LocoRon wrote:If Facebook, Google, and others can successfully track an individual across the web (without relying on an IP address), then I am confident that better methods can be devised for KGS as well.
They track you according to your username. If you go to a new computer they don't magically know its you. You still need to login, at least once, WITH your username. So, if abusive individuals just hop between KGS accounts, I don't see what KGS can do to prevent this short of the IP block.
It's not just your username. Basically, they store a cookie on your computer; other websites then read the cookie and send the data back to... whatever company.
Anyway, the basic technology (cookies) can easily be adapted.
A tech-savvy user can get around these (by simply finding and deleting them); a tech-savvy troublemaker can also get around IP bans easily enough. The IP bans are only good for two things: Punishing tech-ignorant troublemakers, and collateral damage.
LocoRon wrote:A tech-savvy user can get around these (by simply finding and deleting them); a tech-savvy troublemaker can also get around IP bans easily enough. The IP bans are only good for two things: Punishing tech-ignorant troublemakers, and collateral damage.
A dedicated troublemaker will be able to get around any system you put in place short of destroying the server, so they are probably a poor choice to use as reference. Blocking an IP stops the vast majority of troublemakers and in reality (barring the occasional special circumstance like a major go gathering) only has the occasional impact on non-troublemakers. Even in your scenario where you have a neighborhood where the IP addresses rotate every 24 hours, then it would affect you only if you have a neighbor who plays go, they get banned for a serious offense, you get their old IP address the next day, and you try to login to KGS that day. If all of these things collide, the worst case scenario is you cannot access KGS for a day (unless of course you send an email to the admin email explaining what happened and someone restores your access).
Cookies and IPs aren't created equal. I'm reasonably tech savvy, but while I use a lot of private browsing to avoid tracking*, I wouldn't generally take the trouble to avoid having my IP address tracked (of course, it's been years since I was booted from KGS, so my motivation is also different).
* As a Java program, couldn't KGS do more than just use cookies, however? Obviously just about anything would be avoidable by someone determined enough, but couldn't wms make it more of a pain, if he so chose?
LocoRon wrote:A tech-savvy user can get around these (by simply finding and deleting them); a tech-savvy troublemaker can also get around IP bans easily enough. The IP bans are only good for two things: Punishing tech-ignorant troublemakers, and collateral damage.
A dedicated troublemaker will be able to get around any system you put in place short of destroying the server, so they are probably a poor choice to use as reference. Blocking an IP stops the vast majority of troublemakers and in reality (barring the occasional special circumstance like a major go gathering) only has the occasional impact on non-troublemakers. Even in your scenario where you have a neighborhood where the IP addresses rotate every 24 hours, then it would affect you only if you have a neighbor who plays go, they get banned for a serious offense, you get their old IP address the next day, and you try to login to KGS that day. If all of these things collide, the worst case scenario is you cannot access KGS for a day (unless of course you send an email to the admin email explaining what happened and someone restores your access).
Would you not agree, though, that a system that does not block any "innocent" users is better than a system that does, occasionally, result in blocking random users who were unaffiliated with the offense which provoked that ban?
That's literally the only thing I've been attempting to convey; there are *better* ways of banning an individual user than IP bans.
Can you find a better system that does not block innocent users, and most importantly, does not impose more strict registration information gathering for all users?
What I think Mef was saying (and I agree) is that the current system works for the most part, and the effort to make a system that distinguishes better between innocent users and abusers would be impractical.
LocoRon wrote:Would you not agree, though, that a system that does not block any "innocent" users is better than a system that does, occasionally, result in blocking random users who were unaffiliated with the offense which provoked that ban?
That's literally the only thing I've been attempting to convey; there are *better* ways of banning an individual user than IP bans.
I would agree so long as you include the very dangerous phrase "All other things being equal"
