Forged Alliance Forever Forged Alliance Forever Forums 2012-02-24T21:33:51+02:00 /feed.php?f=2&t=696 2012-02-24T21:33:51+02:00 2012-02-24T21:33:51+02:00 /viewtopic.php?t=696&p=7152#p7152 <![CDATA[Re: Broader Mod Support Coming?]]>
However, we could weed out broken mods by GUID screening them in our "vault". Still, someone would need to code it.

As step one, I already plan to put GUID mod support in, meaning you can't join games with mods and you can see from the outside which mods they are using IF YOU HAVE THEM ALSO.

Statistics: Posted by thygrrr — 24 Feb 2012, 21:33


]]>
2012-02-23T22:29:31+02:00 2012-02-23T22:29:31+02:00 /viewtopic.php?t=696&p=7073#p7073 <![CDATA[Re: Broader Mod Support Coming?]]>
Ze_PilOt wrote:
They are far more important things to implement before that.


The TWO things that GPGnet did unforgivably wrong were :

No automod download, so it became annoyingly difficult to get anyone to join a game that required a specific version for 6 or 10 separate mods. It was hard to even get a game with patch 3603...about 1/3 of the people joining would not have even that mod, or would have the wrong version.

No CPU benchmark or way to tell what kind of hardware a player in a game has. So you'd end up with games where a single player would make the game unplayable for everyone else.

And, yeah, custom games were unbalanced. I do like how you've fixed this. But the 2 things I mentioned were the ONLY things anyone I knew consistently complained about. (well, that, and the GPGnet client was slow as a dog)

Statistics: Posted by Habeed — 23 Feb 2012, 22:29


]]>
2012-02-23T22:24:04+02:00 2012-02-23T22:24:04+02:00 /viewtopic.php?t=696&p=7071#p7071 <![CDATA[Re: Broader Mod Support Coming?]]> They are far more important things to implement before that. In fact, you are maybe the first one requesting a mod vault, when replay vault (and a lot of others) is 10x a day :)

Statistics: Posted by Ze_PilOt — 23 Feb 2012, 22:24


]]>
2012-02-23T22:26:05+02:00 2012-02-23T22:20:56+02:00 /viewtopic.php?t=696&p=7069#p7069 <![CDATA[Re: Broader Mod Support Coming?]]>
thygrrr wrote:
Habeed, we'd need the UI anyway because people need to be able to browse mods, too. (where would they get new ones, otherwise?)

But the actual 3-tier data exchange to get a simple mod installed on every client is so daunting, not even GPG (at that time a 100 strong company!) wanted to do it for their own, in-house prestige matchmaking lobby (GPGnet; they could have gone with GameSpy or Steam from the start)


Umm I can think of a trivial way to do it. Just track mods by md5 hash. Any mod someone uploads gets an md5 hash unique to that mod. Before a player joins a game, take a hash of all separate SCD files or mod folders in their mods/gamedata folders and make certain they match every item in a list of every mod a host is using. (you can cache hashes by checking last modified metadata on files/folders so you only have to recalculate them when stuff changes). Any missing or corrupt files you redownload from FAF servers.

Doing it by hash means that you no longer need be concerned about differing versions or corrupt files...a player must have EXACTLY the files that the host has. Any time a mod developer uploads a new version, it is treated as a separate mod. And so on. Once you download the file or folder, packaged as a zip file, you check if a file or folder with the same name already exists on the client. If it does, delete it before extraction (this prevents the nasty problem GPGnet has where you can download 2 versions of a mod, and the older version has more files, so when you overwrite with the newer version you end up with active script files leftover from the old version that will crash the game)


Sketching it out in pseudocode, I do admit it's not 4 lines to make this happen, but it is simple and straightforward.



If Gas Powered Games couldn't do this, they clearly didn't have their competent programmers on it. Given this is a 3d game with high end graphics and an engine so sophisticated it has it's own scripting language, I suspect they just couldn't be arsed.

Statistics: Posted by Habeed — 23 Feb 2012, 22:20


]]>
2012-02-23T22:14:22+02:00 2012-02-23T22:14:22+02:00 /viewtopic.php?t=696&p=7067#p7067 <![CDATA[Re: Broader Mod Support Coming?]]> http://www.quazal.com/index.php?option= ... 45&lang=en

Statistics: Posted by Ze_PilOt — 23 Feb 2012, 22:14


]]>
2012-02-23T22:12:29+02:00 2012-02-23T22:12:29+02:00 /viewtopic.php?t=696&p=7065#p7065 <![CDATA[Re: Broader Mod Support Coming?]]>
Ze_PilOt wrote:
Actually, GPGNet is not an in-house solution :)

True it was contracted to... what was his name? AgentSmith and his crew.

