Of course we often talk about a % CPU usage Every OS even lists it. what it means is what percentage of time the CPU is in use.

Seriously no one gives a crap if you say 5% CPU use or 100% of CPU use 5% of the time. Its just a useless distinction made by someone trying to look smart.

But ask people trying to maintain a ratio (because its required or they are just good sharers) how well 3.3 is going for them.

We get regular complaints about never getting connected to seeds using 3.3 and switching back to 3.2.3 and suddenly they share just fine again. Guess what the problem might by. The only change was client version.

But it must be a user setting im sure……

The other common complaint is with file allocation where disk use hits 100% and the client locks up.

But heck keep telling yourself client doesnt matter.

It’s nice to see wireshark being discussed a bit more, outside of a purely technical context. A thought experiment we suggest to activists concerned with their online security is this: imagine that someone’s sitting just ‘outside’ of your physical network adapter, running tcpdump/wireshark and capturing every packet leaving your machine. What will that logfile look like? What information is available in it, were a determined adversary to sit at a console and do some serious work at understanding what it contains?

That’s an excellent ‘Gedankenexperiment’ despite the fact that, definitionally, it’s an unlikely attack vector in purely practical terms. That is, if one can say with confidence that even an attacker sitting on the other side of one’s network adapter wouldn’t get anything compromising from that packet traffic, one has in place some reasonably competent network security at a baseline level. Conversely, if one can’t “visualize” the concepts in question, it’s perhaps a good indicator that more work understanding these concepts would be constructive.

There’s plenty of tools that will/would subvert such an attack scenario, and do so elegantly. First and foremost, a competent, serious VPN service will result in that packet stream leaving from (and arriving to) the local machine is shorn both of useful payload data _and_ useful header/metadata. That’s a rather impressive traffic profile to present to the world.

(note of course that if someone could run wireshark on one’s local machine, then such an attacker would definitionally have escalated local OS access and the packet stream being encrypted or not would be irrelevant given that access: that’s why it’s a thought experiment, and not really an attack scenario one would expect to see in the wild)

We do notice folks sometimes assuming magical powers when it comes to wireshark. Alas, it can’t capture packets unless it’s in the right topological position on the network. That’s usually the tricky part…

No computer program does that without using 100% of your CPU. You apparently don’t know anything about low level language functions either.

Your computer and your voice are the exact same. If you’d stop running porn.avi.exe and stop shouting, you’d “disable” the unwanted behaviors.

The adverts and statistics are a side effect of, not the intenet of, the software. i.e. it isn’t adware or spyware.

nmap wouldn’t help you do anything other than see how many hosts are on the local network and what ports they’re advertising. Good luck getting scanning all ports above 1024, or even ports in the 6000-7000 range for 254 hosts just to see who is using a BitTorrent client- maybe- possibly- potentially- I guess.

With settings appropriate for the linespeed, client really doesn’t matter.

You might want to remember something – this is what we do. As many times as we’ve heard the ‘client matters’, do you THINK we haven’t checked it out? Set up test swarms, and checked things for ourselves?
Or do you think we just accept the word of random comments left on youtube and activity-monitoring (aka ‘private’) trackers by people who have no idea what they’re talking about?

Think on that, eh?