Our test group flight on Saturday went well with one exception. The link I provided to the voice program we use, Discord, was evidently broken. I have scheduled a second one for this Saturday the 22nd. We almost got to 20 with just our members. Perhaps if some others came by we could get to 40.
For more information:
http://msflights.net/forum/calendar.php ... nfo&e=1734
(Sorry I know it's late for the European folks)
JoinFS multiplayer event 22 October 2016, 2100 EDT
Re: JoinFS multiplayer event 22 October 2016, 2100 EDT
How did you guys get on with yesterday's mass multiplayer, any problems at all?
I would have joined but was fast asleep at the time.
Peter
I would have joined but was fast asleep at the time.
Peter
Re: JoinFS multiplayer event 22 October 2016, 2100 EDT
Peter,
Not getting email notifications of topic replies.
It went okay. We had 11 pilots/users.
-Mike had issues with the area wifi that he has. He had issues seeing others and visa versa. That resolved when Curz suggested that he tried to connect to one of the other users in the session via their IP. That suggestion worked.
-I think we saw last week that when any one user (node) got more than 5 or 6 planes some problems crept in. We used 3 people to connect to the creator this time and had everyone else connect to them.
-We spawned in South Africa (FADM). I used that because I wanted a place where elevation issues were the least likely to be an issue. (who has add on scenery for South Africa right?). I also had everyone turn Vector off. I don't think we had anyone have issues because of a different elevation. I think elevation correction is something the other clients do to work around these things. It solves some problems, causes others, right?
-Once we were all flying things went very well. Very smooth performance. An occasional rubberbanding movement was seen. One user reported a jittery appearence of other aircraft when they were really close. I was very close and saw no jittery movements from that aircraft.
- I did have an issue when we changed from default Barons into a free for all choice that 2 users showed up on the list at the 0,0 point in the Atlantic ocean. I tried changing their model, my model, disconnecting, shutting down and restarting, etc... Both were connected to different nodes/pilots and neither was connected to the same node/pilot that I was.
-I figured out that in testing terms for describing what was being seen where was an issue. I tried to define "pilots" as the number of connected "Pilots" seen in the main program windows. "Aircraft" as the number of items in the Aircraft list window. I called the items in the network list as "GOOEY-D's". That term didn't really catch on. And finally when trying to defined how many we saw actually in the sim, we called those "labels' I think.
I think to the process is going well overall. Naturally I hear it should do this, and it didn't do that. I don't think there was anyone who felt frustrated. I think everyone is excited of what this final product can bring to our hobby.
If you need a venue to orchestrate some testing we do have many people around during the day that have been fiddling with it during the week. Most of our group is in the US and until the weather gets cold, daytime is not the best time. But you are absolutely welcome to schedule an event on our facilities/calendar. I imagine you could coordinate with someone to take the lead thus freeing you up to focus on the aspects that you want to focus on.
Not getting email notifications of topic replies.
It went okay. We had 11 pilots/users.
-Mike had issues with the area wifi that he has. He had issues seeing others and visa versa. That resolved when Curz suggested that he tried to connect to one of the other users in the session via their IP. That suggestion worked.
-I think we saw last week that when any one user (node) got more than 5 or 6 planes some problems crept in. We used 3 people to connect to the creator this time and had everyone else connect to them.
-We spawned in South Africa (FADM). I used that because I wanted a place where elevation issues were the least likely to be an issue. (who has add on scenery for South Africa right?). I also had everyone turn Vector off. I don't think we had anyone have issues because of a different elevation. I think elevation correction is something the other clients do to work around these things. It solves some problems, causes others, right?
-Once we were all flying things went very well. Very smooth performance. An occasional rubberbanding movement was seen. One user reported a jittery appearence of other aircraft when they were really close. I was very close and saw no jittery movements from that aircraft.
- I did have an issue when we changed from default Barons into a free for all choice that 2 users showed up on the list at the 0,0 point in the Atlantic ocean. I tried changing their model, my model, disconnecting, shutting down and restarting, etc... Both were connected to different nodes/pilots and neither was connected to the same node/pilot that I was.
-I figured out that in testing terms for describing what was being seen where was an issue. I tried to define "pilots" as the number of connected "Pilots" seen in the main program windows. "Aircraft" as the number of items in the Aircraft list window. I called the items in the network list as "GOOEY-D's". That term didn't really catch on. And finally when trying to defined how many we saw actually in the sim, we called those "labels' I think.
