This was reflected in a story I did about a Linksys router. Turned out it was running Linux and the company had to be pressured to live up to its responsibility on that. It was also running a Broadcom chip set.
The two facts were related. It's hard to live up to your responsibilities as an open source hardware vendor if your upstream supplier is also reluctant.
Times have changed, and you'll get many of the details at our newest blog product, Steven J. Vaughan-Nichols' Networking.
Not only are developers getting a native stack that supports its most popular 802.11n chips, but there's a framework for supporting future chips as well.
Broadcom had been staging a sort of cold war with developers for years, to the point where some developers were recommending against using Broadcom chips in their designs.
What made Broadcom see the light? I don't think it was the whinging from netbook owners or the open source community. The industry is changing.
Where the big money used to lie in the consumer market, now hospitals and other campuses are building extensive WiFi mesh networks. These industrial-strength networks need to be carefully engineered, with security and redundancy built-in.
Native Linux drivers for Broadcom chip sets lets hotspots become integrated into real networks more easily. There is now a market for that.
Intel has been using open source in its competing systems, and my guess is they were starting to make inroads into Broadcom's business as a result. They were eating Broadcom's cheese -- a big change from the past.
What media and hobbyist and political pressure can't do, a little market pressure can do very well. Which means some of the mesh technology being developed for hospitals may soon become available at a hotspot near you.