Actually they wouldn't have because there was no way that 80x86 in 2009 could have run PowerPC code of just a few years before. That's an inherent limitation of generationless, you're more or less locked into evolutionary technologies and not revolutionary technologies that are fundamentally incompatible for your launch timetable.
You'll note that in the post you responded to, I had specifically said that for this to include PS3/X360, they'd have had to make a change to x86 at the time. That would have been a good thing as the IBM PC was nowhere near as capable as x86 CPUs of the time. And the Cell processor while very power at specific tasks was very difficult to use well.
But to both of your points, I repeat (with some minor edits) a
subsequent post I made that epitomises the bleak situation gamers will face in a generation less system and how having to support the lower tier performance envelope will hold us back.
With more and more processing tasks moving to GPUs from CPUs it's advances in GPUs that are most likely going to impact (or limit, given their absence) complex gameplay systems. Scorpio could throw 1.8Tf (the entirely of PS4's theoretical maximum GPU performance) at physics calculations, AI or other complex systems, and still throw more than a PS4 Pro's worth of performance at graphics. But that game would never run on Xbox One. So that won't happen until Xbox One is mothballed in 4-5 Years time, even though Scorpio will be capable of it at launch. As a gamer, that sounds a shitty scenario.
If Microsoft propose having short generationless cycles of performance tiers then consumers will need some assurances that their hardware is still going to get that traditional 5+ years of support and that support needs to be
mandated by Microsoft. Otherwise nobody is going to buy into a generation two years after launch for 1-2 year of support, ending when the next console arrives. If the console manufacture guarantees such support they inherently limit groundbreaking games that could only run on the new hardware.
Fundamentally it's taking the decision about when to drop support for older console tech away from the developer and publisher. It's worth noting that Sony's current PS4 Pro development policy of "no additional features" on Pro would have prevented EA from publishing Shadow of Mordor on PS4 with the Nemesis antagonist system because PS3 and 360 couldn't cope with this.
This issue is a problem and dilemma and somebody is going to lose and I think the people who will lose are those people who like to buy new technology and see it used ASAP. Their games will be hamstrung by the older version of the console. You'll only ever see your console used to it's fullest years after it launches and not because of a learning curve but because of an arbitrary publication policy. Generationless sounds great in theory and the idea of taking my games to my next console is great, but blurring the lines between generations creates a lot of problems.
First, to address the bolded. We already have games that run at settings (both graphically and WRT to phsyics) that scale from hardware more powerful than Project Scorpio to hardware that is less powerful than XBO. Games, have for years included options that enable or disable both GPU intensive rendering options and CPU intensive options (level of physics simulation, number of AI entities, etc.). Games for years have also included toggles for settings that use technology that is incompatible with some hardware (tessellation, CUDA effects, lots of proprietary NVidia effects, etc.). These games also exist on consoles, but those options aren't exposed on consoles.
So, sure, coming from a console viewpoint what you said makes sense.
Coming from PC, there is no historical precedent for that. Especially if a developer was either exclusive to PC or developed with PC as a primary development platform. You can take advantage of new technology advances WRT GPU rendering without also dropping support for hardware that doesn't support those GPU rendering features.
Yes, some developer's voluntarily limited their engine in order to reach the widest range of hardware possible. Blizzard is famous for deliberately not using the latest advances in technology. OTOH, companies like CD Projeckt Red (prior to starting console development) did not. In fact, prior to the Witcher 3, they had always released a technology patch ~1 or so years after the release of their games specifically to take advantage of newer technology advances that did not exist when they were developing their games (Witcher 1 and 2). Crytek (yes, them again), always targeted technology that was the lastest they could.
You saw things like that with the Dx8 -> Dx9 transition which was represented the largest divergence in how 3D rendering was accomplished on both a hardware and software level. Within the first year of the introduction of Dx9 you saw developers attempting to take advantage of it using hardware capabilities that did not exist on Dx8.1 hardware. Within 2 years you started to see engines which started development either when the hardware was released or prior to the hardware being released focused on Dx9 as the primary rendering method but including a Dx8 rendering path.
Dx10.1 saw implementation within the first year by some developer's despite less than 5% of the market having Dx10.1 cards.
What's most interesting is that from a development standpoint, the closest thing to a console like generational break in the PC world, also results in lower adoption of newer hardware features and software APIs within games. Dx10 is the poster child for this. Dx10 required Vista, meaning it completely broken compatibility the previous generation. Vista adoption was low, again similar to the start of a new console generation. Dx10 ended up largely ignored. Dx12 is facing a similar barrier in that it is limited to Windows 10 which on Steam currently stands at ~51% of the Windows gaming market install base. I'd say that's roughly analogous to where we stand with active users of PS3/X360 to PS4/XBO. If Dx10 (Dx11 to a lesser extent) and now Dx12 didn't face those barriers, you'd have seen them rapidly become the primary rendering API that games engines are based on, with the ability to downscale rendering for lower API levels. Interestingly Vulkan may change all this as it supports prior generations of Windows.
Here's the thing though. Console users have been indoctrinated to think that the breaking of generations is the best thing because that's all they know. Developers don't like it. Publishers don't like it. PC users in general don't understand it (OK, I can't enable X feature, but the game still runs).
Just to reiterate. The introduction of Dx9 represented a far larger change in hardware GPU capabilities and software development API that is unlikely to exist with the idea of a rolling generation of console hardware. And even that didn't stop Dx8 hardware from being supported for years after Dx9 hardware came into being. It also didn't prevent developers from pushing Dx9 as far as they could. The original FarCry is a great example of a game that pushes Dx9 rendering and hardware really hard, yet still ran on Dx8.1 hardware. Sure, the best results weren't seen until years after Dx9 came into play, but that isn't much different than how it took years for PS3 games to come to grasp with the new architecture.
Interestingly enough that time to come to grasp with the new consoles was greatly facilitated by the fact that the hardware was already relatively well understood in the PC space. That concept of each generation already being somewhat familiar due to similarities with the previous generation means it'll be much easier to take advantage of the hardware more fully in a smaller amount of time, results in games that don't need 5+ years of the console being in the market until your knowledge of the hardware and its quirks are finally good enough that you can start to fully exploit the capabilities of the device.
Are some developer's not going to take advantage of a new technology advances in a "rolling generation?" Sure. But that's no different from how it is with the current generation system.
The larger question is whether developers that currently attempt to make the most out of technology advances will still continue to do the same within a framework of a rolling generation system. And coming from a PC background, it seems highly unlikely that they would stop trying to take advantage of the latest advances in hardware.
Heck, we already see one case of that with Doom 3 on PC. Choosing Vulkan over Dx12 due to the ability for it to support prior generations rather than only supporting the current generation. Vulkan also being used so they can take advantage of advances in technology that Dx11 and OGL don't support or don't support well. Choosing a path that simultaneously allows them to push the hardware WRT to the latest advances in GPU rendering technology while simultaneously supporting hardware that doesn't have the latest rendering technology.
The only way I see something like this not happening is if, as you point out with PS4-P, the console manufacturer mandates that you are not allowed to take advantage of newer rendering technologies.
Which is drastically different from mandating that you must support one generation of hardware prior to the current one.
TL: DR - Rendering power differentials and GPU rendering capabilities that are far more divergent that what exists between XBO and Project Scorpio don't prevent game developers from including things in their game that are capable of running on the highest end hardware but not capable of running on the low end hardware at those settings.
Regards,
SB
