norman@sdics.ucsd.EDU (Donald A. Norman) (05/15/87)
Some time ago I sent out a request for information on Macintosh Word Processors. I complained about Word 3.0 and asked for replacements. What I wanted was a professional level system. Here is a summary of responses, plus the results of my own survey. Quick Summary: 1. no easy answers. 2. T/Maker's Write Now. Now available. With minor exception, very well liked by those who have it. But not quite powerul enough for some of us. 3. Ann Arbor's FullWrite. Looks good in demo, but demos are notoriously unreliable, the release isn't ready yet. I saw no signs of user testing. etc. (see what follows) 4. I will stick to Word 3.0, then try FullWrite (but I won't hold my breath). Long Discussion T/Maker's Write Now. Many people wrote to sing their praises. A few voiced reservations. I was able to get a demo copy and to try out a real version (same things, essentially). I interviewed a department secretary who uses it extensively for memos and technical papers. She really liked it, even for technical papers, but the Department Chair (the software reviewer for the SD Mac Users group) didn't think it powerful eneough. Conclusion: a very nice implementation. Think of it as the way MacWrite should have been. If you only want one text editor, get it. The reservations, which seem valid to me, is that it doesn't really have the full power that some of us need. WriteNow may sufice for 90% of the people, and maybe for 90% of the usage of the other 10%, but from what I could tell with my evaluations, it wouldn't do all that I wanted. You may prefer it, however. (Does not suffer from the fatal design desease of creeping featurism. See below.) Ann Arbor Softworks FullWrite Professional. I saw a demonstration at the San Diego MacUsers Group, and I talked with the demonstrater (Tim Johnson) afterwards. It was impressive, with all that I asked for, but. 1. It isn't finished. It may be released in August. (It will cost $300.) It is being rewritten (?) in assembly "to speed it up." Not-yet-released software is always more impressive than actually available software. 2. It shows all the signs of the fatal desease "creeping featurism." Think of a feature you want, FullWrite has it. Reference files. Strike-through text. Automatically marks places that have been revised. The search feature has layered menus, letting you specify lots of different variables (yes, you can search for a font style). It even has a version of MacDraw built in. You can flow text around pictures. Columns. Post-it notes. You name it, it is there. But too many features lead to too much complexity. There is no alterntive. And it wasn't clear that the features were thought through: they were just there. The crowd at the demo went into a feeding frenzy: "Show us X!" they would yell, and there would be X on the screen. I suspect the programming team went into a similar frenzy. A really bad sign, if you ask me. In my private session with Johnson I asked about user testing. Oh yes, he assured me, nothing would be released unless tested. No, I explained, was there user test in the design stage? He looked at me blankly. An essay on ergonomic design -- user-centered design. (Cognitive engineering). You have to design with the end user in mind. The end user who represents the target population. You need to involve them at the very first stages. Not at the end when major design paths have been chosen and it is too late to change. There is a whole profession devoted to this concept, even a whole set of conferences. Alas, major software houses, definitely including MicroSoft, and, I guess, Ann Arbor Softworks, ignore this. Well, ignore it at your own peril. The result is that you will fail in your products. Programmer are not trained to understand users. Arghh. I could go on and on. For a whole book. In fact, that is my book, that's why I want a better text editor. foiled by my own needs. Will FullWrite be satisfactory? Who knows? It looked good. The design philosophy looked good. But skillful demonstraters can make even the most ungainly systems look elegant. The excess features are a bad sign. The lack of user involvement in the design stages is a poor sign. When the final product is released, then a real assesment, with real users, can begin. Word 3.0. My experiments with other editors makes Word look a bit better. "You will get used to it," several people told me. What a horrid phrase for a designer: don't worry, you will get used to it. Good design should not require "geting used to." What Word needs is some better thought through principles, principles which are then carefully followed throughiut the design. And then carefully presented to the user (what I call the "System Image") so the user gets a cohesive, consistent conceptual model of the system. Now, there are many different philosophies intermixed, which makes it confusing for the user. Many design philosophies are hidden from the user, so that the user can't really tell what is happening, or why. Styles is one example of something where the design philosophy is not at all clear. Word should redo itself more in the spirit of the Macintosh. It needs better menus. Layered menus, of which several possibilities exist. See the Macintosh user guidelines (forthcoming). Contrast glossaries with styles. Why are there two separate schemes for doing very similar things? And with different design sepcifications. The glossary seems straightforward: