Phonetool: Integrating . . .

In this paper we argue for computer workstation access to voice telephony functions. We describe several computer based communication tools developed in our laboratories, and explore one, Phonetool, as a detailed example. We discuss user interface issues for collections of such utilities. Finally, we discuss several architectures for interaction with the PBX and the communication requirements between workstation and switch. This paper espouses the need for computer workstations to control conventional telephony functions, concentrating on a graphical dialing tool as an example. By way of this example, we will explore several architectures for software and hardware access to PBX equipment. An underlying motivation for this work is a vision of the workstation as an integrated and personalized comnlunications node, capable of interacting with a variety of media and transport mechanisms. Many portions of this vision are already in place, but not integrated. Users store phone numbers in on-line databases. Computers are being used to send and receive facsimile, including fax/e-mail gateway services. Voice mail system are computers that intercept calls and digitize speech, though tlicy don't yet provide a means to access the stored voice from the rest of the office con~puter environment. A greater variety of services (conference, forward, transfer, pickup, etc.) arc becoming more widely available on digital PBXs; the addition of control interfaces on PBXs opens the possibi1it.y for worlrstations to perform intelligent call setup and routing using these services. Our position argues that as office computer usage increases, greater functionality is gained from allowing computers access to telephone switches than by building more complex telephone sets. This view is not unique to the authors. The MICE project a t Bellcore [2] emphasized user interface and personalization of services in a network testbed environnlent. BerBell, also at Bellcore [3] built a programmatic interface to several switches. Etherphone [S, 91 a t Xerox PARC explored alternate transport mechanisms and expanded use of voice in gffice workstations. Though less ambitious (it is a product, after all) Northern Telecom's Meridian Mail gives networked personal computer users integrated graphical access to voice and text messages. DEC's computer^ Integrated Teleplmny, ClT [7], defines a model for linking telecommunications and computing systems for functional integration. Current work at Olivetti Research Center [I] focuses on developing software architectures for sharing voice and tele~hone functions between multiple processes and applications. 1 Computers or Telephones? Why do we want to interface telephony functions to computer workstations? The reasons fall into two broad categories: improved user interfaces and enhanced functionality. Workstations are already much better suited than telephones for improved user interfaces. They have far superior displays, typically CRT screens; in part because of power requirements and cost, telephone displays are small and difficult to read. Workstations have better input devices, with full keyboards and the ubiquitous mouse. Most important, a variety of powerful window systems provide rich and flexible graphical interfaces (see figure 1). Figure 1: "Iconicn form of Our tools These user interfaces may be customized; recent window systems provide ample opportunity for users to configure the mapping from input to function, as well as appearance (color, font, etc.). There is every reason to extend this personalization to telephony functions, such as: how calls get routed as a function of time of day, calling party ID, personal schedule, etc. the names and number of entries in a speed dialing menu a voice mail prompts and interaction with frequent versus casual callers Such personalization is difficult to control with the traditional 12 button telephone interface. We are already a t the point where new telephone functionality is infrequently used because it is too difficult to access. Personalization may be difficult to implement, in switch software, with its emphasis on state tables and real-time requirements. At this stage in the design of user interfaces, we require flexible operating systems and good software development environments. Design will be iterative and interfaces experimental for some time. Since current workstations already include window systems, there is no meed to duplicate that sesource elsewhere. Even a simple resource such as memory is more effectively used when it can be shared with a number of applications, and allocated according to moment-to-moment needs. Enhanced functionality is the result of new services and, most important, integration at various levels. Integration comes in many forms, such as: 0 linking databases to functions, e.g., dialing from an online directory a communications orientation of many computer-accessible media (electronic mail, facsimile, modems) e increasing potential for use of new media, such as voice in computers and video in telecommunications At least three elements are needed for these new media to be useful: transport, storage, and user interface. By combining telephony with the workstation to achieve all three, we envision a scenario such as the following: While out of my office I receive a call from my co-author; a voice message is taken. On my return, I use the mouse to play (and repeat) portions of the message, then I call him back via my speed dialing tool, using calling party ID if available or his phone number from an on-line rolodex. While on the phone, we both listen to pieces of his previously recorded message, under my control, again with my mouse.