Again, that isn't HARDWARE limiting GAME PLAY.
That's a developer choice to limit game play in order to achieve a certain level of graphical fidelity. As other's have stated there was nothing preventing them from making HZD with flying by reducing graphical detail a bit.
Again, hardware (in the vast majority of cases) limits how much graphical fidelity you can have, it doesn't really limit the type of game play you can have.
There are, of course, exceptions. Claybook has a certain level of physics modeling that directly impacts the gameplay. But even in that case Sebbbi was able to go from thinking it required a certain level of hardware at the start (not possible to port it to NSW) and then later discovering with some work he could greatly reduce the hardware requirements (so it could run on NSW) for the exact same game.
And even then, you could still have a game with the same gameplay with but much lower granularity (larger physical particles for physics simulation). The visuals would obviously suffer, but that's graphics and not gameplay. Gameplay might be rougher operating on larger particles/voxels but it would still be the same game play.
Or to put it as I've been saying for a while now WRT the next gen consoles.
- I doubt we'll see new forms of game play that don't already exist or couldn't be done on current gen hardware.
- What we'll see is greater graphical fidelity (this includes physics in most cases) or more convenience in games.
- AAA developers may or may not take this opportunity to incorporate game play elements that INDIE developers are already doing because graphics are secondary to gameplay for Indie developers. As opposed to AAA developers where game play is secondary to graphics (HZD, for example).
- And just so people don't think I'm saying HZD game play isn't good, it is. But as you said above the HZD developers obviously limited game play (flying) in order to have better graphics.
Regards,
SB
The big question is... When the game is still on paper, you cannot just go wild on what you intend to do. You must look at your hardware specs and make an educated guess, based on some testing, on what you can and what you can´t use. Be it physics, IA, streaming, etc, etc.
At this point if a hardware level doesn´t appear to be capable of doing such a thing, you cannot go on, hoping to be able to manage to optimize stuff, and make it run ok later on. You need to be on the safe side, and this will affect game design choices.
What we are talking about next gen consoles, is for them to take games to a next level... we are talking about next gen improving on current gen quality. Not about having current gen reducing detail to be able to run next gen improvements.
What this means is that, on cross gen games, and, for now, speaking just about streaming capabilities, we cannot hope to create a world with a tremendous high detail using the streaming capacities of the new gen SSD, and then create a bloddy pixel mess on current gen HDDs to make it also work. This would be ridiculous! It would reduce current gen level of quality, and it would make the game suffer to a terrible degree.
What needs to be made is to create the maximum level of quality current gen HDDs supports, and thet improve on it for new gen SSDs. This would create a high quality game on both gens, with the new gen showing benefits.
But althoufh clipping, view distance, and even detail could increase, some legacy would be present. And the biggest one would be movement speed. It would be the lower specs that would define it.
This is true for everything else... You cannot create a physics model that current gen does not support, and then use a simplified model on current gen. This would not be the same game. If your game requires the advanced physics, you just cannot port it to the current gen. You could create a new game with the same name, add the "Lite" extention or something to the name, and then use a simplified model. But not on the same game! It would be ludicrous, since the nuclear part of the game is not the same!
A good example is IA, you cannot use an excelent IA model, tht requires high CPU usage, and then stupidifie it for current gen. How ridiculous would this be?
It must be the current gen who defines the IA level.