Statistics: Posted by thygrrr — 23 Feb 2012, 22:12


]]>
2012-02-23T22:10:52+02:00 2012-02-23T22:10:52+02:00 /viewtopic.php?t=696&p=7064#p7064 <![CDATA[Re: Broader Mod Support Coming?]]>

Statistics: Posted by Ze_PilOt — 23 Feb 2012, 22:10


]]>
2012-02-23T22:12:51+02:00 2012-02-23T22:09:40+02:00 /viewtopic.php?t=696&p=7063#p7063 <![CDATA[Re: Broader Mod Support Coming?]]>
But the actual 3-tier data exchange to get a simple mod installed on every client is so daunting, not even GPG (at that time a 100 strong company!) wanted to do it for their "own", branded matchmaking lobby (GPGnet; they could have gone with GameSpy or Steam from the start)

But this whole discussion has given me a bunch of ideas, and we could incrementally look into "broader mod support", as opposed to "pervasive mod support", what this thread seems to be about. 8-)

Statistics: Posted by thygrrr — 23 Feb 2012, 22:09


]]>
2012-02-23T22:03:25+02:00 2012-02-23T22:03:25+02:00 /viewtopic.php?t=696&p=7062#p7062 <![CDATA[Re: Broader Mod Support Coming?]]>
Detecting the mod in the lobby and auto-downloading/upgrade it if you don't have it IS NOT something GPGNet does.

In fact, I'm not really sure what you mean exactly with your explanation :)

Statistics: Posted by Ze_PilOt — 23 Feb 2012, 22:03


]]>
2012-02-23T22:01:19+02:00 2012-02-23T22:01:19+02:00 /viewtopic.php?t=696&p=7061#p7061 <![CDATA[Re: Broader Mod Support Coming?]]>
Ze_PilOt wrote:
Auto-mod download for GPGNET ? We are not talking of the same things, because GPG doesn't have that at all.

They've got a mod vault, that's true. But that is not a high priority for us. But if you are a developper, you are welcome to help us doing it :)

There's no difference between a manual click by a user and an automatic download by the lobby software. The code to make it happen is 99% identical. In fact, the manually clicked version requires more code because there has to be a GUI.

Statistics: Posted by Habeed — 23 Feb 2012, 22:01


]]>
2012-02-23T19:25:49+02:00 2012-02-23T19:25:49+02:00 /viewtopic.php?t=696&p=7056#p7056 <![CDATA[Re: Broader Mod Support Coming?]]>
Habeed wrote:
What are you going on about? I am a programmer by trade, and I can also see when another implementation works.

GPGnet has allowed for auto-mod downloads and installation since Supreme Commander has released. This feature has MOSTLY worked the whole damn time.(it doesn't handle updates to a mod very well because it doesn't delete all the files in a folder before installing updated files) The ONLY thing GPGnet doesn't do is check the mods enabled in a game and force you to download those mods before joining. That is NOT a difficult feature to add!


On the contrary: GPGnet has purely manual mod downloads, just automatic install (which is trivial). The hardest part is getting the right list of mods (and their versions) to use from the running game lobby, and reliably map that to a download. To perform this mapping, the GUIDs of the mod need to be extracted, put in a DB, etc - both on the client and the server (though the client's filesystem can be abused to do that, but it's one of FAF's design goals to be compatible with GPGnet content). The mods then need to be parsed at submission and install time by client and server, and verified at launch time by the FAF client, in addition to the engine side code (which harbors SupCom's heavily modified Lua engine). The actual download is 3 lines of code in python. The mapping and communication of mod ids, a small project of its own, including building a database, mod reception and listing code, etc., spanning all three components of FAF: Lobby, Server, and Game.

By the way, the Map, Mod and Replay Vaults were added a quite some time after the initial release of Supreme Commander. :)

But it's great to hear you are a programmer, would you be interested in contributing a mod vault with auto downloads? I can spawn a new module in the lobby code for you to work on. Maybe we didn't put enough thought into it and you have an easier solution in mind - that'd be terrific and the community would be grateful. :)

Statistics: Posted by thygrrr — 23 Feb 2012, 19:25


]]>
2012-02-23T18:33:30+02:00 2012-02-23T18:33:30+02:00 /viewtopic.php?t=696&p=7053#p7053 <![CDATA[Re: Broader Mod Support Coming?]]>
They've got a mod vault, that's true. But that is not a high priority for us. But if you are a developper, you are welcome to help us doing it :)

Statistics: Posted by Ze_PilOt — 23 Feb 2012, 18:33


]]>
2012-02-23T17:57:47+02:00 2012-02-23T17:57:47+02:00 /viewtopic.php?t=696&p=7052#p7052 <![CDATA[Re: Broader Mod Support Coming?]]>
thygrrr wrote:
Habeed wrote:I mean that users could submit their mods to be qa'd, and you'd hopefully review the mod reasonably promptly. Kind of like the "app store" for Apple products.


