Microsoft and Intel Promise Interoperability through New Home API

It's a little daring to express enthusiasm for an ambitious new home automation initiative that promises-like so many others before it-to "usher in a whole new industry resulting in the unprecedented control of residential environments." That's just one of the claims (attributed to Deb Massof, vice president of Honeywell's Home Vision group) profered in an upbeat press release about the new Home API, an open industry specification being defined and developed by Compaq, Honeywell, Intel, Microsoft, Mitsubishi Electric and Philips Electronics.

The goal of the the group is to provide a technical foundation-through application programming interfaces, or APIs-for enabling the control of a wide range of consumer devices. These APIs will be protocol- and network media-independent, enabling software developers to more efficiently build applications for multiple communications platforms.

Sound familiar? Where else have we heard protocol promoters promising that their contribution "will stimulate the development of hundreds of new applications that integrate… devices in a way that adds real value to family and home life."? (attributed in this case to Ed Arrington, manager of Intel Architecture Lab's Anywhere in the Home initiative.)

But really, this one is different from the other initiatives-CEBus, Home Plug and Play, LonWorks, FireWire, etc.-that have promised similar results. "When you look at all the pieces you have to put in place to extend computing power around the house, one of the things that's been missing is a single API," says Dave Redelfs, technology marketing manager for Intel's Anywhere in the Home effort. Widely embraced APIs can quickly turn weak, fragmented industries into cohesive, vibrant entities with healthy product lines. That's what TAPI's done for PC telephony, SAPI's done for speech recognition and MAPI's done for mail management.

Why can't HAPI do the same for the perpetually "almost-there" industry of home automation/networking?

It's not quite so simple. Other popular APIs are more or less oriented to a single function and one or two standard communications media, e.g., telecommunications. By contrast, a viable Home API must address a wide range of functions, such as telecom, security, entertainment, and Internetworking. It must consider products as diverse as light switches, televisions, HVAC systems and Internet appliances.

Because the phenomenon of "convergence" came long after the proliferation of dramatically diverse home electronics, after-the-fact networking and automation specifications have been particularly difficult to implement.

Thus, applications developers have had to write all aspects of home device control from scratch, including the handling of network interface hardware and device control protocols. With Home API, the application developer can ignore these low-level details and focus instead on adding the features that provide direct benefit to the user. "It frees the applications programmer from having to care about what the transport media is," says Redelfs. Adds Intel engineer Dave Doerner, "Developers can write an application once instead of writing it for every protocol," meaning Home API-compliant applications can operate in virtually any household, rather than a limited number of homes that employ one or two of dozens of automation platforms.

Attractive applications environments and wide-open markets, it is assumed, will spawn a prolific community of automation software developers, and with it receptive retailers, enlightened consumers, and a growing movement among PC manufacturers to package free networking software with their computers (which will in turn drive sales of networking modules and other peripherals, much like free gaming software spurs sales of joysticks and the like).

To support developers, the Home API group plans to provide a Software Developers Kit for Windows, expected to be available by the first half of 1999.

Because of the openness of the specification, applications developers can implement HAPI in other operating environments such as Sun's Java, real-time operating systems from WindRiver or Microware, and other non-Windows platforms.

Vendors of home networking and control products will provide HAPI-compatible driver software to expose features of their hardware to developers of home applications. Since Home API does not define communication protocols but rather provides a framework for integrating and exposing existing protocols, it is not necessary for device or network vendors to modify their hardware products to enable them to work with Home API.

Home API, its promoters stress, does not compete with such technologies as HomePNA (phoneline networking alliance), HomeRF, CEBus, LonWorks and X-10. Rather, it is intended to help such technologies coexist.

Privately, Home API supporters say they were initially unclear as to how the API would mesh with the digital A/V networking standard Home Audio/Video Interoperability (HAVi). But it is clear to them now that the two technologies are compatible, and work at different levels.

"We believe that Home API and HAVi complement one another and target different clients," explains Eddy Odijk, Philips Consumer Electronics' technology director working on home networking initiatives. HAVi focuses on IEEE 1394 A/V devices such as TVs, VCRs and set-top boxes. Indeed, Philips and Sony-both members of the HAVi working group-are working with Intel and Microsoft to define a bridge between Home API and HAVi that will enable each to use the other's services to expose home devices to their respective clients.

The size of the initial Home API working group was intentionally kept small in order to speed up the development process; however drafts of the spec will be widely circulated to gather input from other parties, and to generate quickly a critical mass of supporters that will be key to the success of the initiative. "We want to get in there and get the job done," says Redelfs. "The goal is not to make this group last forever."

For Intel's part, Home API is just one layer in the company's vision of home control for the masses. "We won't stop at the development of a programming environment," says Redelfs. "We're not done with a capability until we have lots of users using it. We look at the whole stack-what needs to be done with the hardware, operating system and applications-as well as things like end-user marketing."

That's why Intel is also involved in workging groups such as HomeRF and HomePNA that address other layers of whole-house control.

Everybody Get HAPI?

So, then, should we all rally around yet another home-control initiative? If nothing else, suggests Home API supporters, look at the players involved. "When I hear that there are standards happening, I look very carefully at who's on the standards committee. If it's an API, and you're looking to get it into Windows, you have to wonder: Is Microsoft involved?" says Redelfs.

"Having Microsoft and Intel behind it would've been enough for me," says Tim Shriver, president of Home Automated Living which markets the HAL line of PC-based home automation systems. "But Home API really does give the industry some glue to pull this fragmented industry together."

Just as HAL has recently made its software TAPI-compliant in order to work with virtually any voice modem on the market, the company plans to make future generations of its software HAPI-compliant. "With interoperability no longer an overriding issue, consumers will be free to shop for the best technology and that is where we think the competition should take place," says Shriver.

Ditto for Savoy Automation, whose next version of CyberHouse software (3.0) is expected to be HAPI-compliant.

Other home systems manufacturers, like AMX's PHAST, manufacturer of the high-end Landmark home control system, are a little more cautious. "We, like other manufacturers, are trying to figure out what the best utilization of our time and effort is," says Dave Hill, AMX vice president of residential technology. "It's somewhat of a confused marketplace. We are evaluating several activities in the marketplace. One of them is HomeRF, one is HAVi, one is VESA, one is Home Plug and Play…. I've got people full time on FireWire. I'm on two or three standards committees; Tom Hite [PHAST CEO] is on two or three committees….We aren't blameless for the confusion. Of course, we'd like to continue to push PHASTLink [the PHAST networking protocol], too."

http://www.homeapi.org