FAF - post-mortem.

[Ze_PilOt]

As we are slowly going toward the first release, I will do a little post-portem of the development of FAF.

You will discover a story made of emotion, joy, suspense, and community splitting.

[Ze_PilOt]

PART ONE : The silly idea.

I was playing Starcraft 2 with kekouse. It was bad and boring.
We regret FA.

Me : "It's sad, Galactic Wars got a lobby, okay the GW part wasn't working, but with the lobby, some patch, it was enough to make the game live longer !"
Kekouse : "We will never have another game like FA in our lifetime !"
Me : "Hmmm.. Wondering how hard it would be to make something like GW..."
Kekouse : "You will never make it !"
Me : "Look at TA Spring.. We make a simple irc client for chat, with pyQT I can prototype a quick interface.. My god, it's really possible !"
Kekouse : "Ok let's (you) do it !"

After contacting some old modders on GPG, get some "you are silly" answers (I particularly remember deadMG saying it's too much work for a couple of guys, for a dozen players , I start trying to get a connection with FA working...

What you will see is maybe a big part of our recent history :



What you see if the first FAF server. And more amazingly, that part of the code is still inside the current version !

We are the 25 June. You can say it's the official birthday of FAF.

What you see is FA connecting to the server, and the server asking it to create a lobby.

"Hurray !!!" "It's totally possible and not that difficult !"

What a silly thought...

After one or two day, I manage to have a working lobby for a single player.



(you can see how happy I am : "hurray, it's working" said the chat.)

We are the 27 June.

Ok. Server done !

Now, the lobby.

Here a screenshot of the first version of the lobby, ever. It connect to a IRC server and do nothing, but proudly.

We are the 29 June.




30 June, it now connect to the server, with a password and all the fancy stuff (actually, without the fancy stuff).



Here, I thank the patience of the early alpha tester, who spend evenings trying to connect to my server, over and over again : Kekouse, Ephez, Olivarius, Nesar, I hope I don't forgot anybody.

4 July :



The first connection between two players, and the first game played on FAF ! (I won)
FA was launched manually

You can recognize the old game lister. He lived long
The penis icons didn't make it.

Two days later, the first alpha was out (with auto-launch of FA), running on my mediacenter.

[FunkOff]

Two days later, the first alpha was out (with auto-launch of FA), running on my mediacenter.

And you started advertising it on GPGnet, right? That's about when I latched on to your idea. I had wandered back onto GPGnet for the first time in at least two years and I was thinking, when you talked about FAF, "What a great idea?". And since then, you haven't been able to get rid of me!

[Ze_PilOt]

That was a little later, when I got a first real server. It will be in the second part

[Ze_PilOt]

PART TWO : The others.

As I said in part one, I've left FA for about one year, playing other games.

FA wasn't on steam yet, and the stats of GPGNET were quite poor when I've started the project :

66a9a802c227e8ea9a45560d685e3.png (37.45 KiB) Viewed 3347 times

It was now or never to try to resuscitate this game !


Others projects like FAF were tried during the long life of FA.

Galactic War

As previously mentioned, it was the idea that GW would have been a wonderful lobby that motive me to do it.
But my "design" blueprint was TA Spring : Efficient, fast and simple.
For those who don't remember GW, only a beta came out.

Sadly, the client was leaking memory, the server wasn't holding the charge, and the beta ended after a few weeks.
After that, many promises were made.

But as the GW team admitted it, their ambition was 1000x bigger than the team was able to do.

They did a wonderful job and what is following is not a bash of any kind, but my analysis of their mistakes in order to not make the same.

- They only had one coder, who (seems) to sadly listen to every demand of the team, even the extravagant ones (3d galaxy map,....)
- While the beta was promising, they've closed it. Having people supporting a project and see the improvement over time is really important and motivating for that kind of project.
- The server was not efficient enough, the client wasn't work that well, but instead of fixing that, they've added features over a not working version.

So, with that in mind, I've fixed myself some realistic milestones, and a precise roadmap.
Here is it : https://bitbucket.org/thepilot/falobby/wiki/RoadMap

The plan was to release it to public as soon as possible.
It's old and not updated since a while. I've changed the order of priorities over time, but retrospectively, I'm happy : I've met all the deadlines with a LOT of advance on the schedule !
The version planned at the end of the year was ready on september.

Server-side, my "goal" was an efficient and simple server :
- Must run on a single core Atom (cheap server).
- Must take at most 1 go of memory (the current server is running at 4-500 mb with 300 players online).
- Must be able to support 1000 players (it's optimistic, but it help keeping the code optimized).

For the code, I've decided to prototype the thing with python. I've finally kept it because it's simple and fast to code, and the speed drawback is irrelevant for that kind of project.

I've nevertheless asked the GW main coder if it was possible to have access to the GW code.
All I got, for unknown reasons (I didn't insist), was a single old file, doing what I've described in part 1. So, useless to me.
It still quite sad because all their work is lost forever now. (I don't think anybody else will try to take over the GW project if I fail).

