Feasibility of an upgradeable or forwards compatible console *spawn*

[Shifty Geezer]

In that case it's rational but the market is flooded with expensive (and inexpensive) tablets that dont double very well as labtops ...

Debatable. Plenty of stands and keyboards exist to make tablets into all sorts of configurations. Furthermore the functionality of the laptop can often be served by a tablet without it having to be a laptop. There are lots of reasons to own a tablet. I can agree with your sentiment regards people upgrading to tablets than they perhaps don't need because they already have a functional device, as a new shiny, smaller tablet adds non-functional value to the device, but that's far from the full picture

 

[wco81]

The other thing is tablets are a personal device, something that you're likely to use more during the day and on a daily basis.

Not always true of consoles.

And cheap or free useful content, vs games that cost up to $60 for 10-20 hours of use, if that, for consoles and games.

But it's true about tablet saturation. iPad growth may have leveled off but they may still sell more $500 tablets a year than $400 consoles.

 

[Nesh]

Debatable. Plenty of stands and keyboards exist to make tablets into all sorts of configurations. Furthermore the functionality of the laptop can often be served by a tablet without it having to be a laptop. There are lots of reasons to own a tablet. I can agree with your sentiment regards people upgrading to tablets than they perhaps don't need because they already have a functional device, as a new shiny, smaller tablet adds non-functional value to the device, but that's far from the full picture

Yes there is an overlap between labtop and tablet of some functions. And yes there are many configurations. That is suggestive of the wider possibilities and various consumer profiles but not suggestive of the usage volume of the different configurations or how well they function compared to the complete compatibility of the laptop

In addition I think you missed the perspective I pointed to which is related to the purchase and upgrade of the $500 tablet. The most commonly used functions that make tablets appealing are found across a wide price range though, they are not exclusive to the high end expensive tablet mentioned earlier which someone suggested people are upgrading at a faster rate.

My phone and my tablet are far from being high end, they have reduced my laptop usage, but they are far from doubling as laptops. But most importantly most things most people do with the expensive tablet are also common with the cheaper ones. They have become so efficient that a cheaper tablet suffices very easily

There are the exceptions ofcourse that buy the expensive tablet for other more demanding functions. But these people dont need to upgrade their higher end tablet every year unless we move deeper to other exceptions

 

[Squilliam]

I think this is the last console generation if we use the definition from previous generations. I think that it is simply too costly to launch entirely new hardware with the same dilemma of expensive hardware + expensive games with no install base. Instead it makes more sense to have an iterative hardware design, so I think that is where the Xbox One is heading. I would not be surprised if in 2015 they release a die shrunk Xbox One that is ~50% faster than the current generation of hardware.

 

[Nesh]

I think this is the last console generation if we use the definition from previous generations. I think that it is simply too costly to launch entirely new hardware with the same dilemma of expensive hardware + expensive games with no install base. Instead it makes more sense to have an iterative hardware design, so I think that is where the Xbox One is heading. I would not be surprised if in 2015 they release a die shrunk Xbox One that is ~50% faster than the current generation of hardware.

I dont think they will need or want to make it perform better in a year. For marketing purposes it will make the communication of the product's capabilities a bit more complex than necessary since the performance wont be uniform across customers.One of the benefits of consoles is that they are stable, fixed products where everyone can enjoy the same experience. Consoles mature in software with minimum hardware changes in terms of what the customer can get from game performance.
Instead if they go that route they will be ripping the cost benefits of the die shrink and launch with "minimum" R&D cost in 4-6 years time

 

[Squilliam]

I dont think they will need or want to make it perform better in a year. For marketing purposes it will make the communication of the product's capabilities a bit more complex than necessary since the performance wont be uniform across customers.One of the benefits of consoles is that they are stable, fixed products where everyone can enjoy the same experience. Consoles mature in software with minimum hardware changes in terms of what the customer can get from game performance.
Instead if they go that route they will be ripping the cost benefits of the die shrink and launch with "minimum" R&D cost in 4-6 years time

There are two major problems with the current strategy:

1. The console hardware gets stale.
2. Renewing the cycle is expensive.
3. Anticipating the future is problematic, and they can get blindsided without adequate means to respond to moves by their competitors.

If they move to a 2 yearly cycle then they can solve the above problems. If they release an updated console then they have the opportunity to solve the complaint that their hardware is weak by comparison to the PS4, and they get the opportunity to release new features to increase the market for their console. Why should they leave the performance advantages on the table when they move to 20nm finfet?