Cool idea, we'll earn billions of dollars that way... no, wait. :lol: We won't. But at least we'll triple our profits. Oh. Wait. We don't have any. :?

The key motivation behind featured mods is to have somewhere between 5 and 10 really cool mods, maybe phasing some out when they become less popular, to showcase the very best that FA and FAF can be. For that, we try to make it as easy and "click, click, play!" as possible to get these mods. That takes special care on the side of the lobby, the server, and the mod maintainers.

Featured mods are carefully selected, the creme de la creme. You can always make suggestions, but chances are, if it's not rock-solid, popular as hell, or truly extraordinary - the mod is not gonna get featured. Your chances are likely to improve if you sponsor a mod with your own development time, because it's not guaranteed that a mod will work with FAF out of the box. It may need tender loving care with each new version of FAF, too. Maintaining mods is hard work (kudos to all those who do!)

That said, there WILL be extended mod support meaning that one day, the lobby will at least be able to list the mods used in a game. But we can't provide downloads for them, at least not easily. A real "mod vault" is a distant dream, also considering how badly the GPG one got cluttered with crappy mods and crappy versioning. Brrrr.

So, it all boils down to: If you have it in you to be a contributor to FAF, by all means, step forward and help us build a feature like this. It's just not a priority seeing how low on time our current team already is.


What are you going on about? I am a programmer by trade, and I can also see when another implementation works.

GPGnet has allowed for auto-mod downloads and installation since Supreme Commander has released. This feature has MOSTLY worked the whole damn time.(it doesn't handle updates to a mod very well because it doesn't delete all the files in a folder before installing updated files) The ONLY thing GPGnet doesn't do is check the mods enabled in a game and force you to download those mods before joining. That is NOT a difficult feature to add!

Statistics: Posted by Habeed — 23 Feb 2012, 17:57


]]>
2012-02-23T17:50:28+02:00 2012-02-23T17:50:28+02:00 /viewtopic.php?t=696&p=7051#p7051 <![CDATA[Re: Broader Mod Support Coming?]]>
Habeed wrote:
I mean that users could submit their mods to be qa'd, and you'd hopefully review the mod reasonably promptly. Kind of like the "app store" for Apple products.


Cool idea, we'll earn billions of dollars that way... no, wait. :lol: We won't. But at least we'll triple our profits. Oh. Wait. We don't have any. :?

The key motivation behind featured mods is to have somewhere between 5 and 10 really cool mods, maybe phasing some out when they become less popular, to showcase the very best that FA and FAF can be. For that, we try to make it as easy and "click, click, play!" as possible to get these mods. That takes special care on the side of the lobby, the server, and the mod maintainers. Offering mods that "simply work" instead of a huge list of mods that "should work" is the key priority here.

Featured mods are carefully selected, the creme de la creme. You can always make suggestions, but chances are, if it's not rock-solid, popular as hell, or truly extraordinary - the mod is not gonna get featured. Your chances are likely to improve if you sponsor a mod with your own development time, because it's not guaranteed that a mod will work with FAF out of the box. It may need tender loving care with each new version of FAF, too. Maintaining mods is hard work (kudos to all those who do!)

That said, there WILL be extended mod support meaning that one day, the lobby will at least be able to list the mods used in a game. But we can't provide downloads for them, at least not easily. A real "mod vault" is a distant dream, also considering how badly the GPG one got cluttered with crappy mods and crappy versioning. Brrrr.

So, it all boils down to: If you have it in you to be a contributor to FAF, by all means, step forward and help us build a feature like this. It's just not a priority seeing how low on time our current team already is.

Statistics: Posted by thygrrr — 23 Feb 2012, 17:50


]]>
2012-02-23T17:47:31+02:00 2012-02-23T17:47:31+02:00 /viewtopic.php?t=696&p=7050#p7050 <![CDATA[Re: Broader Mod Support Coming?]]>
Ze_PilOt wrote:
We can do almost anything, the problem is the time.

Right. Is the system call to retrieve 2 strings (CPU ID, current operating frequency) a 15 minute task or a big project? The only obstacle I can imagine is if the language you are using (Python?) doesn't allow you to make those calls.

The second task though would be to automatically look that CPU up on a benchmark site and report an actual expected performance score. THAT would be kind of complex (you'd have to search the page for the closest match, and then also scale to correct for the performance boost from overclocking. Well, don't search the page in realtime, search a data table scraped from one of the web pages for a CPU score site)

Statistics: Posted by Habeed — 23 Feb 2012, 17:47


]]>