Reports from the Electronic Frontier: Writing Tools, Part 2 Tom Maddox Donald Norman is interested in the ways people use technology and how it uses them. In Turn Signals Are the Facial Expressions of Automobiles(Addison-Wesley, 1992) he says, "Much of modern technology seems to exist solely for its own sake,oblivious to the needs and concerns of the people around it, people who, after all, are supposed to be the reason for its existence." This common enough lament acquires special significance as part of Norman's attempt to refine the human dimension of technology. From this book and others by Norman a common theme emerges: it is difficult to design technology so that it works easily and well. Ubiquitous technologies such as VCRs illustrate the theme all too clearly. As comedians are fond of pointing out, in millions of home across the United States VCRs sit blinking in idiotic repetition because no one in the vicinity has a clue about how to set the time on them--or, more subtly, because no one remembers how it was done six months ago, on the previous occasion of the power going out. And though instructing theVCR to record programs while its human owners are asleep or absent is one of the primary uses of VCRs, actually doing it is so beyond most users that a rather complex support industry has sprung up to simplify the process, involving code numbers lifted from TV Guide and clever black boxes that can speak universal VCRese. And we are all familiar with a multitude of other design flaws. Some make our lives more difficult in subtle but irritating ways--such as doors whose construction cries push but whose function demands pull or vice- versa--while others can cause death and destruction--such as confusingly laid out airplane cockpit instrumentation. Such concerns came to mind as I was working with several kinds of software intended to aid the writer. As desktop and portable computers have become more powerful, the word processor has been augmented by a number of subsidiary programs focused on particular writing problems. Among the more numerous freeform database managers, and I recently looked at several recent entries into this market for the Macintosh. Spiral (Techworks, Austin, Texas) is the name of one that has been widely advertised and reviewed lately. It is designed around the metaphor of the spiral notebook, a simple and powerful metaphor, but one it inflicts on the user without let up or remorse. In particular, it presents information on utterly inelastic "pages" entirely inappropriate to an electronic medium. This peculiarity requires the user to reformat a document using some strange dodges detailed by the program's manual, which shows that the company is aware of the problem. Spiral also allows categorizing the pages according to sections, which is nice, though not particularly innovative. However, even at this task it does things peculiarly, through an interface that still makes no sense to me. Spiral is also slow. In short, while it theoretically performs the admirable task of preserving information in useful form, so far as I am concerned, practically speaking it constitutes a kind of software torture device. NoteTaker (NoteTaker Software, P.O. Box 10270,Truckee California 96162), a less-publicized entry, uses the combined metaphors of the outline and the notecard. "Topics" are its outline headings," notes" its notecards, grouped under the headings. This seems to me a much more workable scheme than Spiral's pointless "pages." And in fact one can work one's way around a NoteTaker file much more easily than a Spiral file and can enter data in more straightforward fashion. In short, NoteTaker is not a bad program. However, it suffers from a certain slowness as its data structures are wheeled into play--in particular, generating a list of topics takes more time than it should, and one begins to dread the appearance of the "Generating List of Topics" bar. Still, however, it is a decent attempt at a freeform database. Working with Spiral and Notetaker caused me to recall a now rather antique program called Tornado Notes, once (and possibly still) available for MSDOS machines. It performed what seemed to me when I first encountered it a fairly magical feat: once invoked (by pressing alt-j, I believe) it filled the screen with notes, electronic Post-It note, really. Depending on certain constraints, it could present hundreds of them in sheet after sheet, cleverly tiled so that one could see a whole screenful at once. And the program was fast. One could quickly scan the tiled sheets or locate specific words or phrases. As a result, despite a command structure that seemed entirely arbitrary(which required that so its less-used commands be relearned periodically) Tornado Notes was an exceptionably useful program. It allowed me to dump into it stray facts, conjectures, ideas, questionsany old text fragments at all-- and to find them again when I was working on something else. In short, in a very simple and straightforward way, Tornado Notes accomplished much of what its much more sophisticated successors work at. But in my opinion it did its work much better. Free form databases tend to demand too much structure and to work too slowly. Perhaps more crucially, they tend to demand that you know what you are looking for, unlike Tornado Notes, which would show you all your little bits and pieces of information at lightning speed (those tiled sheets of notes) or let you search for what you thought or hoped might be there. So as I sit at my relatively powerful Macintosh, whose capabilities far exceed those of the old Leading Edge D on which I ran Tornado Notes, I long for a program that will do what that old program did and do it as quickly. Unfortunately, for reasons I will discuss below, I doubt that anything so simple will emerge from current software development environments, but I would love to be proved wrong. Dictionaries and thesauruses have also emerged lately as a viable software category. While word-processing programs have employed spell-checking functions for years, only lately has the combination of large hard drives, fast processors, and efficient compression schemes made possible genuine dictionaries and thesauruses on desktop and portable computers--that is to say, ones with definitions, guides to pronunciation, and etymological information attached to words. And unlike the freeform databases I mention above, these programs do something new (at least on a desktop computer) and do it with a certain elegance and speed. I am particularly fascinated by an electronic dictionary's capability to find all instances of a word quickly; for instance, to show the user every definition in its store in which the word quantum appears. With the electronic thesaurus, one