"Don´t tell me that it is not possible in such or a similar manner, because in the First Person Shooter genre that is commonplace, see a server in the list, download the map if you don´t have it, join." Exactly, I think CStrike does this since 10 years?... however this has some flaws as well: 1. we used to run a CSS server with custom maps (gunmode) and even though the maps were small and the server not very much known we had gigs and gigs of auto-downloads... this exploded our priceplan and we had to disable it... > we were only a few friends and did not want to invest a lot. > even today high quality bandwidth has a cost... and auto-download is kind of a risk for the guy running the server. 2. the server running the session with autodownload must have a really good bandwidth... > "I understood that he meant that there would be some way to give files without crashing server..." the server might not be crashing but the ongoing race might be disturbed... one sec in the wrong corner is enough to ruin the whole race... Even today it is very difficult to properly "throttle" or shape traffic. If you do it, you should do it on the firewall or router. > hence the "server auto-downloading function" should at least use a dedicated port, this would make it much easier to put some QoS management on the traffic. 3. The server autodownload function could be an issue for people hosting races on their home line. I often just do a little race with 2 or 3 friends on my own dsl line... my upload is however limited to 800kb.... so downloading from there will take a while. But to be clear: any proposed solution in this thread is better than what we have now... For me it just seems the most user friendly thing to add a download function to the mod manager and integrate this with "GET MOD & content" function in game... So do we all agree that we would like to have better management of content and some kind of autodownload function? IF SO, shout out loud
Cracheur you are fixating too much on problems that aren´t any. The built-in HHTP server can be disabled in its current state already, if you have it enabled, you can limit the maximum file size. People, that don´t have the means to host the downloadable content, be it lack of bandwidth or traffic volume, will have those problems either way, so they should simply disable that functionality. Or as I suggested in the post I linked to above, redirection to mirrors could be implemented. The built-in HTTP server uses its own port already. And I have to disagree on traffic shaping being done on the router or firewall. Who best to prioritize the traffic than the application generating it. We set the bandwidth in the Multiplayer.ini. Any traffic not directly related to running the race should be queued with the least priority before even sending it to the network interface. The critical traffic should simply bypass that queue. You cannot expect the average rFactor 2 user/admin to do such things on their own. Anyway, all your points look pretty much like they are really one and the same, concerns about people lacking bandwidth and or traffic allowance. That is a problem you cannot change on the application level anyway. Using redirections to mirrors or to the bittorrent network are workarounds, that could be implemented. Cheers, Marcus
Yes, redirection to mirrors or torrent managed downloads would be a good solution for bandwidth abuse, but in my opinion, the greatest problem today is people making vmods wich include 10 or 15 tracks, often including one or two out of date. It's so simple to create a vmod, so, why not just create one when you wish to change track or car (or both) on your server? Even more, if every server admin would learn to make 5 Kb vmods, racing life would be much easier byyyyyyyyyye
And what are the crappiest tracks you would know that be a great way to promote confidence for young modders putting a persons tracks in the crap list. clap clap clap .......................... How about blaming the right person. If GET MODS does not work..........BLAME the MAKER ! ffs lmao
ISI failed to come up with a good design for this. And since the software is now 'not beta', you can be sure they won't do anything significant as regards changes to the structure/design they currently have. The right approach for 'Get Mod' would have been for ISI to come up with a protocol for finding the name/address of a server to contact to download the Mod files, and a mechanism to transfer. This need not be the server hosting the race. For example, the server hosting would give you the vmod which basically says, you need component A which you can download from www.modA.com/modA and component B which you can download from www.modB.com and so forth. Thus, no need for ISI to host all the mods (for which it would get nailed for licensing violations). They could also provide and manage a whitelist or blacklist if license holders find servers with illegal content, that they want the matchmaker to block. No need for ISI to host the actual race server or mod file server, no bandwidth needed for these 300Mb components etc. All they do is run the matchmaker. The mod files would contain everything needed (single file, or one file with the content and one file with install config) and the means to self-install. This would require standardization of the game paths/folders (or parameter based, where things should go). But, this still requires the mod makers to find someone to host every single version of a component that they ever publish. And it still doesn't allow someone to create a fix/patch to one of these versioned components (as trivial as their own logos for billboards for example, or a fix to a pit position so the track will load/be raceable, where the original mod maker never issues such a fix). At present, this requires you to unpack the MAS file, include the fixed/updated file, and repack into a new MAS file. But now you've created a fork from the original modders work, and in fact, you've no idea what that MAS file really contains or changed. No system will be perfect to solve these problems.
