Hey Relight, here you go - transcript from GDS 2010, presentation by Jarek Kolář, held in Prague.
Here is the video:
Video, exclusive for www.mafia.czech-games.net
Translation starting from 5:00:
Thanks for the introduction. I have worked on Mafia II from 2008, first as a manager, later as a producer in a game design and gameplay. This presentation should be taken as the first comprehensive summary of Mafia II development.
So...I'll tell you some things, then I'll maybe show you a few videos and at the end of the presentation we *unable to understand, terrible echo*
=====YEAR 2002 screen
So...in 2002, after five years, almost five years, things finally went as they should and the game Mafia I, developed by Illusion Softworks, shipped for PC platform and scored a great sucess. Most of the team began working on PS2 ans XboX ports afterwards, and that guy over there in the back, Daniel Vávra, welcome on the presentation, started gathering ideas for Mafia II script.
In 2003 Mafia I console ports were finished, it was rather difficult as the game was not meant to run on PS2 nor XboX, it was heck of a work and if you had played it you know outcome was not the best, we had some problems with data loading so the goal was, since the company decided Mafia II is the deal, the goal was to make the game primarily for these platforms, to ensure it will run smoothly on them and will be released simultaneously with the PC version.
The goal was to make the sequel quicky, take advantage of the drive Mafia I has caused and quickly punch a sequel together. That means, as you know, games are always depended on technologies, they need some engine which "runs the game". That means, since the original engine used for Mafia I was not console-friendly, the company decided to buy a licence for an engine, it will licence an engine and it was a software named RenderWare. It was a rather popular engine, the whole GTA series, Vice City and San Andreas used it so it was quite suitable for PS2.
So, technology was licenced so we could start with a design.
=======Year 2004 screen
Then 2004 came, we finished the pre-production, had our technology ready for use and it looked like Mafia II can shift to production phase and be finished in no time. But what happened...I'll tell you in a minute, first I'll present you my video, showing how the game looked in 2004
====FOOTAGE of RenderWare-powered Mafia II alpha.
So you can see RenderWare logo, we used *unable to understand due to the awesome music*
Here you can see a multiplayer *unable to understand the rest*
=======Year 2004 screen
So we almost had it, just create missions, characters...But what happened in 2004, Electronic Arts bought Criterion company, owner of the licenced engine, and the next year EA completely stopped the promised engine development, so the agreement voided, Illusion Softworks didn't licence RenderWare, so...you know the rest. So what you have seen here ceased to be acceptable in 2005.
=====Year 2005 screen.
2005 was also special thanks to the next round of "Console Wars" - and it happened this way: Microsoft, establishing itself on console market with their XboX, needed something to outrun so far leading PS2, so they released their XboX360 already in 2005, of course Sony reacted by official statement regarding PS3 console. That meant PS2 consoles, Xbox consoles...these won't last long and it is needed to adapt to the new standard.
So...thanks to this "war", and also due to the unability to use RenderWare anymore, the company decided:
- for first, need to find some technologic solution other than RenderWare
and
- for second, the game will be "next-gen", aimed for XboX360 and PS3.
That meant complete development restart, re-estabilishing project goals again, all art done up to that date needed to be redone, as next-gen, that next-gen technology, simply the platform allowed vast variety of possiblities art-wise, and we needed corresponding next-gen technology.
Of course Unreal (engine) was a possibility but as you may remember, maybe not, during that time Unreal omitted Xbox 360 support and only PS3 support was planned, so based on the experience with RenderWare, the company feared Unreal could be again victim of some marketing speculations and the project could end up with no technology again, so the company decided to create its own engine instead.
During that time Illusion Softworks finished some other projects and the company realized if they want to go for the next-gen, maintaining parallel-development of several titles is unbearable, so it decided to focus on smaller number of projects, some people moved in so suddenly on Mafia II worked a few dozens of employees instad of previous twelve or so.
The company also realized, we got some managers from abroad, they were far more experienced than anyone from us, they started introducing us to some new methods of development management and so on.
Then 2006 came,
===Year 2006 screen
it was important, the technology we were speculating about, it was originally developed for some other project here in Prague NOTE(this presentation was held in Prague, Mafia II was developed in Brno)/NOTE , in Prague department, it was so called "Unified Engine", the company tried to use all of its experiences with videogaming coding, unify all used coding methods; later it was renamed as Illusion Engine, which is now used by us.
===Year 2006 screen
In 2006 we began with storyboard creation for all animated sequences as well, we began with americanisation of the script, that means we took what Daniel had written and it was translated, in U.S. JAck Scalici started to rewriting it to suitable style of U.S. English. We begain our work on scenes, creating the city once again, levels were worked over as you could see, and we continued with design and so on.
