They did a clean room reverse engineering of Windows and it's still legal, it works, and it can run quite a good lot of Windoze applications! I wouldn't be surprised if it could run Freelancer. And that's Microsoft's god-damned operating system they copied for Pete's sake! If we did that same thing with FLServer we'd have the ultimate, stable, bugfree, free implementation of the software and not have to pay Microsoft! (that I know of.)
They did a clean room reverse engineering of Windows and it's still legal, it works, and it can run quite a good lot of Windoze applications! I wouldn't be surprised if it could run Freelancer. And that's Microsoft's god-damned operating system they copied for Pete's sake! If we did that same thing with FLServer we'd have the ultimate, stable, bugfree, free implementation of the software and not have to pay Microsoft! (that I know of.)
I know about ReactOS and its predecessor. I also know of the time used in developing ReactOS.
If we only look at ReactOS from the time they began targetting NT based versions of Windows, and not from the inception of the project (FreeWin95) their kernel rewrite started in Q1 2007, and hasn't finished yet. They haven't even denominated a version as a beta test, all their available code revisions so far are in an alpha state.
The only reason ReactOS have a userland part (The GUI and means to interact with the kernel) is because of the Wine project, itself running for more than 16 years now. And none of those projects has in that long run managed to cleanroom reverse engineer their intended target parts of the Microsoft Windows OS to a dot.
FLserver is ofcourse not as complicated to code, as a complete OS, but still, a complete cleanroom reverse engineering of FLserver will take lots of man-power and man-years.
IF such a project got started, a result would not be ready for the first couple of years or even more. Remember, FL took several man-years to develop. A cleanroom reverse engineering will also take several man-years to make.
I just don't see neither the Discovery Mod community, nor the complete Freelancer community have the required critical mass to have talented codeanalysts and programmers enough to do such a sissyfosean task within this decade, nor the next.
IMHO a reverse engineered fixed FLserver is a nice dream to have - But thats all it is : A dream.
But again as I said in my previous post - Feel free to prove me wrong with code - But for the love of Sirius, stop using the "we just need to rewrite FLserver" (while looking in Cannon's direction ??!?) as a solution to any and all limitations the current flserver has.
Cpt. Miller, of the BHG|Core vessel "Miller's Draft"
Out of bats, Out of bots, Out of torps - Down to harsh language...
I was thinking about it and there may be a way to use FLhook and allow more limit on the same server. Though on the back end it would be multiple machines.
This is how we do it.
we split sirrius into multiple systems..... (6 of em I guess one for each house and one for the Outcasts and Corsairs) . Eash system with a limit of 200 people.
once you jump from one system to another. flhook uses the info to disconnect you from one system and connect you to a corresponding system (logging off of one server and log onto another). If that server is at a limit or its down, system can say "jump failed" or something like that. with sirrius divided into multiple secotrs, server load would go down on individual machines and we can potentially increase the size to 1200 people.
' Wrote:I was thinking about it and there may be a way to use FLhook and allow more limit on the same server. Though on the back end it would be multiple machines.
This is how we do it.
we split sirrius into multiple systems..... (6 of em I guess one for each house and one for the Outcasts and Corsairs) . Eash system with a limit of 200 people.
once you jump from one system to another. flhook uses the info to disconnect you from one system and connect you to a corresponding system (logging off of one server and log onto another). If that server is at a limit or its down, system can say "jump failed" or something like that. with sirrius divided into multiple secotrs, server load would go down on individual machines and we can potentially increase the size to 1200 people.
Yeeehaaawwww.....
What do you guys think ?
*cough* Uh... that's Majkp's second proposal in the first post of the thread, 'cept 3 times more expensive, and unnecessarily so. THis does seem to me the most attactive idea, though. Admittedly, it's also somewhat more difficult to implement.
Second option is more realistic although, as we've all read, harder and more expensive to implement. The fact that everyone's aware of is that number of players going higher each and every day. But to split it in 2 parallel universe? In that case, it's better to do nothing.
Given that each FLserver can only use one core, I doubt it would be significantly more expensive to split into two, four or even eight seperate threads - I doubt very much more that significantly more expensive iron would need to be hired to accommodate it.
However, this is absolute pie in the sky. The idea in the original post about pseudo-cloud computing would be more along the lines of having two distinctly separate servers sharing a single database (or having frequent sync between two) and having users assigned to server 1 or server 2 when they click to join the server - Thoughts about splitting regions into separate servers are pretty much asking for a complete rewrite of the flserver.
What we have here is a gap between the attainable and the ideal.