You are missing the point. The very reason we have this problem are vMods themselves, and their definition, which requires at least one vehicle and track. That is compounded by the fact, that rFactor 2 requires the client to have the whole mod installed not just what´s relevant for the current session. I certainly don´t want to go login to the root server and create yet another vMod for such simple thing as changing the car and track selections. But I am repeating myself. Cheers, Marcus
calm down man! In my opinion the crappy tracks are the badly made conversions from RF1 tracks. stuff like 70's glen and such. So yeah, i want a list of badly made conversions to discourage people from making them. They don't even bother to at least make the grass textures a bit darker. I don't have anything against new tracks from new people - flooding the community with badly made conversions is a completely different thing.
I am with Durge on this one, I´m afraid. A platform like rFactor 2 should be all about diversity and customizability. The reason people consider making such lists is that the current mod delivery system is in contradiction to such goals.
I am just expressing my personal taste. I am judging no one. Lol, this will just end in a flame war with you guys. I will just keep out of this without defending me any further or expressing my opinion any more.
and again a polite thread is being turned by some rF2-TALIBANS to a war. let's ignore it... the issue we're discussing needs to be adressed properly. You concentrate on: "people lacking bandwidth and or traffic allowance" YES, that's right because it's an issue for most users. For me it's one of the most important factors to consider when looking for a "content distribution solution". Regarding traffic shaping, well Yes there are many discussions on how to do it proporely in general but there is no 100% way. As exemple do the following: start a download client bittorrent, usenext, edonkey or whatever, limit the bandwidth in the software, start a download. GO online and play, for instance, a shooter: check your latency... > your ping will suffer anyway heavily anyway from the background download... "The built-in HHTP server can be disabled" > and that's what will happen for the majority of servers. > the solution will not work for many people > not really a solution in my eyes. "Using redirections to mirrors or to the bittorrent network are workarounds, that could be implemented." > why not going directly this way? I agree it would be better to have direct download from servers but you agree that we might need another "backup" solution to serve all users. (again the one not owing a nice, unlimited high bandwidth server...) "No system will be perfect to solve these problems." that's what I mean but the proposed solution should, in best case, be a solution for the majority of users. Sorry to insist but I have kind of a strong opinion on the "download content from server running game" solution. Check this thread: http://isiforums.net/f/showthread.php/1183-Minimum-Recommended-Server-Specs-for-rFactor-2/page2 Post from Tim: “Ok, finally clear in my head... 768Kbps *30 = 172.80 MB/min = 2.88 MB/s = 23.04 Mbps = 2.75 MiB/s = 10.37 GB/h My home connection is 0.92 Mbps upload, to put that in perspective. So I won't be doing much hosting. But I am on the cheapest package my ISP offers right now. ” And that's without any downloads...
And the reason for your ping suffering is, that there is no QoS solution or traffic shaping in place. Also, filesharing network clients tend to use a lot of connections, so even though you have limited the bandwidth, you might just be giving your home router a hard time managing all those connections. NAT and packet filtering for very many connections requires computing power as well, which home routers are in constant lack of, if you try to use them hard. No matter which solution is the best, should there actually be an answer to that, unless rFactor 2 provides it with their program and make it usable by an average user, QoS and/or traffic shaping won´t happen on a large scale. I think we are on the same page, that direct download from the server alone with no option for redirecting to other download sources, is a soultion that does not work for all users, I am just not sure if it really is the majority of them. But here is where discipline of those admins can go a long way. If you stick to popular content, which often is of high quality too, the probability that someone wanting to join your server doesn´t have it already will decrease with increasing popularity. Many will already have such content, either downloaded whilst joining another server or from the original source, because it will have gotten some praise on the simracing media. And this is how it should be decided which content is good and which isn´t, by popularity not by some commitee as some seem to propose. That would be democracy at its purest. Which birngs me back to the OP. With the system as it is now, we know, one is required to have everything that is referred to by the (v)Mod selected on the server, no matter if it is even selected in the server´s cars and tracks filters, let alone which track is currently running! So if one happens to have unpopular or exotic content in such (v)Mod, that automatically renders the server itself unpopular and/or exotic as well. Exaggeratedly speaking, nobody will want to download a backyard dirt track and the accompanying tractor just to join the currently running GT race at Portugal circuit. But that´s what the current system requires, which is fundamentally wrong and should be fixed. Everything else is just icing on the cake. Cheers, Marcus
we're getting closer... a combination between auto download, definition of musthave content and some minor change on the modding system....hmmm i have doubts isi would go there.... some communities, like bockbierbude, offer update/sync clients.... maybe we will have to rely on an addon....or another sim for casual racing....