one enters terms into a file and calls them back with comand-backspace and a name. Styles are not so clear. The "System Image" is never well specified and badly thought through. Where do styles reside? Are they attached to the document or in a general file for general use. Do new styles get attached to old documents? Does modifying a style affect those old documents? (Don't tell me the answer -- that isn't the point. The point is that the philosophy should be apparent in the use: one shouldn't have to read the manual or ask a guru. The usage should reflect the underlying design principle. Causal users can't always refer to the manual. The story goes on and on. The complexities only occur for complex things, which is good, I suppose. Some words of praise: Word's spelling corrector, especially its ability to make appropriate suggestions, is the very best I have seen. Whoever did it should be rewarded. The help system is quite good. Lots of the features are well done. Which makes the deficiencies even more frustrating. Word has been very good about updates. The announced bug corrector for Word 3.0 is a good sign. Note that I am not complaining about bugs. Bugs don't bother me, for bugs can be fixed. I am complaining about design decisions: they are not so easily fixed. Conclusion: It is a tough decision. Nothing is quite right. But I am sticking to Word. As one person correctly pointed out, it would be madness to switch word processors when one was rushed. Comment: Am I being inconsistent? Complaining that Write Now doesn't have enough features and that FullWrite and Word have too many? Partially. But carefull, consistent design, design based on a good conceptual model with a consistent system image can overcome the problem. Word fails badly at this. FullWrite looked quite good in demo -- better than Word -- but one cannot tell without a complete evaluation: by the appropriate set of users. Donald A. Norman Institute for Cognitive Science C-015 University of California, San Diego La Jolla, California 92093 norman@nprdc.arpa {decvax,ucbvax,ihnp4}!sdcsvax!ics!norman norman@sdics.ucsd.edu norman%sdics.ucsd.edu@RELAY.CS.NET
jlc@goanna.oz (J.L Cybulski) (05/18/87)
In article <360@sdics.ucsd.EDU>, norman@sdics.ucsd.EDU (Donald A. Norman) writes: > Some words of praise: Word's spelling corrector, especially its > ability to make appropriate suggestions, is the very best I > have seen. Whoever did it should be rewarded. I do not quite agree. I used Mac Lightning with a dictionary and a thesaurus, its interface was much more advanced than that of Word 3.0. The most important feature was its interactive mode (ie. spell checking in the background - as you type). It WAS because Microsoft changed the way Word handles events (sort of delays them) and now MacLightning similarly to many other DAs does not work with Word 3.0. I agree with Donald that Word 3.0 is a bundle of badly designed and unrelated features. I actually switched back to Word 1.05 (after spending $$$ on v 3.0). Jacob
chuq%plaid@Sun.COM (Chuq Von Rospach) (05/18/87)
>> Some words of praise: Word's spelling corrector, especially its >> ability to make appropriate suggestions, is the very best I >> have seen. Whoever did it should be rewarded. > >I do not quite agree. I used Mac Lightning with a dictionary and a thesaurus, >its interface was much more advanced than that of Word 3.0. The most important >feature was its interactive mode (ie. spell checking in the background - as >you type). Interactive spell checkers drive me up the wall. I guess it is a matter of taste. I, personally, used SpellSwell until I upgraded to Word 3.0. The Word 3.0 Spellchecker is not as functional as SpellSwell, but it is a lot more convenient to use the checker inside the program. There are a few glitches in the spell checker, but generally it does what it is supposed to do quite well. >It WAS because Microsoft changed the way Word handles events (sort of delays >them) and now MacLightning similarly to many other DAs does not work with >Word 3.0. I could say something about programs like MacLightning that depend on mucking with undocumented interfaces and people who depend on compatibility where compatibility is not guaranteed, but I'd probably just start a religious war. So I'm glad I'm not saying it. >I agree with Donald that Word 3.0 is a bundle of badly designed and unrelated >features. I actually switched back to Word 1.05 (after spending $$$ on v 3.0). I'll agree that some of the modules don't fit together well -- I'm not sure why you would want to put an outline processor or a spellchecker inside a word processor (except, of course, for the fact that you're trying to simulate a multi-processing environment). They may not be as functional as the state of the art widgets you can buy separately, but they weren't designed to be, either. They are simply functional additions to a program, not designed to be whiz bang programs. If they aren't functional enough, then go out and buy the whiz banger (as I did with Acta, since I need an outliner that's available in any application). But I don't see why people complain becuse they settled for "good enough" instead of designing "the best" for a given function. And I, personally, think that anyone who goes back to 1.05 after working with 3.0 is insane. I couldn't live without 3.0 these days. chuq Chuq Von Rospach chuq@sun.COM [I don't read flames] There is no statute of limitations on stupidity