30 Apr
2012
30 Apr
'12
1:18 p.m.
On 04/30/2012 04:06 PM, Richard wrote:
Or perhaps "half-duplex", although USB 3.0 claims full-duplex
capability.
None of this is making sense to me.
Please clarify what you mean by these terms in this context:
"polled nature"
"asymmetric"
"half-duplex"
...and it would help that instead of just rephrasing in another
euphamism, you would say exactly what it is about USB that matters
here.
Richard picks a fight, once again.
He was, I'm pretty certain, referring to the fact that USB is a fairly
complex (at least in comparison to RS232) master/slave network
interface. You cannot (usually) plug a USB peripheral into a PeeCee and
have the PeeCee become a slave peripheral.
Dave's original assertion was poorly worded, but it was absolutely
correct. If you really want to nitpick the wording, rather than the
(friggin' OBVIOUS!) intended meaning, well...you've got more free time
than I do. Or my cats, for that matter.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
>> [...]
because USB is
>> one-way
>
> Since when?
I think David is referring to its polled nature.
Yes, I suppose "asymmetric" would have been a better term. It's
not peer-to-peer like serial ports or 1394 are, which is what would
make it nearly useless for kernel debug; that's before you get to
issues of protocol stack, which 1394 gets away with by using DMA on
the controller, which is questionable from a security standpoint but
great for kernel debug.