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

[Kieran Mansley]

On Fri, 1 Mar 2002, DASILVA    VINCENTE    BIT wrote:
> I was thinking that there are a lot of people working with/on lwip. A lot of
> them are developing drivers for Ethernet hardware. There are a lot of other
> tcpip stack projects. What differentiates lwip from the rest? Why should a
> developer choose lwip over some other stack?

It's not clear to me that in all cases they should, and I can't speak for
anyone else, but the reason I chose to use lwip was the clean, modular
code that would allow me to adapt it to my needs, the simplicity of it,
the fact that it was being actively developed, it came with a BSD style
licence so it wouldn't exclude it from use in a commercial product.  In
short, it provided a good general purpose *base* for me to do my own
research.

> Firstly, I think we should not forget the main reason lwip was developed in
> the first place, to run on small embedded devices with a small amount of
> memory. Whenever adding code to lwip we must do our best to make sure that
> it is added in a way that allows developers working on small embedded
> systems to easily configure the stack to include or not include modules.
> (DHCP,FTP etc..) I think the HAL layer in lwip is excellent. This allows it
> to be ported easily. What is lacking is documentation. Many more developers
> will be attracted to lwip if there is more documentation.

I didn't find this a problem.  While more documentation wouldn't hurt,
there is good deal out there already.  The important things, such as the
API's and the sys_arch layer are covered in reasonable detail, and if you
want to understand the details of what's going on inside the stack then
firstly it's fairly standard TCP, and secondly the code is quite easy to
understand.  The mailing list is quite responsive too, but the ability to
search it would be useful when you first have a problem.

> Secondly, instead of developing Ethernet drivers for devices that other
> people have already developed drivers for, why don't we create a common goal
> of for example, developing a cheap wireless Ethernet driver/hardware that
> lwip will be able to run on. We will be one of the first to do it, and I'm
> sure you will all agree that this kind of functionality will give any
> products we are working on using lwip a competitive edge, and will also
> attract other developers to lwip.

While I'm sure a lot of people are using an Ethernet there are some who
aren't.  (I'm one of those!).  It would be reasonably easy to use lwip
over a wireless ethernet, but this is a fairly narrow application of it -
there are many other applications that people are working on, and this is
a good thing.  If you do decide to develop this I'm sure some would be
interested in helping, and many would be interested in the results, but
others are using lwip for quite different applications.

Good luck!

Kieran

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