Paul wrote :
>
> I work for a major CE company, and in a show environment we had a small
> enclosed room (with metal walls) into which we had 802.11 at 11MBps and
> streaming BlueTooth working simultaneously right next to each other. No
> problems, no missing packets (at least as far as we could see).
Of course, you won't have any missing packets. Why do you
think both 802.11 and BT include MAC level retransmissions ?
You need a better understanding of how the technology work to
really know how the two network impact each other, because both MAC
layer compensate for the effects of the Phy layer. You won't see any
packet losses, what you will see is both delay and reduced throughput.
Furthermore, my simulation was looking at the worse case, when
both link are fully loaded. You are probably aware that the current
limitations of the BT demonstrators prevent to fully use the
link. Fully loading 802.11 might also be tricky if you don't know how
to do it.
Now, to get worse for 802.11, you need both 802.11 nodes far
apart and the two BT in close range of one node. If the BT nodes get
away from the 802.11 node or the two 802.11 get closer, 802.11 can
capture packets and the throughput degradation fall to only a few
percent.
> However, I absolutely agree that there is a LOT of work to do on how all
> the UI stuff is handled. This is going to take a while. On the plus
> side, it seems to work and there are a LOT of companies working on it.
It's not because there is a lot of companies that thing will
go right (IrDA was having also a lot of companies, including MS, IBM
and HP, when both IBM and HP were still developping technology), but
that didn't help much.
And I claim that some of the way the MAC is designed (with so
many modes, high latency and exclusions) will prevent to offer a
consistent user experience : sometimes it will work as expected,
sometimes it will take time, and sometime it won't work because
whatever...
Example : if you have a voice channel at rate 1/3, you can't
perform "discovery" and you can't be discovered, and you can't open
any data connection. Boy, you can't do more stupid than that !
> Maybe we should focus on improving the UI for IrDA, so that by the time
> Bluetooth is perfected we can describe it as "Beaming without
> pointing".
I agree that we should improve IrDA user UI.
For beaming without pointing, you make me laught. Pointing is
an essential part of the IrDA user interface (at least, it should be
if things were designed properly). Let's say I've got two printers on
my desk. With IrDA, I select one by pointing my Palm at it. If I use
beaming without pointing, the printer next door will print my
document. Clever !
Ad-Hoc connection will remain with IrDA. BT is good only for
long term associations between devices (like MY phone and MY handset).
> After all, by the time Corel, RedHat, Debian, Mandrake, KDE, Gnome et al
> have finished a simple UI will allow normal human beings to install and
> use something as astoundingly complicated as Linux! :)
For installing Linux, the concept is simple, you just copy
file on the hard drive, but the implementation is complex. For
networking, the implementation is quite simple but the underlying
concept are complicated.
Jean
_______________________________________________
Linux-IrDA mailing list - [EMAIL PROTECTED]
http://www4.pasta.cs.UiT.No/mailman/listinfo/linux-irda
