why did you use fairly low-polygon models and low-resoultion textures? Engine limitations, system requirements, artistic direction or limited resources? I'm not sure how much extra work it would be to make higher-polygon models and hi-res textures.
Yes to all but the artistic direction side of things. Although there's part of that too. I wanted to make sure older computers (up to 5 years old) could play the game, first off. And the engine is... well, it's a frankenengine build upon 10-year old technology. There are still comments in the code referring to where they used it for the game Tribes. That was, what, 1999? Anyway, it's not a stellar performer. More modern engines have better schemes for things like texture and object management, scene management, etc... but they also get to assume more modern systems (which is why, if you have to dial the detail down on the newest games to get them to run on your machine... something I do a lot... they actually look crappier than older games)
Another problem - and this is something of considerable note with modern games: What works well when you've got only one enemy on the screen can suddenly go to crap when you've got six or more high-LOD (level of detail) enemies packed together in front of you. Modern games tend to be very aggressive with this, preventing too many high-detail enemies from being visible at once. They use all kinds of tricks to fake it, so you probably don't notice it. I'm still in awe of how they made Left4Dead work... they did some really cool things with that one to be able to display zombie hordes.
But that's another reason why shooters are so popular... you are at a distance with your opponents and so they can use a lower level-of-detail, imposters, etc. That's ALSO why you tend to get worlds that are much more linear in modern games - it's is a technical constraint affecting design, but it allows them to really aggressively control their scene management and therefore show the highest-quality models they can.
Anyway, lots of answers, most of them technology-based, but the final answer is: That's what worked. We did a bit of experimenting and kinda-sorta found out what worked. That's why we're hoping things will go faster with the sequel, if we don't do something crazy and change engines... we know what works in advance and how to avoid the hundreds of pitfalls we encountered the first time around.