Wednesday, August 13, 2003

Telirati Newsletter #7

Of my old newsletters, this one is still almost fully relevant today. Telephone user interfaces are as awful today as the first horrid attempts at GUIs for placing phone calls. Even the extra features on mobile phones don't integrate well with the fundamentals of dialing a call. Progress? It can be summed up in one phrase: on-hook dialing.

Address books, calendars, call logs, and messaging are as un-integrated as ever, in both the widely used mobile user interfaces and in the finally-becoming-mainstream desktop IP telephony interfaces. Smartphone UI, with a completely flexible touchscreen interface never delivered progess, either.

What does this tell us? For now, it means that, even with 400 million handsets delivered each year, nobody is putting a serious effort into fundamentally improving the way existing mobile phone features integrate with one another. It means that SMS and MMS are no closer to integrating with voicemail than they were in 1997. It means that there isn't even one application that uses call control in an innovative way on the desktop or in mobile phones. Bah! And people wonder where the bad karma of the telecom crash came from?

Newsletter #7: User interface

I like nice artifacts. A car that looks right and drives right. A kitchen implement that adds to the pleasure of making and eating a meal. A house that speaks to you about the life inside it. We spend so much time in front of our PCs that user interface is a matter of the quality of one's life these days.

No wonder then, that some people look at the Web as the work of the devil. In newsletter #6 I spoke of the apparent end of history when Windows 95 was announced, and how the Internet unraveled that end. One reason why Windows 95 was such a milestone was that it marked the point where Microsoft had effectively eclipsed the Macintosh in every important way, especially in user interface.

Up to that point, it was the Macintosh that lead the way to a future where strict rules would ensure that every application in the desktop metaphor was similar enough that, just like the papers and files in our offices, every document and tool for manipulating documents was obvious. Operations would be standardized -- made generic. Pick an object, perform an action. That's it. No wrong choices are possible because once the object is selected, only the valid actions are highlighted.

The world on screen was headed for a kind of super-reality where direct manipulation of on-screen objects made us unerring draftsmen, arithmetic geniuses, and creative typographers. History would be a smooth continuum where the remaining ugly "dialog box" inheritances of our menu-based cave-man life would be gradually replaced by ever more powerful direct manipulation of documents on screen.

Sure, there was some backsliding. "wizards" are no more than a hideous rebirth of those lame-brain suggestions we've all encountered early on in GUI evolution: "Gee if you just made a sequence of dialog boxes appear, we could port our menu-screen software to the Mac!" But all in all, the virtual-desktop super-office world was coming to life on our screens. Homogenous office documents, all happy at home framed by powerful, yet familiar and uniform tool-bars. Then wham! The Web.

In contrast to the uniform, business-like world of the desktop metaphor, the Web is more like the pop-music magazine metaphor. No standards. Intentional violations of convention. Loud, clashing graphics. When you hear Microsoft's principal software architects dissing Java, the subtext is that the whole Internet has put a wrench into the works. Object models, messaging architectures, document formats, the architecture of multi-protocol networking, and, not least, the user interface are all brought into question by the Internet.

Why is the question of user interface interesting to telephony software developers? Part of our industry's problem is that, so far, user interface is a stepchild issue among the big issues of computer telephony. To take one example, the state of the art in interface to the on-screen dialer is generally shameful. A picture of a telephone on a computer screen? Is that the best we can do? Fake LCD displays? Heck, I've seen fake wood-grain on PC phone interfaces. Can an industry that created an icon of modern life, an artifact that is as emblematic a part of civilization as the water faucet or the automobile steering wheel, be so ignominious in its failure to translate the basic ideas of telephony into the computer user interface milieu?

The dawn of the Web as central focus of user interface ideas has brought us to a crossroads: one way lies Gomorrah, using the Web as an excuse to return to the bad old days of pictures of telephones on the screen. The other road leads to a shining city on a hill where for the first time in a century, we may have a fundamentally better user interface to the telephone. Viewed in the context of the number of telephones and their far greater historical span, the telephone user interface looms large among computer user interface problems.

