Developers' Diary #1  
From the creation of the game concept to the development of a next-generation graphic engine.

Jeremy Smith Lara Croft is back!When starting to develop Tomb Raider 1, I don't think any of us thought we were working on a franchise that would go on to sell nearly 30 million units and have a successful movie launched, with another one ready to go. We were just doing what we enjoyed - making games. The reason for stating what is pretty obvious, is that we created four further games from an idea which was only really for one game.

In truth we probably made two games too many, but that is now history and we can only learn from that and look to the future on the new consoles.

When we all sat down three years ago to start planning the next generation Tomb Raider, we were faced with some tough decisions. There were key things we didn't want to change and other areas which needed expanding on. The big advantage we had was knowing that it is one of the largest franchises in the industry and that this time, we would be doing more than one game - the fact is we sat down to design a trilogy.

The design team consisted of the lead members of the team plus myself and my brother. The team was adamant that this game would be the first of the Tomb Raider games where the player will actually understand the story line. To make this work, we need to ensure that the plot unfolds as you play the game, guiding you through the story right to the end.

So, our initial work was to come up with a story and locations. We knew we didn't want to go into "sandstone coloured textures" again, so the settings had to be more real life without losing the elements of fantasy that the original Tomb Raiders had. We also had to expand the character of Lara. To help us communicate the plot, Lara was going to have conversations with other characters in the game. To continue with new attributes for Lara, we also wanted to build into the game Lara having far more interaction with the environments.

The final key new enhancement to the game was for us to introduce a new partner for Lara. We really felt that to expand the whole Tomb Raider experience, it would be really cool to introduce a completely new game dynamic with a new character for you to play.

After about three months of many design meetings, we arrived at a pretty solid game design which included a strong story line, new Lara attributes and a new character called Kurtis.

The Angel Of Darkness arrives after a long period of development.

Whilst the game design was being worked on, the programmers were assessing the capabilities of the PS2. In the early stages of development, this was difficult as we didn't have PS2 development kits so there was a lot of theories being designed which of course can create problems when you are telling your level designers you can have 'x' amount of memory to build levels, then 12 months later you actually find out you can only have half the amount you were originally told, which then means all the maps that have been built have to be rebuilt. I am sure you can imagine the reaction from our level designers!

Not only during the design stage does the story need to be locked down, so does the technical capacities of the engine. We know that gamers would expect a state of the art engine to go with the new design and the first thing the coders agreed upon was that the game had to run in one frame (or sixty seconds). This without doubt separates us from all the other games. Having a game running in a frame gives you such a smoother game experience. This is one of the biggest issues we have had with the development, trying to keep it in a frame and there have been times when we have considered dropping it to two frames, but now it's finished and runs in a frame, I'm very pleased we didn't.

Kurtis, the new character created for TAOD.
Lara, sexy & dangerous.

Throughout the design stage we continually looked back at the previous games and tried to analyse which parts people liked and disliked. Again, this is not as easy as it sounds due to the fact that the games have sold so many copies, trying to get a subjective opinion is sometimes very difficult. When you talk to gamers who played Tomb Raider 1 and 2, many of them don't like 3 and 4 and if you talk to players of 3 and 4 and ask if they liked 1 and 2, most would say yes. We found that people who bought the original Tomb Raider 1 and 2 are far more hardcore gamers than gamers who bought 3 and 4. 3 and 4 gamers seemed to be more mass market. We did change the game dynamic of 3 and 4 deliberately aiming it at a more mass market consumer, but we hadn't reckoned on losing the gamers who bought 1 and 2. So, our dilemma for Next Generation was which consumer do we aim the game at? Of course, the only real answer was both - which I can assure you is easier said than done - and the truth is, we will not know if we got it right until the game is shipped and on the street and both sets of gamers have played it. We did however try to cover all bases.

Designing a completely new engine on new hardware is not a task to be taken on by the faint-hearted. There are many talented developers who have to date tried and it hasn't worked for them.

We have written a new engine in three years with a programming resource of 7 people and a total team size of 35. Metal Gear Solid 2, which I consider to be the only other real next generation engine to have been written, took 5 years with 100 people. The lessons we have learned moving forward is that our team size needs to increase to ensure a smoother transition of engines plus we broke a golden rule - never develop a game while you are developing your technology because you can't be efficient continually chasing your technology. Lessons to be learned for the future.

- Jeremy Smith (18 Jun, 2003)