In 2007 PS3 was finally available as well
===Year 2007 screen
and during that time, we realized, or it began to be realized by our programmers from different platforms that our engine is not optimized well, we had problems, so general revision of "Unified Engine" began. We started with implementation of streaming, memory management, we worked on rendering so the engine is ready to run smoothly on all platforms. Parallel to that creation of graphics, cutscenes continues as well, we can see the first results which look fantastic, pictures we could see indicated the game will be neat, will look superb. That's why, despise the project was quite delayed, the technology was still not ready, the company as a whole and T2, who supported the game finacially, had good feelings the game will be fantastic even tough it takes too long to create.
However, thanks to the still-not-ready technology, our programmers were not able to start finetuning various design aspects, gameplay design aspects, so we were not able to to iterate our ideas, simply, we were still creating the core of our technology, that lasted, the whole technology creation process took its time, scene viewer...but so far no scene was created.
That changed at the end of 2007 when we...when my co-workers put together the first demo, which showed player movement, character movement...gameplay mechanisms used in demos were created so they can be used in the actual project, it could be seen how the game's going to look and play like.
Also, during that time, it was a bit complicated as I have said because all payments and financing was done by milestone system, due to technology delays there were payment delays as well and the situation was not easy for us. It lead to the deal
===Year 2008 screen
between company's management and T2, and solution was the aquisition we all know about - at the beginning of 2008 Illusion Softworks company merged with 2K, the family which is held by T2. This aquisition showed T2's trust to the whole team and the project as well, they were blamed by public but it took some time before the game was finalized from that point, but they trusted us thanks to our presentations. Thanks to that, thanks to the aquisition our work significantly simplified, because ther was no longer need to compete with time for milestones, but whole time management as well as financing was T2's matter. We, developers, could fully focus on the game, on the development itself, we ran detailed analysis of project's parts: code, design, script, completely without any politic intetions, aim was not to lie but to talk on the line.
T2 participated by giving us vast sources of experienced people, we were given some pros directly from 2K, experienced in videogaming creation, either they directly helped or they were consultants, we had there producer of the Ghost Reckon, before he worked as a lead programmer, we had about 15 guys from BioShock team, we had lead designer from the original Splinter Cell, these all came, saw problems of the project and adviced us how to solve them and what to do, they stimulated the project, which was by no means in a full working status, to move forward, to advance sucessfully.
We also had major team restructuring, originally we were divided into graphics, animators, coders...graphics, animators, coders, designers; the new team structure was more pre-production based, that means we had city team, mission team, A.I. team,player team, graphic team, act team, animation team, technology team, test designers, testers. The point of this new structure was to empower central management, to focus on each component of the game individually and finish pre-production, which was in that time still not over, everything "existed" only on paper.
===Year 2009 screen
So we are starting 2009, when according to the original plan Mafia II should ship yet at the beginning of the year we were still missing some technologies. ??? technology (unable to understand), it had always some problems, development problems, it was always late; whole city traffic section was being rewrited as previous version was still the one designed for RenderWare, still not acceptable A.I. and we were yet to begin with optimization for each platform, XboX360 and PS3. So due to all these problems it was not until 2009 where the team went into pre-production as a whole (all parts of the project). Also the team grew bigger, about 150 people, they were drafted from 2008 and on, after some training they joined the project - coders, animators, graphics, scripters...simply put, huge pack of guys working on production so the Mafia II can be finished in time.
Due to various changes in technology we were able to do regular checks how the works progresses, first they were on roughly month basis, later we had them each week and in the end we had them each day. Also in the U.S. there started *unable to understand*. In 2009, as you already know, the game was presented to journalists for the first time, for us developers it meant excessive amount of work as we had to create all those demos; it is unbelieveable how many demos we were forced to create: in March 2009 we showed demo for ???, then we did E3 demo, then GamesCom, so it was third demo, and other one as preview version and ??? demo so we had to do all this extra work.
It was for good in some sense as these demos were the first prototypes gameplay-elements wise, showing how should the final product look like. Even so, it was unpleasant as many features were still defunct and had to be somehow "emulated" for demo purposes in order to show the bright side of the game.
===Year 2010 screen
In 2010, at the beginning of the year, after Mafia II didn't launched in 2009 as originally planned, it was not possible to continue with various features development, so all unfinished parts of the game were left unfinished, so the team could focus on finished and working parts of the game, to test and debug them. All adding of the content was terminated and we focused on debugging. Also the first focus tests took place, that means we had some people play the game to learn whether the game is player-friendly enough, tutorial is understandable, it can be played well. We also worked on localized versions as we planned simultaneous launch of all localized versions.
And what is also important, due to consoles the game had to be ready with enough time left so it can pass its rating in time, before announced launch date, so finalization took place in May 2010, we finished the game in May 2010 and had enough time for administrative things, to have this game authorized by Microsoft, Sony.