FA-Ladder

One of my biggest mistake, and I still asking for forgiveness about that.
As I've said, I've left FA for quite some time.

When I left, FA-Ladder was really young, and was just a mod and a replay vault with a ladder prototype.
When I've started FAF, I've honestly didn't though it was so advanced (mostly maybe because the name didn't changed even if the possibilities were bigger). It wasn't really advertised either, so I didn't though much about it.

The problem was that FAF and FA-Ladder goals were the same. After launching the first public beta, someone said to me "Eh, you should contact Moritz !".

And I did saw what FA-Lobby has become, and I was impressed.

At that time, FAF had more community tools (chat, player list, more obvious game-lister), but FA-Lobby was way more advanced on all other areas.

So I've asked Moritz to join effort. Sadly, as it used Java, it was a hard/impossible job to implement the code.
And I've killed FA-Ladder.

That was not the goal, I'm not happy about it. Moritz is still part of the FAF contributors, so I hope no hard feelings.

PlayerTrack

More a player database than a real lobby alternative.
While I've loved the stats (and FAF will propose at least that amount very soon , I didn't liked the ratings.
I understand the goal, but it was really biased.
I was also very interested in the rank icons in TA-Spring.

People like awards and stats. I've put some though about all that and came with the idea of implementing a real rating system for all games, replacing the biased user rating.
That's where trueskill integration came from.

The community

I skip ahead in the story, but again, when I've left, 3603 was just made as a mod, all hopes of a official patch was gone, and I've stayed with the idea that everybody was eager to see a real 3603 patch for FA.

When I came back with FAF, I was so happy to tell everybody that it was finally possible.
Then I've discovered the horror.

It kind of remind me the movies where the hero leave town for a crusade, and discover a twisted version of the town he loved when he came back, just because the artifact he have seek, needed by the town, was not there at the right time.

I was truly horrified that two cults were born : 3603 haters and 3603 lovers.
Not everybody was happy about a patch. I still not understand why and how it became like that, but again I've skipped one year of FA history.

Koecher launched the term "Community splitter" toward me, Moritz told me that he didn't advertised FA-Ladder too much because of the 3603 haters,... "WHAT HAVE YOU BECOME FA COMMUNITY !" I was screaming in my head.

I've left nice bunch of players, hoping for GW, 3603, shiny new things.. to discover a bunch of integrists in both side.

Well, guess what, I'm still happy to give patching possibilities to FA, and seeing how diluted the "clans" are now, I'm think I was right to not give up (was really close to) at that time.

[pip]

This is too good a story to be put in the forum. You should make it into the wiki!

[Kekouse]

GG ZeP!
It's quite amazing to see how far FaF went from our little conversations and frustrations over Starcraft2, 1 year ago. I never said you couldn't do it....my only fear were on the long-time commitment needed for such a project. And see...I helped a little bit in the beginning but now I've done nothing for over a year. Same for Ephez, Olivarius and the other "frenchies".
Now you have Raging_squirrel, Thygrrr, FunkOff, Moritz and modders to help you and that's great because it means the project will go on for a looooong time.

So really big kudo for the hard work!
And I will come back someday...when I will have some more time for FAF

P.S:
Et bordel t'écris mieux en anglais que moi désormais. Petit con va!

[Ze_PilOt]

Hey ! You've just spoiled the next part !

[TA4Life]

Thank you for your stubbornness, staying on this project in spite of a divided community. The greatest thing about FAF is the simplicity and the user friendliness, as soon as you install it, it takes care of you with a single click. This will have a huge effect on the community I believe.

The new golden age of Supcom is just around the corner.

[Ze_PilOt]

I'm not the only one

For the first alpha version of the lobby, a lot of people helped me testing the functions. Ephez, Kekouse were already cited, but the whole Canard PC (a french community) FA community helped too.

But several other person helped me coding FAF.
As previously stated, one of my goal was to use the more straight-forward solution : IRC for chat, web for as much things as possible.

Live replay server
The first live replay server is quite older than FAF, and was made by Mancer. Batto helped adding some functions like the delay and debug others : http://code.google.com/p/openlivereplayserver/
I later replaced it by my own python version.

Trueskill
Not a direct help, by it's a direct implementation of moser code in python (and now lua) : http://www.moserware.com/

Map vault
The first version of the map vault (web) was made by Olivarius. Same for the first leaderboard.

Icons
The tanks tech icons was done by Ephez.

FA Lobby
By Moritz, obviously.

The server
The first FAF server was hosted by Wolfmic.
That was the opening of the first public alpha version, the 05 July 2011.

That version was very simple. No auto-update, no matchmaker. The chat was there, the custom games too, and that's it.
You can see what it's looked like in the previous post. (minus the d*** icon - by Ephez too