the pen (which makes sense for a pen-based application, but certainly not for a keyboard-based application, since it makes it impossible to navigate by keyboard alone), but I have absolutely no idea why it happens in Works.)

4. Frequently, the help text for the Newton OS doesn't even fit into the provided help window, resulting in the help text being both truncated by several sentences, and overlapping with the window's buttons.

5. Sometimes, the on-line help text is just plain *wrong*. For example, one help page refers to tapping the "Classroom Connect" application; I can say fairly confidently from the context that they probably meant to tap the "Classroom Dock" application.

6. The results of a Find command frequently appear as blank options with checkmarks, and are only irregularly selectable.

7. Frequently I attempt to readdress or modify a message in my OutBox, I get a numerical error message of -48204. Waiting three seconds sometimes results in the error message number turning into a -8007; not a new error message of -8007, but rather, the number simply changes to a -48204. Huh?!?! Of course, neither of these errors are explained anywhere in the documentation; is it unreasonable to even include textual error messages rather than meaningless retro-mainframe numerical errors with eight megabytes of ROM, especially when the -48204 error appears at least once every 15-20 minutes during E-mail tasks? (mostly with large messages, so I assume it is some kind of low memory error)

8. The input buffer in the InOut box's text editor can be trivially overrun (in large documents (e.g. 20K), it won't accept more than 1 character every 5 seconds...!), with the computer sometimes inserting random garbage into the text document after the actual input has ceased.

9. Changes made to outgoing messages in the InOut box's text editor are frequently not actually saved upon closing the document and reopening it --- and yet somehow are still present in the document in a Print Preview. I haven't sent the messages in question to see whether the text was actually missing, or if it just wasn't being displayed.

10. Sometimes, tapping on a message in the eMate's InOut box will pull up a completely different message --- subject and all --- than the one that was actually selected. This is normally preceded by an unusually long pause in pulling up the message.

11. The alternate Time Zone selected in the Time Zone application (e.g. for keeping track of your home time zone while away) is randomly being reset to a completely different time zone than the one it was set to. (Or, randomly enough for me to comfortable classify it as random, any intentions of the designers aside... ;)

12. The bottom section of the Works scroll bar region can easily be corrupted and left improperly redrawn simply by extended the document to a second page, scrolling down, and deleting the extra lines that resulted in the second page being added; the scroll bar thumb will exceed the bottom border of the scrolling region, and will leave garbage underneath the scrolling region, until the Works application is closed and reopened.

III. Questionable user interface design

1. Selecting a document while Works is in overview mode and clicking on the File or the Envelope icons returns a misleading error message, that "Nothing is selected", also violating Apple's own guidelines for writing good error messages. Bullshit; there's certainly an item selected if its name is in inverse text and Works pulls it up when I press the return key! What it presumably meant to say is that nothing is "checked"; redefining what "selecting" something means between pieces of the user interface is inconsistent design and unnecessarily confusing.

2. Different portions of the user interface are drawn in a free-form anti-aliased grayscale style, whereas others remain drawn in 1-bit, without anti-aliasing, presenting an interface that appears to be in the middle of an unfinished graphical transition, much like the Macintosh shortly after System 7.

3. Likewise, despite references in the documentation (and user interface) to settings preferences for sound recording functionality, I cannot find out from the documentation whether the functionality is just present in Newton OS 2.1 or also in the physical hardware itself, perhaps through the Newton Inter Connect port. If the eMate can't support the feature, it shouldn't be in the user interface; this isn't Windows!

4. The "undo" button is misleading at best, and only undos text operations in most parts of the user interface. In fact, after I unintentionally changed a word with the Works spell-checker and pressed "Undo", it responded by erasing the last block of text that I wrote, rather than undoing the spell-check correction --- while the spell-checker was still
running!

5. Fundamental user interface gestures are inconsistent between different packages burned into ROM, making the eMate's learning curve more steeper than it should have to be. For example, holding down the pen on a word in Works selects the word, whereas in the Notes application, it simply invokes a highlight ball, and does not highlight the selected word.

6. In terms of the physical user interface, the row of keys that invoke specific functions (e.g. turning the backlight on and off, or undoing a text editing operation) are located directly behind the row of number keys, with no physical separation; e.g. half the time I press the delete key, I also switch on (or off) the backlight. Especially while typing in the dark. A simple raised bar could have easily fixed this problem. This might not be a problem for elementary school kids, but I doubt my fingers are any larger than those in the older portion of the K12 demographic group.

7. Pressing the "dock" button defaults to invoking a Classroom Dock package, which only connects to the Classroom Exchange application --- it won't also connect to the Newton Connection Utilities dock application, unless specified in the Teacher Setup package (which, of course, only teachers are likely to run, despite the fact that every non-teacher will want to change the dock setting to NCU, and will not easily find the magic switch).

This usability issue could have been easily solved by giving giving an error message with enough information to let you know that you're running the wrong connection utility on the Newton. Certainly it's not rocket science to look on the other end of a serial connection and see if NCU is running, and offer to auto-launch the Dock package if it turns out that the user is running the wrong app, or to at least inform the user why the connection isn't working. It's small points like these that used to differentiate Apple's user interface design from the competition, and it's disappointing to see that these points are being ignored.

8. Upon replying to a message in the InOut box, the eMate doesn't even place a cursor/caret into the message reply to allow text editing. The user must pick up the pen, place a caret (normally on the first line of the message, of course), and put the pen back down. Nominally, a 10 second and wholly unnecessary task for each message; I've never seen a Mac mail client that requires the user to actually click into their reply message window to place their cursor before they could start typing. ;)

IV. Documentation omissions

1. The 650+ pages of documentation, which has a higher-than-normal number of typographic errors, didn't even appear to contain a specification sheet, which is an inexcusable omission for any computer product. For example, despite mentioning facts like the limited lifespan of the backlight, the user cannot find out just how limited the lifespan is in order to weigh their usage of it --- 200 hours? 2000 hours? What kind of battery does the eMate use, and how many charges is it rated for? Is it a type that has a memory effect? (etc)

2. The on-line help is also littered with a distressing number of typos, spacing errors, and formatting errors; I saw over a dozen in a few days of casual usage alone. Never mind actual content errors.

I haven't touched the spreadsheet, drawing application, or graphing calculator; perhaps these components have fewer bugs.

Compounded, these quality issues are enough that I would only recommend an eMate as a consumer product for high school or college students with significant trepidation, as the software and documentation strike me as being early beta quality at best, and certainly feels rushed into production, perhaps for the Fall '97 school season; most of the bugs I've encountered are ones that even any competent QA engineer would have isolated very quickly.

The end result is that the current eMate is just not an especially usable product as shipped. The high-level user experience, in my opinion, is extremely solidly designed: the implementation just needs a lot more polish, with much more attention to quality and usability details.

Apple has demonstrated the practical feasibility of a full-featured, rugged, and inexpensive kid-specific computer, and once again created a new computer system category. It won't be that long until one of the Win CE manufacturers choose to slap their system into a cool-looking enclosure with equally rushed software and outmarket Apple with Microsoft's marketing machine behind them, too.

The only question left is whether Apple will capitalize on Yet Another market that they've singlehandedly invented before their competitors once again snatch it from their grasp.

(c) 1997 Eli Goldberg. Commercial distribution prohibited without permission. All contents of this message are the solely opinions of the author.