Alright, I usually am able to play with pretty unstable but useable connection. EVERYTHING else (other games, web pages, streaming, etc.) work flawlessly. But why would SK work? The game disconnects out of nowhere, and you'll see everyone running off the map.
Must be my computer right? Nope, youtube works fine. My only guess is that the servers are pretty weak, and can't handle the traffic, but I have no idea. Any ideas on if this can be fixed, or is it a GH server problem?
What part of the world do you live in? If you live in Asia or Australia then you may have some connection problems because there are only two servers that run Spiral Knights, and they are close to Ireland (EU) or the US East Coast (US). If you're not close to either of those servers, then the game will be laggy or have issues with connection, however the servers that you get other information from (news outlets, youtube) do have servers near you, meaning they'll work just fine.
Ive had this problem for years now, and from talking to people it happens to them too. I am in US east coast, 42ms ping, and the game has 4 bars which mean nothing. I am not near any microwave ovens or other radiation emitting sources.
I have tried SK from browser, client, kongregate, steam, on mac and pc, at different locations with different Wi-Fi (college, home, etc.). I have no idea what it could be on my end, if anything.
Either the servers are designed for extremely low traffic, or something in the code is causing the disconnect. I'd rather play a game with 15 fps and 200ms ping with an actual stable connection than this game at 100fps and 42ms ping with random pointless disconnects.
The netcode is getting worse and worse as time goes on. Random spawnables causes the game to hang as all threads wait upon a server response, (examples: IMF boss begin, and opening monster boxes) as does loading the inventory from the server, server automatically logs out people with a variable latency, sprite ability usage is based on some two-fold netcode rubbish that stops the ability from working outright in a lot of common circumstances, and position correction can often be rapidly incrimental causing freaking obnoxious and inexplicably large and incorrect rubberbanding. (Video was an extreme case but highlights the errors in the system)
Interesting...how exactly does the netcode deteriorate over time? If GH resets the servers daily, wouldn't the data that causes the hangups be flushed? Or is this something that is due to hardware physically changing? Or, is it suboptimal design in one or more aspects of implementing their entire system? Or a combination...perhaps the servers are just getting old, either their software is deteriorating and/or the hardware is beginning to fail. I would imagine servers dealing with such high traffic would be difficult to maintain. I am not too informed about networking design though, but if daily resets do not solve the problem, then they definitely need replacements.
Maybe they just need new servers. Natural decay of machines is inevitable....how much do servers cost? I'd be down for an SK group funding page for new servers.
Thanks for the post, extremely informative.
Feature creep, I presume. As different things are added, sometimes functions have to be re-written to accomodate. And this isn't always done in an efficient manner.
Have you changed anything in your setup recently? Upgraded the operating system, drivers, Java VM, etc.?