Before we delve deeper into the telephone user interface problem, lets see where battle lines are currently drawn. Up to this point, the infrastructure of our desktop super-reality has been an increasingly powerful set of generic operations accessed by selecting menu items or pressing toolbar buttons. These toolbars became attached to the standard architectures of the underlying programs. In Windows COM applications, embedded objects could take along pieces of their "home" menus and toolbars and cunningly and complexly merge them into the menus and toolbars of the applications editing the documents in which these objects are embedded. This in contrast to the Web world, where designers slang about "blowing away the chrome," their term for the gray menus and toolbars that encrust documents like user interface barnacles.

Trouble is, so much of what passes for user interface design discussion these days is just so much "anti." Anti "chrome," anti desktop metaphor, etc. Add up all the anti and no wonder flitting from Web page to Web page seems at times regressive. Back to the bad old days of one screen of menus leading to another screen of menus. Some contemporary commentators on user interface go on about speech recognition, or 3-D, without realizing that it's not 2-D that makes interfaces the way they are. It's not even bitmap displays. The desktop metaphor is rooted in virtual coordinate systems first invented for CAD applications so that objects could be moved without disturbing their internal data about their dimensions. Until the would-be Web-based alternatives offer some real alternative to the logical structures underlying the desktop metaphor, it cannot replace it with something more powerful.

The hope of replacing the architectural underpinnings of the desktop metaphor might well be in vain. Let's look at the way humans have consumed information for millennia: marks on flat surfaces. Paper and ink, the objects modeled in the desktop metaphor are but a recent adaptation of a fundamental human communication form. So while modeling the office work environment might well be on the way out, it would be premature to think that marks on a flat surface, whether smudged by the thumb of a cave man or traced by electrons on a screen, will cease being our main non-verbal communication.

Post-literacy has been cited by some as the downfall of desktop metaphor. We will return to the tribal village of emotion and physicality, conveyed across space by our video images. Faugh! Even McLuhan who first elucidated the relationship among types of media did not think that new media and their evident enabling of the return of tribal structures and interactions was a good thing -- it was just a thing, a fact, and one he could explain uniquely well.

So stuck between the imperfect implementation of the desktop metaphor, and the ill-understood changes driven by the Web, the would-be designer of the next telephone user interface stands, far worse-off than before. What can he hold on to? What should he let go of? Where will the new world of user interface design lead?

Modeling reality vs. the document metaphor: This is first hurdle of user interface design in computer telephony, and one that few contestants overcome. Should a telephone user interface model the look of a telephone? The answer is yes and no, but clearly and resoundingly very much mostly "No!" (N.B. we cannot in this space answer the deeper question of whether a telephone user interface can be entirely put on a computer screen -- that is, whether the receiver should be as fundamental a part of the physical complement of a computer's user interface as the mouse and the keyboard.) But we can say a hearty "no" to pictures of phones on the computer screen. The telephone operates a network, creates a paper trail in the accounting software at the CO, and takes input from our knowledge of who we want to call, when, and at what number. These are the deeper fundamentals of telephony, not the keypad and especially not the feature-buttons of a telephone. Translate these deeper fundamentals into operations -- document oriented operations -- on the computer screen, and you are on the right track to a telephone user interface.

Windows look vs. Web look: Well, if we've cleared the hurdle of fake LCD displays and flagrantly breaking the generic-operation model of the computer interface by putting a telephone's button's on the screen, we now need to check which path to take. Do we translate the phone into a Windows COM program, with multi-tab dialog boxes and neat rows of toolbar buttons? Or do we step into the intimidating open spaces of Web design? The latter looks more risky, but consider how many things Microsoft has thrown overboard in their pursuit of dominating the Internet scene. Will Windows go "chrome free?" Will the Web kill dialog boxes that remained impervious to the desktop metaphor drive toward direct manipulation? My bet would be with the Web. Kill your toolbars. Slaughter your dialog boxes (and good riddance). Annihilate your gray beveled buttons.

In other words, do things the Web way for the right reasons. We should have the freedom to shape our user interfaces distinctively. But don't abandon the elements of the desktop metaphor that are rooted too deeply to be moved by the Web: The act of phoning people produces a document, a record of who you have phoned, a plan of who you will phone, and notes on what was discussed. Phone calls have relationships to other documents -- mail messages, address book entries, etc. A picture of a telephone on the screen captures none of these relationships. Do this right, and you have a chance at making a genuine improvement in the way people work.

Copyright 1997 Zigurd Mednieks. May be reproduced and redistributed with attribution and this notice intact.