finds the corresponding ability to follow a semantic trail, to skip quickly from word-group to word-group without having to search headings or turn pages. Here, as with word processing, the electronic page seems to me to contribute something that its non-electronic counterpart cannot duplicate. While a sheaf of pages (the book, remember?) still provides the most effective random access, the electronic book makes possible a more effective directed access. However, even here, where technology provides special advantages, issues of design remain difficult. For instance, the most widely available of these dictionaries is the American Heritage Electronic Dictionary, which features quick searching, a broad database, and a pleasant enough interface, except for a few elements of design that result in chronic annoyance. The most telling of these occurs every single time one uses the dictionary. One invokes the dictionary from within a word processor, let us say, with a word selected. In other electronic dictionaries, the selected word automatically pops into the dictionary's search window, and so it takes one keystroke at most to find it. However, the AHED requires the user to go through a little series of commands, none of them complicated or difficult, but all entirely pointless. And, let me repeat, the user has to repeat them every single time. What I used to find mystifying about such deep and obvious design flaws is that they should have revealed themselves to anyone who bothered to use the program. But Donald Norman has an answer to this puzzle. He points out that design of such products is in many ways like writing and that people in general are not very good at either. He says that both require an empathetic identification with the audience-- the user or reader--in order to create a product or piece of writing that can be easily used or understood. As a writer and writing teacher, I will second his comments. Novice writers often have the most difficulty getting into his reader's head. They thus fail to provide rhetorical and logical connections and, more generally, are prone to say things without regard for the puzzlement caused to a reader. Of course it is not enough simply to be empathetic: the software designer and writer must put their empathy at the service of considerable skills to create a workable piece of software or effective piece of writing. However, Norman does not concern himself greatly with other, equally or more important factors that shape (and often distort) the development of software. Most software is written to make money, and with an eye to the fact that a quite alarming amount of money has been made in recent years from software. So a dictionary, for instance, is made to sell, above all. Let me quickly point out that I am not claiming that this should not be the case; I am merely pointing out that market pressures are at least as significant as formal design pressures in shaping software products. To briefly illustrate what this means, let me invoke the analogy of automobile design, manufacture, and distribution. Those of us aged enough to remember the cars of the '50s recall with ironic fondness the tailfins and chrome excrescences (such as portholes) that seemed essential to a Really Fine Automobile. Many such devices now appear in software as "features,"capabilities of such unusual, subtle, or arcane nature that few users will understand them, fewer still employ them. Recall also the now time-honored practice of planned obsolescence, by which this year's model is forcibly distinguished from last year's. And of course devices such as tailfins have been used to make the distinction. Such practices, while perhaps understandable,even admirable in some circumstances (they contribute greatly, after all, to the dynamism of both the automobile and computer industries), can have awkward and unforeseen consequences. 20 The '59 Chevrolet, I recall, had tailfins of such sweep and grandeur that in certain situations they would lift the rear of that automobile right off the roadway.I leave it to the reader to discover his or her own candidates for a software equivalent. To sum up the point of this analogy, current marketing practices do not encourage the development of stable, powerful software. I might wish to see the development of Tornado Notes-like programs, greatly improved and updated, but so what? Perhaps there really is not a lot of money to be made providing simple, effective tools for writers -- an impecunious group anyway, generally. Finally, I should point out that really powerful metaphors (which seem necessary for creating this kind of software) do not come along very often. To create a genuinely new, useful, and powerful kind of software, I suspect that one needs not only the tools of the programmer and the empathy of the writer but also the luck and brilliance of the artist--you need to be there at the right time with the right metaphor. So it isn't hard to understand why mostly we find sizzle, not steak, in new software. News From Online: The major online conferencing systems-- Compuserve, GEnie, Prodigy,and America Online--have all changed their pricing schemes lately, which no doubt portends something or other but I don't know what. As to Prodigy, I cannot say much, and the only favorable thing I can say is that it remains alone among the services in offering 9600 baud access without an ugly surcharge. Otherwise, I can justsay that serving as Your Reporter, I joined Prodigy but quickly fled in horror from it. Not only is its interface of surpassing ugliness (it seems to have been designed by a color-blind person), not only does its communications program grab control of your computer in a way suggestive of demonic possession, but also it does not provide anythingof interest sufficient to offset the relentless flow of advertising. At 9600 baud or any other speed, I find it a nightmare. Your mileage may vary. GEnie appears to be switching from a two-tier to a one-tier system and from a largely flat-rate system to one that charges by the hour, while Compuserve has gone the opposite direction, and America Online appears to be giving more connect time for less money. However, if you are interested in what the rate changes mean to you, I would advise you to take a close look, in detail,in terms of your usual pattern of online use. Do the arithmetic, in short, because I can assure youthat they did. For instance, if you have been used to hanging out hour after hour on GEnie, you will almost certainly find your bill greatly increased. Internet address: tmaddox@netcom.com ======================================================================= This document is from the WELL gopher server: gopher://gopher.well.com Questions and comments to: gopher@well.com .