Greetings,
I'm an absolute newbie to this Linux world.

After a bit of faffing about, I managed to get debian 3.0.r4 installed and working on a fairly venerable hardware set. The fun bit was getting the wireless interface to work (Atmel chip on a USB device).

Thanks to the guys who wrote the driver and supplied the info on ifconfig and iwconfig this worked pretty well.

Having been through this route, I discovered that the usual way around was to use ndiswrapper and the ndis drivers from Windows.

Does anyone know if this method would work with the Atmel drivers?

Regards
RB

To be honest, ndiswrapper should be an absolute last resort. It is a nice piece of work that allows a lot of people to use wireless cards manufactured by complete bozos, but it really is an act of desperation to use Windows code in linux. Since Atmel chips do have native Linux drivers, in the long run you will be much better off using them than ndiswrapper.

Just a little aside on the "Bozos" part (and I agree with Hangdog42, by the way).

The lack of availability of native Linux drivers for many recent chipsets and the need for work-arounds such as ndiswrapper or Linuxant has been primarily the result of a technological advancement - software radios.

In the good ol' days, control of the radio was hard-coded into the chip. Firmware upgrades could affect some things, but most of the critical stuff was inaccessible.

Atheros, Broadcom, et al, introduced designs that lessened the reliance on hardware and allowed feature/function to be controlled via software - the driver. Advantage to manufacturers: reduced fabrication costs, reduced time to market and simpler fixes for design flaws. Advantage to consumers: lower cost, better features, quicker fixes.

The problem with this arrangement is that the "owner" of the software also owns the card's behavior to a large degree. Opening the door completely to the inner workings would mean that creative programmers could make the card do things the manufacturers never meant them to - operate outside of their assigned frequency or power range, for example. The FCC and other agencies charged with controlling the air waves tend to frown on such things.

Even Linux wireless drivers developed with full or partial support from the makers - Intel, Atheros - still rely on a closed piece of code that keeps malicious or careless coders from wreaking havoc.

The Bozos at Broadcom, in contrast, have steadfastly refused to offer any support or assistance. Necessity being the mother of invention, that intractable position forced programmers to develop ways to wrap the proprietary Windows drivers in a shell that could talk to both the kernel and the driver. The open source community fix is ndiswrapper. Linuxant took up the challenge as a commercial venture.

I'm sure there are other chipset manufacturers that deserve a head smack as well, but Broadcom stands out because of their market share. Some of the recent cards made in China seem not to work no matter what amount of head-standing is applied. Some of the card makers have at least attempted to supply a Linux driver, although only in rpm or so badly written they wouldn't compile. I guess they get a gold star for effort.

Sorry for the long-winded ramble. This has probably been covered before, but seemed like a good opportunity to add some detail behind the mess. My advice is always to do your homework before buying a card and support those companies that are supporting your desire to run Linux.

Thanks for the explanation, I've never really understood why the current situation exists other than blatant greed/stupidity. Actually, it sound eerily similar to what happened with real hardware modems and winmodems.

I guess the part I find most frustrating is that with the prevalence of wireless connections, a new Linux user is going to get smacked with this problem almost immediately, and almost certainly before they've developed the understanding and skills needed to solve the problem. And this is a big enough obstacle to cause people to give up. I mean really, who wants to use a computer you can't connect to a network? To be honest, I've never understood Broadcom's position on this. Nobody is asking them to develop or support Linux drivers, only release enough information so someone else can. They do it with their wired chipsets, why not wireless?

Greetings All,

Thanks very much for the posts.

Yesterday I tried ndiswrapper (which sounded like a good idea to me because NDIS is defined) plus Belkin F5D6001 Windows Drivers. It looked as if it might be persuaded to work but did not.

Today I tried the Linux native driver from Andrea which worked very well. I agree with the comments about lack of drivers making it difficult for newbie's to get Linux systems going but seem to remember that it's only relatively recently that getting drivers working on Windows was straight forward.

Sooner rather than later as the GUI Linux community continues to grow it should be obvious to the manufacturers that they could be missing a trick or two.

In the meantime, many thanks for the reply; I am better informed if not wiser.

Roger Bowen