Groucho said:
theword said:
Groucho said: ^^ Wait... sorry.. the Unreal Engine has been in development for less than 4 years? This is news to me.
The man-years excuse is silly, especially since you can spin it to favor whomever you choose. I guess at least the usually blind 360 fanbase can at least appreciate KZ2's excellent visuals. |
The man-years reference is relevant because man-years quantifies the amount of work and is definitely not the same as the number of years a game was in development. In the case of KZ2, a team of over 100 people put in 4 years of work translated to well over 400 man-years. In the case of Gears 1 a team of roughly 50 people put in about 2 years of work resulted in about 100 man-years of effort. Epics also said that they invested about 30 man-years (with a team of no more than 18 persons working on the multiplatform engine at any one time) into Unreal Engine 3 at mid point of Gears 1 development. Consider that UE3 was a 3 platforms effort, the 360 portion of the engine was at most 15 man-years at the time.
115 man-years for Gears 1 vs. 400+ man-years for KZ2 is a big difference. One could say that it took nearly 4 times the amount of work to make a game that looked marginally looked better than Gears 1 in terms of graphics.
Keep in mind that all this time, KZ2 developers had focused mainly on the PS3, while the EPIC team (smaller) had to spread their man-hours among three platforms.
All numbers I mentioned above were publicised in developers interviews and press conference.
|
You sir, are grossly underestimating the work which goes into a game engine, which has nothing to do with the platform, and the power of the X360's proximity to PC development when you're talking about "crossplatform". Killzone 2 was likely rewritten from the ground up for the PS3 (as opposed to being based upon Killzone 1) -- Unreal was almost assuredly not rewritten (not even close) for the 360, simply because the 360's architecture is so much closer to the PC, from the game engine persepctive. There is a downright titanic savings in not having to rewrite your tools pipeline alone, let alone all the mechanisms which comprise a game under the hood. Implying (as you are) that it took 4 times as much manpower to get the visuals of KZ2 to be superior to GeoW2 is absolutely ludicrous.
A significant portion of the GeoW2 team was probably working on improving visuals specifically for the 360. I wouldn't be surprised if Guerilla's engineering (and art) depts actually spent less time on their visuals, than the detailed texture/shader work took in GeoW2. KZ2's visuals are better because the platform allows for a rendering technique that greatly enhances the ability of artists to use lighting, amongst other things -- not because the shaders were better, not because the textures were better, etc. etc. In other words, it probably took *less* time to get KZ2 to look better than GeoW2, not more -- thanks, in part, to the hardware architecture they were working with. The rest of the team was almost unquestionably devoted to developing the single player game (scripting, physics, object management and actions), the networking and multiplayer components, the toolchain (a downright huge amount of work without middleware), etc. -- all of which the GeoW2 team likely leveraged from previous iterations of the Unreal Engine easily.
On top of all that, Guerilla produced Killzone Liberation (PSP) during that timespan as well -- and I'm pretty certain they did not use the same PS3 engine/tools/etc. They likely leveraged the PS2 Killzone engine for it, but it would have still required a decent amount of reworking.
In an entertainingly ironic twist, your implication that overall manpower is the reason that KZ2's visuals are generally considered (by reviewers) to be superior to games like GeoW2, suggests that you can always get a linear work speedup by increasing the number of workers on a project. As if the workflow of a large engineering project were... embarassingly parallel. You are, in an indirect manner, suggesting that, since the PS3 has a significant amount more parallel horsepower than the 360, it will inevitably be more powerful in any application -- when in fact, the parallelism comes in handy at certain points in the game engine pipeline, and not in others, which is why you get "we're only using 60% of the available power" comments from developers. Back to game development, there is a very large iterative, effectively serial, component, when it comes to toolchain development (and other elements, but tools is a great example), and the Unreal engine has had eons to develop in this area. Developing a new toolchain, in parallel with a game engine, requires a *lot* more horsepower-over-time than you would need given many years of slower development.
As I said, the your "man-years" line of reasoning is totally backwards. Any professional game engineer in the console industry can, and would, tell you so. Why on earth do you think middleware (like the entire Unreal engine, which is extremely advanced after long years of development, especially with regards to tools) is so popular these days? Think its mostly just a wash, buying (really expensive) middleware licenses, as opposed to doing the work yourself?
|