There's an idea: offer your pool over Tor
Would be a nice alternative to using Tor exit nodes and would reduce latency (bigger choice of intermediate nodes) ^^

Can you explain more?  I realise that an attack is very unlikely, but I'm curious.  Do you keep a record of which work has been assigned to which worker?  Everything else is the same - the data, the target etc.  Only the work (nonce? is that what it's called?) changes, so it would be enough for an evil tor exit node to connect to the pool with it's own worker credentials and submit my successful nonce.

Unfortunately, there is a firewall here, so I have to use tor to connect, unless you can open a port<1024 and forward to 8332 on your machine.  So far latencies don't seem to be causing any problem.  I reduced the ask rate to every 10sec, so it pauses less often.  I'd experiment to find the best values, but I only ever expect to earn about 80BTC through generation, so I'm not very motivated.

I'm using m0mchil's python miner, and everytime it submits a getwork request, it stops hashing, opens the connection, gets the work, and then starts again.  This causes me to lose much more that 2.5% of my hashing power, more like 15% or so (note, that's hashing power, not hashes).  Like I say, if I were worried I'd modify the miner, but between now and when the sun goes out, I can only expect another 70 or 80 bitcoins with my meager 20Mhps if the bitcoin network keeps growing as it is for a while longer.  So it'll do as it is.  Thanks for the reply.

Looks weird. How is you network going? Any packet losses, timeouts?

It's calculated from all found blocks in this day, so total reward. And yes, yesterday was not good, but it is how it works. Two days before were fine, for example.

No problem with restart, but maybe will be good to receive an email with at least 6 hrs in advance.

Yes, I did mistake in network configuration, so there was small outage, too. But it was not full restart with all stuff bootstrapping.

Tell me your phone, I'll call you!

This reboot was unplanned outage, so sorry, I cannot tell you 6 hours before in this case.

I have email notification about crashed miners almost done for many weeks. Unfortunately bad guys are inventing more and more bad things, which makes me busy on fixing it

Well, I believe that you may count them on your side, too.
But I also think that stats were useful, I think Slush needed to take them out due to pooling cheating.