[Prism54-users] Re: [Prism54-devel] Prism54 development update

[Rainer Weikusat]

Feyd <feyd at seznam.cz> writes:
> On Wed, 14 Sep 2005 19:39:53 +0200 (CEST)
> Rainer Weikusat <rainer.weikusat at sncag.com> wrote:
>
>> In Germany (and at least in the US, too) you may own a radio that can
>> be tuned to arbitrary frequencies, but you are legally forbidden of
>> ever using it.
>
> Is then the limitation imposed by the driver sufficient? The device
> itself can still be tuned.

You seem to be very fond of always trying to ask the wrong
questions. See, I can (of course) take a conventional radio and modify
the circuitry to overcome any artificial barrier that it tries to
impose on me. But just because this is possible does not mean that
the people who try to impose this would be very font of a switch at
the frontside that everybody can easily use to overcome the
limitation.

>> > Unless you can offload expensive operations like encryption, the
>> > difference will be close to zero.
>> 
>> First, you are using the verb 'offload' in the wrong way --
>> non-crippled hardware usually implements its functionality while
>> crippled hardware offloads this to the host CPU (which is supposedly
>> owned by some idiot running Windows who doesn't deserve all those
>> cycles, anyway).
>
> Not true. The verb doesn't specify particular "from" and "where".

Certainly true. For your "reasoning" to make any sense, you must first
install the conception that hardware should do as little as possible
on its own as 'natural order of things' in the minds of your readers,
despite traditionally (thing of 'WinModems' versus 'real modems') it
used to be the other way round. Therefore, you talk about 'the host
CPU' 'offloading' certain tasks to devices, while I use the
conventional meaning of 'the device' offloading functionality to the
driver that used to be handled by the hardware itself.

>> Second, the only meaninful information the phrase
>> 'the number is close to zero' conveys is that this number is larger
>> than zero and that you can imagine another number that is large than
>> the first one, which you (supposedly) would not describe with 'close
>> to zero' anymore. So what precise numbers are we talking about?
>
> Again, not true. The meaningful information that you apparently missed
> is that the overhead of maintaining the 802.11 stack is negligible,
> tens of packets per second while data transfers generate thousands.

I don't care for two cents about what you consider something to be
that you are not going to name itself. I have a multi-CPU system and
want to use all CPUs in this multi-CPU system. If you don't want to
use all of them, that's fine with me. Apart from that: What numbers
are you talking about?

[BTW, please view this as a rethorical question. You are a BSD-bigot,
therefore, you try to talk everybody into using BSD-code (generating a
living for the people who have written those code as a side effect)].