1. is attainable
2. is ideal
"Man's reach must exceed his grasp, or else what's a dream for?"
this is six months since peeps were saying 'leave it and see if it calms down' but it hasn't.
1. we already have. there is rp usa, rp uk, etc etc. we can choose to come to other servers if this one is full. if the servers were magically linked so that the chars were shared, there'd be a lot of other servers popping up.
2. would be sweet. could we implement a version of the /beam command to move from server to server.
would reducing the no of npc's reduce lag, and thereby enable us to increase the no of players without frying the processor?
a 50% reduction in the no of npc's/asteroid rocks spawned would make life a lot easier
maybe only have npc's spawning in a certain zone around bases (like asteroid spawn only in certain areas)
in the short term, i recommend that there is a limit on no of players that can be in any one system (20?). so if your a lolwut capzor and you last docked in NY, you have to play other chars instead.
simples
' Wrote:maybe only have npc's spawning in a certain zone around bases (like asteroid spawn only in certain areas)
in the short term, i recommend that there is a limit on no of players that can be in any one system (20?). so if your a lolwut capzor and you last docked in NY, you have to play other chars instead.
simples
What if I am not a lolwut capzor and last docked in NY? Because my character's ZoI lies in a high trafficed area I don't ge to play? Sure, that sounds perfectly fair.
As for reducing the NPCs/other enviromental stuff to reduce lag and thus be able to increase the number of players.
While I would love the server to be able to allow more than 200 people on at once, I don't want to sacrifice the game environment just to have more people online. Whats the point of playing with a ton of people when the environment is empty?
To be honest, the only idea I have liked is to install a server password. One that is only obtainable by a member of this forum. This would free us of a lot of the lolwuts that don't know the rules and prevent those that don't care for our rules and such from logging on and taking up a slot when an actual Disco player is trying to get on.
' Wrote:To be honest, the only idea I have liked is to install a server password. One that is only obtainable by a member of this forum. This would free us of a lot of the lolwuts that don't know the rules and prevent those that don't care for our rules and such from logging on and taking up a slot when an actual Disco player is trying to get on.
This sounds pretty reasonable of an idea to me. In the server info viewable at server select, simply explain that there is an application process to access the server which involves registering for the site. You could easily implement an application form for the site that walks new players through basic rules, purpose of the server, point them to various forums such as the story and bio's section to get a feel for what goes on and to start brewing their own stories.
Then ask a few questions perhaps, gather a demographic of the player population. Keep it RP-relevant. Ask them what past RP experiences they've had, why they're interested in joining our community. It doesn't have to sound like an interrogation, however I have used this style of application in guild recruitment for other games in the past and it serves as an excellent filter for those who are serious and interested in the RP, and those who are not. An english test may also be applicable here. Those who blatantly flag concern can be denied at this point with no further hassle of warnings, sanctions and bans due to people with no intent of playing with the rules in mind.
Once the form is submitted it would require admin review. Upon acceptance, the password is sent to the user's new forum account inbox, and forum posting access is granted to them.
The downsides I could see to this approach are: - Admin manpower. As I understand admins are already swamped with reports, valid and inane. I am not sure if the staff has or would consider increasing the number of admins to accomodate. I would also counter that this sort of filtering would drastically cut down on the number of reports created by players uninterested in following rules to begin with. - Disinterest. Some players are not interested in applying for a job to play a game. It could turn some prospective members away knowing that they need approval in order to try it out. Some could see this as a positive note and in a way I would not disagree. Those with the drive and intent to bring unique character content to the community and manifest it in-game will be those who stick it out for approval, and may very well understand the process. Most RPers come from previous RP communities where similar issues are recognized. - Keeping the password hidden, but regularly updating current members. Using a password does no good if the password is promptly displayed or spoken about. One of two combinable approaches to this would be to create a sub-forum to the accouncements forum, where weekly server passwords are posted by the admins. On a pre-determined day, perhaps whenever the server is given it's scheduled restart, the password is changed.
Secondly, the password could be sent to all member inboxes, or another option could be setup to have it emailed to all users, to the email registered to their forum account. Players would know on Such-and-Such day every week, to check for their new password to play. This can also be used as a hook to get players on the website more often and participating.
Of this issue, the suggestion of population control seems to be the most reasonable in terms of ability and time. I know I have been in the situation several times thus far in my last three weeks or so of playing where I cannot log in due to a full server. When I do log in, the servers are swamped with lag. Cut down on the people that you don't want on the server to begin with, and it will allow for more time in coming up with a permanent solution.