I think to the process is going well overall. Naturally I hear it should do this, and it didn't do that. I don't think there was anyone who felt frustrated. I think everyone is excited of what this final product can bring to our hobby.
If you need a venue to orchestrate some testing we do have many people around during the day that have been fiddling with it during the week. Most of our group is in the US and until the weather gets cold, daytime is not the best time. But you are absolutely welcome to schedule an event on our facilities/calendar. I imagine you could coordinate with someone to take the lead thus freeing you up to focus on the aspects that you want to focus on.
Re: JoinFS multiplayer event 22 October 2016, 2100 EDT
Hi Karl,
I appreciate the feedback, as always.
Not sure about email notifications, I have the subscribe option set on the forum settings. I'm a bit new to forum administration so bear with me
I think different scenery can still cause problems. I've had this a lot when using FSX multiplayer. It doesn't help that the aircraft get shifted to (0, 0) if they happen to go underground on your scenery in JoinFS. Perhaps the best thing is to get in the air before checking on who is where.
I've seen the strange shifting back and forth. It only happens to a few people and is consistent. I think one possibility is the system timings can be inaccurate on certain machines. I've increased the system timer resolution for 1.0.7. That may fix it. There's something else that I can try as well for a future version.
I will come and join some of your sessions this week. I think it will help getting fast feedback and understanding of some of these issues.
And, I think I've just fixed some network stability issues that crept into 1.0.5/6, perhaps they didn't affect you much yesterday, but from looking at the code it is likely it has contributed to some of the problems you mentioned.
1.0.7 is out tomorrow with some good updates: fixes, more aircraft info, loop twist fix and pilot nicknames.
Best regards,
Peter
I appreciate the feedback, as always.
Not sure about email notifications, I have the subscribe option set on the forum settings. I'm a bit new to forum administration so bear with me
I think different scenery can still cause problems. I've had this a lot when using FSX multiplayer. It doesn't help that the aircraft get shifted to (0, 0) if they happen to go underground on your scenery in JoinFS. Perhaps the best thing is to get in the air before checking on who is where.
I've seen the strange shifting back and forth. It only happens to a few people and is consistent. I think one possibility is the system timings can be inaccurate on certain machines. I've increased the system timer resolution for 1.0.7. That may fix it. There's something else that I can try as well for a future version.
I will come and join some of your sessions this week. I think it will help getting fast feedback and understanding of some of these issues.
And, I think I've just fixed some network stability issues that crept into 1.0.5/6, perhaps they didn't affect you much yesterday, but from looking at the code it is likely it has contributed to some of the problems you mentioned.
1.0.7 is out tomorrow with some good updates: fixes, more aircraft info, loop twist fix and pilot nicknames.
Best regards,
Peter
Re: JoinFS multiplayer event 22 October 2016, 2100 EDT
Sorry I missed it but real life got in the way here.
On the rubber banding and connection issues, is RTT the same as ping in the network window? Maybe setting background color to red or green for those dropping packets or with long ping times? While there isn't much cure for bad network issues between pilots, it could highlight if that is the issue or if users should look for other explanations?
On the rubber banding and connection issues, is RTT the same as ping in the network window? Maybe setting background color to red or green for those dropping packets or with long ping times? While there isn't much cure for bad network issues between pilots, it could highlight if that is the issue or if users should look for other explanations?
Re: JoinFS multiplayer event 22 October 2016, 2100 EDT
Hi Robert,
Yes, RTT (round trip time) is effectively ping. I've renamed this to Latency in version 1.0.7. It is a different measurement to normal network pings, so don't expect really low values like 10 ms. As well as time spent in transmission it also includes time being processed by JoinFS.
As you suggest I could add some extra metrics to highlight network inconsistencies. That would help in identifying unavoidable connection problems.
Thanks,
Peter
Yes, RTT (round trip time) is effectively ping. I've renamed this to Latency in version 1.0.7. It is a different measurement to normal network pings, so don't expect really low values like 10 ms. As well as time spent in transmission it also includes time being processed by JoinFS.
As you suggest I could add some extra metrics to highlight network inconsistencies. That would help in identifying unavoidable connection problems.
Thanks,
Peter