The only people who consistently talk about 360's 'frailties' etc.. are people who don't know what they are talking about. Quite amusing tbh.
Environments in file size are actually a lot smaller than you would expect, a fraction of a DVD. A large environment is a simple collection of polygons, obviously the more complex geometry requires more polygons and results in a higher polygon count, however as a percentage it's almost unnoticable for the vast majority of games when compared to textures/audio/video.
Textures take a lot more space, but you can only go so big with textures (1024x1024) due to system limitations, not file size - each texture takes up a chunk of system memory, more textures, more memory. When games have to run a Diffuse Map, Normal Map, Bump Map, Specular Map and very soon complex Parallax Mapping, on top of this sprites and Particle effects. People wonder why current games suffer slowdown or run at 30fps it's because running all these effects murder the systems memory, run all these at the same time and it doesn't matter how big your disc is, it simply can't do it. You could fill 30gb of a BD-Rom up with textures quite easily, you just simply couldn't use them.
Sound takes up another good chunk of disc, although this is mainly taken up by music, sound effects aren't too bad at all coming in at minute file sizes to save on memory for repeated sounds (think engine roar or gun-fire). Audio that is consistently heard is compressed as much as possible (before obvious degredation), voice overs require more space - a reason not all games come with lots of voice talent (something Oblivion doesn't get enough recognition for). Music is the biggest hitter, games like Gears of War or Resistance: Fall of Man get away with set-piece music allowing for a cinematic feel, the placement of audio is placed specifically into the game whereas Grand Theft Auto or Project Gotham Racing have music that is freely choosable at any time and as a result more variety required - more variety means more space.
Video is a massive hitter, often incredibly compressed, WWE Raw vs Smackdown 2007 is the best integration of video into a game I have ever seen by using compressed 30 second looped video footage for character introductions. In the past many games have been guilty of using Full Motion Video (FMV) too much, Final Fantasy VII and Parasite Eve come to mind, FFVII is well known to have fit on 1 single CD, only reason it had 3 was to fit all the FMV's onto disc.
The only area that could be causing problems for developers of Grand Theft Auto IV is video or audio, as I don't expect any video the thing that rings alarm bells is fitting as much audio as possible into the game, notably music. Looking at how much music has been in past games, starting with GTA3 which had a variety of music, GTA:VC went further adding a selection of well known 80's tracks, GTA:SA had a massive selection of music. I would guess they (Rockstar) want to add even more music into GTA:IV than they put into GTA:SA.
Just whinging about disc space is ridiculous, you want to be aware of what is and what isn't possible before complaining and jumping on a bandwagon. Remember, through clever coding from good programmers almost anything is possible and we are just touching on high quality compression and streaming techniques, DVD9 is far from over, the only companies complaining are those without the forsight or ability to get the best from a format (and the same stands for those complaining about Playstation 3 btw). Space doesn't mean it can and will be better, being able to store a million files and utilising a million files are worlds apart.