'Feature creep' imperils Bluetooth, execs warn

'Feature creep' imperils Bluetooth, execs warn
By Peter Clarke, EE Times
June 8, 2001 (3:58 p.m. EST)
URL: http://www.eetimes.com/story/OEG20010608S0053

MONTE CARLO, Monaco — Senior executives within the Bluetooth Special Interest Group stressed the importance of shipping products based on version 1.1 of the core Bluetooth standard and told developers to beware the problems of "feature creep" at this week's Bluetooth Congress.

Feature creep occurs when additional features are continually added to a product specification during development, which delay a product's completion and market introduction.

The incremental nature of Bluetooth development, previously expected to go from the present version 1.1 through a series of versions up to 2.0, is now being cast as dangerously similar at a time when Bluetooth is under pressure to fulfill its promise as a ubiquitous short-range wireless communications standard.

The new high-level philosophy could indicate that the tide is turning against plans to develop a higher data rate option for Blueto oth that was due to arrive at version 2.0 of the standard. That version of the spec was expected to support 10- or 20-Mbit/second data rates.

The Radio-2 working group, chartered with development of the higher data-rate standard, has yet to show a compelling market need that cannot be met by another standard, a senior member of the Bluetooth Special Interest Group (SIG) said. And without such a market need, technical work cannot begin, he said. At the same time, the IEEE 802.11a wireless LAN standard has somewhat eclipsed the need for a broadband data rate version of Bluetooth.

"At the start of the [higher data rate] activity there was some hope of beating 802.11a to market. That didn't happen. So there has been some repositioning," the senior executive within the Bluetooth group said.

Indeed, the SIG is so anxious to build on what has already been developed, that it has decided to change terminology and avoid the use of the Bluetooth-2 or Bluetooth 2.0 terms.

Simon Ellis, strategic dev elopment manager at Intel Corp.'s mobile communications group and chairman of the marketing working group within the SIG, said the previous terminology had suggested too big a break between what was presently available and what was planned, and that could lead to companies opting to wait for later versions.

Instead, he said, SIG management wanted to freeze the Bluetooth core standard version 1.1 for "more than 12 months" to give companies a fixed platform with which to design. Ellis said that additions will be made through application profiles.

Stable spec?

But it is not clear that the Bluetooth 1.1 core is stable enough to satisfy many companies, and others will always seek to be standard-compliant while offering proprietary enhancements. For example, Scatternet operation, though covered in V1.1 of the standard, is widely held to be incompletely described and in need of additional clarification. And Infineon Technologies AG cause d a stir at the Bluetooth Congress when it put a proprietary 4-Mbit/s capability onto its Bluetooth chips.

Standing in opposition to the danger of feature creep inhibiting adoption is feature freeze, which could lead to fragmentation of the standard.

Motorola Inc., which is one of the nine "promoter" companies at the heart of Bluetooth, pledged at the Congress that it would lead the development of 10-Mbit/s Bluetooth. But it admitted that the community was divided on the need for the 10-Mbit/s option.

"I don't think there is a clear assessment," said Bill Dunnigan, general manager of the wireless local connectivity division of Motorola's Semiconductor Products Sector. "IEEE 802.11a can't handle the mobile piece of it.

"We see some applications that will hit the sweet spot [of 10-Mbit/s Bluetooth]. You have to trade off power and cost," Dunnigan said.

But, challenged that there was not, as yet, a compelling need for 10-Mbit/s Bluet ooth, Dunnigan said, "You'll probably hear half and half around the Congress."