Vice City was a mad dash to get done in time. Everybody was worn down and it was clear another one year game was not an option. In addition; it was obvious that Vice City hadn't innovated that much and another 're-skin' wasn't going to fly. This is why San Andreas got 2 years of development time.
Our confidence as a team was sky-high and with the longer development cycle we felt we could be ambitious with the new features for SA. And ambitious we were.
Gta1 had 3 cities and we wanted to do the same. There was one particular meeting that I still have nightmares about. In R*N we would usually make the big game decisions with 5 people. We had a meeting where 3 people wanted the cities to be on different maps. This would save memory as the models for the skylines of the cities wouldn't need to be in memory at the same time. Gta1 had 3 cities on different maps. The player would take planes/buses/trains to travel between maps.
I wanted the cities to be on the same map as it is important to be able to drive between cities. Even if this meant we'd have to find some memory elsewhere. The 5th guy wasn't there that day. I wasn't able to convince the others. It was super frustrating as it seemed we were about to make a massive mistake.
The following day we had another meeting. The 5th guy agreed with me and was much more persuasive than me. He talked them round and the cities ended up on the same map. Disaster averted.
(I am often spectacularly wrong. For a while I wanted gta3 to be topdown. I implemented the topdown camera and argued for it to be the default)
Usually when Leslie Benzies had a strong opinion on something, that would be just what we did. He seemed to have the best judgement.
We took the code from Vice City as the foundation for SA. Many things were improved but no structural improvements were needed.
R*NY organized another 'research' trip. They hired a bunch of town cars and we drove between Vegas, LA and San Fransisco. The artists took pictures. It was loads of fun.
The RPG elements caused real problems. If the player ate too much, CJ would be too fat to climb certain walls or run fast. This made certain missions impossible. QA had to catch all these cases.
When the player got hungry, his stomach would rumble. Players started going onto Mount Chiliad to look for big foot (one particularly persistent rumour). Many players thought they could hear big foot not realizing it was their rumbling stomach.
So why did we include so many features? I particularly remember sitting in a meeting where we decided to include stealth in the game. We didn't need it as we had so many new features already. Nobody was particularly passionate about it but we did it anyway. We made things unnecessarily hard for ourselves as we were struggling to get on top of the bugs at the end. Ambition and overconfidence.
The 3 cities with the country side in between, almost killed the map artists. gtaIII and Vice had maps of 4x4km and even then there was loads of water. The SA map was 6x6 and densily packed.
New York was particularly active with research. They came with large lists of things that the various gangs in LA do. They insisted different gangs would have different walk animations. Apparently a recognizable gang walk is a thing.
The team kept growing as we hired more people. On top of this, the Manhunt team joined halfway through. That was a massive help. At the end of SA we had about 80 developers plus the test department in Edinburgh. New York also had testers and production people.
For most of the time development was manageable. Certainly less stressful than Vice City. To finish it off a good stretch of crunch was required. The company got a chef to provide a good meal for those working late. That really helped.
On the day of release, stores opened at midnight. It was a cold, rainy night but I went to have a look anyway at one of the big stores in Edinburgh. (Virgin on Princess Street). I didn't expect it to be too busy but there was a line of 50 meters or so. I bumped into 2 artists from the office doing the same thing. We were standing there looking at all these people being so excited. People got their game and hurried straight home to play the game we'd made, all through the night. It was magical.
That's the thing about making games. You're just working away in an office and it's hard to gage the excitement of players out there. Especially before the internet, you just didn't have that feedback.