So, this is how the development actually looked like. Now let's have a look... What went wrong. The first and the most painful thing was a timing of the project. Originally, our goal was to make the sequel quickly, about two years after the first one. This trend to do it quickly was still present, so coders were forced to do their work as fast as possible so we can have representative parts of the game done and ready; thanks to that we had no time for solid project background. Another major issue was licencing, we had problems with RenderWare as well as with other parts, for example physics, licence for it was passing between companies, we had problems with licences for sound and memory management software, mainly for PS3, as we developed our game, Canvas and that other company, sound software company, were finishing PS3 versions of their software, so were ready and were simply waiting for them to give us required software, so we had quite a lot of problems with it.
Another problem was editor, originally intended to be used for all Mafia II versions, but as time went and new and new features were added it grew into a monolith which would be impossible to maintain. Also various people, who came and went away from the project, created their own tools to ease their work with the graphics, animations and so on.
Another major issue was platform optimization. Due to the decision to ship the game on platforms PS3 and Xbox360, we had to make sure it will run smoothly on them; we have created rather complex packing/unpacking system for the game data, so when somebody made some change, he had to compile whole chunk of data over and over again just because of some small change, in a long run it got very time consuming, be it ???, ???, graphics, animations. And another huge problem was development of the technology itself, it was not until 2009 A.I. started working as it should and other parts as well.
What went right. Thanks to the restructuring and various revisions, we were able to cover up for early mistakes caused by not very much planned-ahead software development, also we set up technology team structure with each team leader being held responsible for the according part of the development and code. We also established new positions which were nonexistent at the beginning and the whole team started working a bit different than back at the beginning of the development. Thanks to all of this, the code, the program has finally formed its shape, all teams maintained technologic standard of their part of code, so when we sent the game for authorization, after such a long development, it was good enough to pass Sony and Microsoft validation for the first shoot even tough it is not typical for such a large project. Also, by creating our own technology we have it completely under control, ready for our future games.
What went wrong design wise. The whole script work was done independently, not in coordination with designers, who were even scattered over the country. Nonexistent technology, as I have said you cannot make revisions and refine your ideas without some engine, everything was only on paper. Another really major issue was finished graphics and animations before implementing them into game design. So by the time designers got A.I. code and gameplay mechanics to work with, graphics, animations and cutscenes were done, so under these circuumstances even ordinary revisons of the code were extremely difficult. Other projects have usually several iterations, but in our case we had environment modelled, detailed and ready, so any significant changes would mean heaps of extra artist work and other problems.
What went right design wise. Everything we have planned in 2005, after the project restart, was in the game, everything we wanted to be in is in. Regarding focus testing, the game was playable for all sorts of players without random difficulty shifts, unlike Mafia I, friendly for all players. Also we had time for debugging.
Random hate&rage&flame attempt. Ignored big time.
Regarding development as a whole, the biggest mistake was underestimation of the next gen platforms and lots of the work had to be done twice in fact. Also we failed at pre-production phase, it took too long since 2003, ideal pre-production is when everything is ready and tested, and another big issue was milestone-type development at the beginning of the project. Everybody just wanted to achieve another milestone in time, but the project idea as a whole was slowly shattering. After the aquisition we had similar problem as everybody wanted to finally pass the pre-production phase, so programmers we forced to do their work as fast as possible, often paying too much attention to the audiovisual side, ignoring core of the technology and other elements; these of course heavily suffered from it.
What went right with development was the aquisition; it literally saved the project, when I look back, in 2007 the project was on the edge, close to be cancelled, so aquisition and subsequent actions play significant role in a fact this game has shipped. We also successfully passed production phase, took about a year. Regarding debugging, testers found unbelieveable 60 thousands of bugs, so we had to kill them all, send back to testers, sometimes the bug appeared again but in the end all 60k were successfully eliminated. We also prepared eight different language versions, seven being released here (he probably means Europe), with dubbing as well, considering the script has about one thousand pages it was extraordinary amount of work, multiplied by the fact the game shipped on three platforms, so in the end we maintained about twelve Mafia II games. All of these we had to create, test, send to the U.S. for test and approval over and over again. And for the whole time, we heavily supported 2K marketing.
That's all for the development. Now I would like to show you some videos, at the beginning you saw some motion capture sessions with real life actors and scene settings, those films have about two and half hour. If you promise me you won't record this, like you, dude, I will show you a video...
*camera turned off*
DISCLAIMER: THIS IS A FAN-MADE TRANSLATION. I HEREBY APOLOGIZE IN CASE I MISINTERPRETED SOMETHING, NOT INTENDED.
Andrashi
![[GST]GhoST0613](/data/avatars/m/1132/1132257.jpg?1337038451){kind=avatar}
.