Rob Lockhart wrote :
> Hi Jean
>
> The definitive Ultra spec is part of IrMC, not IrWW. The IrWW version is
> specific to Shimokura-san's watches.
Ok. I guess those are not really that different anyway...
> Speaking as one of the authors, though, please don't base a new
> implementation on Ultra unless it's specifically to talk to watches. Ultra
> was a solution to a problem that didn't really exist, we just didn't know
> it. At the time, we hadn't been exposed to IrDA Lite's ability to cut the
> main stack down to really small model device sized byte counts. (Ultra
> predates Dave Suvak's IrDA Lite by two meetings.) If we had, we'd've never
> written Ultra.
Speaking as the guy who pushed Ultra in the Linux-IrDA stack,
I beg to disagree. We are using intensively Ultra in our CoolTown
setup and we will continue to do so.
Our microcontroler is so dumb, slow and cheap (actually
cheaper than the Ir transceiver itself) that it would really not be
able to do IrDA lite, so for us Ultra is the only way to go.
Other added benefits :
o Ultra is *vastly* more robust than normal IrDA. We have
many, many cases when normal IrDA doesn't work because of cheap
tranceiver and fluorescent light, but we never found a single case
where Ultra would not work.
o Ultra is way faster. Normal IrDA has horrible latency for
connection setup.
o Ultra is much more power efficient. In fact, if we could do
Ultra at 115 kb/s, we could save even more power...
o Ultra is simpler to implement. Enough bugs in our code...
To conclude, we are really pleased with Ultra, and no way we
would go to IrDA Lite. In fact, for our applications, I can't see a
single benefit of using IrDA Lite over Ultra (I'm lying, the benefit
would be compatibility with various IrDA stacks that don't support
Ultra). In fact, we are planning to move to Obex over Ultra soon...
> Rob
> Chair - IrMC
Have fun...
Jean
_______________________________________________
Linux-IrDA mailing list - [EMAIL PROTECTED]
http://www4.pasta.cs.UiT.No/mailman/listinfo/linux-irda
