By the way, regarding the patch you mentioned, it hasn't been implemented yet, has it?
No, it hasn't. (But, I plan to send theymos a summary of all the outstanding patches at some point. Hopefully that'll help. I think that a few of them are just waiting on theymos to come into enough free time for him to be able to really sit down and focus. Right now, theymos and I have a pretty stupid relationship... I'm often blocked on the thing that he has in abundance:
access, and he's often blocked on the thing that I have in abundance:
bandwidth. At some point, we'll figure out some compromise that makes better use of both of us.)
This could be fun:
It occurs to me that keeping a counter on what percentage of posters in the thread are ignoring a poster could help, but that's getting offtopic.
Yep, I noticed the "Implement stabbing" topic.
(I was working on a response to do with superconducting loops, quantum states, and just how unreliable a face-stabbing implementation in PHP would be, but, yeah... I ran out of energy after considering my wife's constant warnings about my deadpan sense of humor being very hit-or-miss with most people.) At some point I'm going to (try to) convince theymos to let me replace the whole ignore system, and when I get around to that, I'll think carefully about how to lay things out so that these kinds of suggestions can be looked into (that is, suggestions that depend on the availability of ignore-related statistics).
I've got a few ideas (and some guesses about why the previous approach didn't scale), but, I'm not going to put my thinking cap on just yet.
The current posting delay is almost as old as Bitcoin itself:
waittime = 360;
if(activity >= 15)
waittime = (int)(90 - activity);
if(activity >= 60)
waittime=(int)(34.7586 - (0.0793103 * activity));
if(activity >= 100)
waittime = max((int)(14-(activity/50)), 4);
Reporting counts as posting. I think deleting the red parts would solve it
(I had to replace code tags by quote tags to make it visible).
Hmm... I wonder if that pseudocode is still correct? (Only theymos would know, but, I've been bugging him a lot lately, so I'm not going to PM him for a while.)
I mean, one thing that occurs to me is to replace all of that with something simpler:
waittime = (int)min(988 / max(activity - 12, 1) + 1, 360);
Then, instead of it looking like this (red is when
activity is < 15, blue is when it's >= 15 and < 60, green is when it's >= 60 and < 100, and purple is when it's >= 100):
It would look like this:
(That way, the important features are matched, and, with that cast-to-int truncation, the wait-time would become 1 second once
activity is > 1000.)