 

[RudeCurve]

People need to remember they don't need to stop selling the older model. They could continue to sell it at a lower price. That is the whole point. The old console could still play newly released games just not with all the bells and whistles that the new console provides.

 

[AlNom]

People need to remember they don't need to stop selling the older model. They could continue to sell it at a lower price. That is the whole point. The old console could still play newly released games just not with all the bells and whistles that the new console provides.

On the one hand, you'd think flipping a few switches would be easy ala PC - essentially low to ultra presets. On the other, perhaps it's easy to ignore PC drivers kinda fixing a lot of crap behind the scenes - QA problem.

Or something like that.

 

[blakjedi]

I think in order for that strategy to work, Ms would have to provide upgrade paths for their customer base. I still have both an iPad 1 and a surface rt. I use both.. (surface more than the iPad) I will probably get a surface pro 3 this year. With universal apps apps will work on both rt and pro... which is nice.

Ms would do well to try to apply/extend the universal app strategy to Xbox one and its successors. All Xbox apps work on all xboxes moving forward until a technological break forces incompatibility.

Damn that was a lot of blubbering.

Tl;dr Xbox one is the new base of all Xbox software going forward. All new Xbox branded console games have Xbox one as the minimum spec.

 

[dobwal]

In that case it's rational but the market is flooded with expensive (and inexpensive) tablets that dont double very well as labtops and arent called SurfacePro. SurfacePro is a drop in the ocean. Its market performance did not meet MS's expectations either which again shows that people dont buy much to that utility but look for other gains that come with a tablet, they dont think like you.



Those references were made for the $500 tablet and later the iPad.

People buy tablets for the same reason people buy sandals in the summer time. It offers the necessary utility and functionality for lightweight usage. Steel toe work boots offer a ton of protection and function. Doesn't mean people want to wear them on a beach while on vacation.

A keyboard and a mouse is basically overkill where a touch screen, a tabs simple form factor and UI is more than sufficient for light use that doesnt require a bunch typing and heavy ui interaction.

That being said the iPad was the first successful tab and it benefits from brand loyalty and no need to reinvest in ones app library. Hence why it's able to continue to sell relative well at $500. My issue with surface is why buy it when I can buy a high price tab and a decent laptop for the price of a new surface pro.

I agree that console don't need an upgrade schedule akin to tabs. I think the reason the growth rate of the ipad has stopped is because current tabs offer performance and functionality at a level that's more than sufficient for their usage. I still have my iPad 1 and it chokes on some websites or aggregate news apps, but mobile software requirements don't increase at a level that requires yearly updates.

Consoles don't make any real money at initial price points. Mass adoption doesn't happen until consoles prices are fairly cheap. Accelerating the upgrade schedule makes getting to lower price points more difficult.

 

[steveOrino]

There are two major problems with the current strategy:

1. The console hardware gets stale.
2. Renewing the cycle is expensive.
3. Anticipating the future is problematic, and they can get blindsided without adequate means to respond to moves by their competitors.

If they move to a 2 yearly cycle then they can solve the above problems. If they release an updated console then they have the opportunity to solve the complaint that their hardware is weak by comparison to the PS4, and they get the opportunity to release new features to increase the market for their console. Why should they leave the performance advantages on the table when they move to 20nm finfet?

Steambox is already going to do this without the licensing overhead, limited architecture, and small install base.

 

[Rurouni]

One of the biggest problem in doing a shorter cycle is making profit from the console hardware. Right now the console hardware value is relatively high because the platform owner either subsidize it, no margin, or tiny profit. If the cycle is shorter, making profit from the hardware is much harder, thus my guess is that the console will be sold with a higher margin thus more expensive for consumer. Now you've got more expensive hardware and shorter cycle. If that's the case, then might as well going the PC route. Sure the console might have the advantage on hardware optimization, but for me personally, if the cycle is shorter, might as well buy a PC (which probably would still cost more than the expensive console) but you can just upgrade the CPU/GPU to reach next gen, while on console you'll need to buy a new hardware. It may lack the optimization, but with DX12 and the like, brute forcing the performance would be cheaper.

 

[RudeCurve]

I've already explained the solution but nobody listens!

In general new consoles with more powerful hardware will always cost more than old consoles...therefore if MS releases a XBO V2 in say 2016 it will cost between $400-$500. At that point the "old" XBO V1 will have already come down in price to $300.

Nobody "needs" to upgrade if they don't want to...obviously...

It's an optional upgrade just like how I skipped the Iphone 5 and am waiting for the Iphone 6...no harm done.

 

[Squilliam]

I've already explained the solution but nobody listens!

