[lwip-users] RE: [lwip] Wireless Ethernet

[Stefan Beyer]

On Fri, 1 Mar 2002, DASILVA    VINCENTE    BIT wrote:

> Thanks for the input Kieran...
>

<snip>

> When I started with lwip I had not done embedded programming for years, had
> never done anything with tcp/ip stacks, never used the Keil compiler, never
> used a real-time embedded operating system and had never worked with
> multiple threads, semaphores and mutexes on a low level. For people who
> start like this the more documentation or references to documentation there
> are the shorter the learning curve and the quicker the adoption of
> lwip...the quicker the adoption of lwip the quicker we will have more
> utilities/bug fixes for it...

I agree, that some things could be better documented. I managed to find
out everything I needed to know from the mailing list archive, but would
have prefered to read it in anice indexed document. However, the problem
with documentation on projects like this is, that no-one bothers to write
up on what they have found out. I myself could not be asked write
things up, because I had to move on after porting lwip to my platform. A
TCP/IP stack is for many people a bit of infrastructure they need to move
on to their applications. Once it is out of the way, they don't tend to
communicate their experiences back.
Also, some of the things you mentioned above (semaphores, threads, etc.)
are not lwip speciifc and I can't see why they should be documented with
lwip. A decent Computer Science text book is probably abetter way of
finding out.

> Also, as far as developing a wireless Ethernet device/driver I was targeting
> the people who are creating Ethernet devices/drivers, not the ones that are
> not. I'm sure these people would agree that in most circumstances Wireless
> Ethernet is better than Wired Ethernet, Wireless Ethernet is the future of
> networking...

I fail to see the difference. Any ethernet driver does essentially the
same thing: moving packets onto the network and reading packets of the
network. Any decent skeleton code can guide people in the implementation
of drivers for their hardware. It should be a goal to encourage people to
publish any driver implementation they have done, wether wired
or wireless.
Maybe I am misinterpreting what you want to do regarding wireless
ethernet.

Stefan Beyer
-- 
address@hidden
http://www.cs.man.ac.uk/~beyers8


[This message was sent through the lwip discussion list.]