In general new consoles with more powerful hardware will always cost more than old consoles...therefore if MS releases a XBO V2 in say 2016 it will cost between $400-$500. At that point the "old" XBO V1 will have already come down in price to $300.

Nobody "needs" to upgrade if they don't want to...obviously...

It's an optional upgrade just like how I skipped the Iphone 5 and am waiting for the Iphone 6...no harm done.

I think Intel's tick/tock cadence would work quite well. In 2015 they can release a slightly upgraded Xbone slim for $299 with say 20% faster CPU/GPU clocks, and enable the fused off part of the GPU with say 8 DDR4 8Gbit chips with greater bandwidth. Not huge changes, but it should be enough to give some of the hardcore people a reason to upgrade whilst at the same time giving them a reasonable competitive advantage against their competitors. I think this is doable on the 20nm process node. Then in 2017 they can release their new 'tick' console with upgraded CPU architecture, more esram, more ram (say double) and an upgraded GPU with the same relative cost as the previous 'tock' using the 16nm finfet process.

I think that they could improve performance by say 25% and then 60% which would give them double the performance overall in 4 years. It would also enable them to have new features in both software and hardware, and it would keep the console fresh so that it stays relevant both for current console owners and for new entrants who want the best deal they can buy.

 

[Arwin]

But you lose performance first, due to the overhead of supporting meaningful differences in hardware?

 

[Shifty Geezer]

But you lose performance first, due to the overhead of supporting meaningful differences in hardware?

I think those losses already exist in most games, especially if the console is running a degree of VM abstraction. We've been told Sony lets devs hit the metal harder, suggesting HW is already distanced from devs on XB1. Throw in middleware and there's already lots of potential lost to make development and porting easier.

As came up before, you lose say 10% performance from the hardware in order to gain portability of games across generations. Overall I'd say that was a benefit in the same way it is on other devices.

 

[JPT]

I have always wonder, middleware, does it abstract a lot? I mean it seems everything is middleware these days, but when I read dev interviews, it looks more like what they call middleware are libraries you include in your code.

To me middleware is UE, Unit and Cryo engines, but there is always lot of talk about physics middleware and sound and etc.

 

[Shifty Geezer]

Middleware covers everything from full on Engines to libraries.

 

[joker454]

What you lose in abstraction you gain in development familiarity. Because the new machines are of similar nature and architecture means there is no long period of being stuck in a learning curve like with traditional consoles, instead developers can hit the ground running right away. For example when a new console comes out you always have to wait years for developers to come to grips with it, get their tools sorted out and finally start to really take advantage of the hardware. That's not the case on non console hardware, developers can make full use of the new hardware right away. So when a new pc gpu, iPad or whatever non console hardware comes out you don't hear how it will take years for coders to understand this new hardware. Instead new stuff supports it right away, more so because backward compatibility means all your old games run better and suddenly all those 60fps console games are actually now 60fps rather than 45fps. The idea that hardware abstraction causes loss of performance is a myth in the real world once you take development needs into account. By the time developers "come to grips" with typical console hardware you would have been able to release multiple versions of hardware abstracted bc hardware to where any technical advantage of the old world console hardware is totally lost.

 

[Shifty Geezer]

So when a new pc gpu, iPad or whatever non console hardware comes out you don't hear how it will take years for coders to understand this new hardware. Instead new stuff supports it right away.

That's not completely true. Old techniques are used until the new hardware has a large enough install base to worth targeting directly. Ergo, your new hardware runs the old code just fine, and devs can make new content for it, but it's not used to anything like its full potential for some time after release, as the lowest common denominator is last gen HW. That's not true of traditional consoles that, as closed boxes, also get targeted as a complete unit.

By the time developers "come to grips" with typical console hardware you would have been able to release multiple versions of hardware abstracted bc hardware to where any technical advantage of the old world console hardware is totally lost.

Except that comes at added cost buying new hardware. In the traditional console, you certainly get your money's worth when every ounce of performance is squeezed out of it. With progressive hardwares, improvements come more with buying new hardware than devs advancing the software, although of course software developments do help.

Hypothetically, let's imagine MS released a BC console. Moving to compute based rendering would probably be slowed down versus releasing a completely new hardware with no legacy ties.

Now if continuing consoles are fully BC and get refreshes every two years, as new techniques are developed (like tessellation was) they won't be fully implemented on the new hardware as the old hardware is the primary target. This is the problem with PC and mobiles - new hardware with new techniques goes unused. That's where console's greater hardware utilisation is a big win.

Obviously the change in the market and cross-platform middlewares means new boxes aren't viewed in isolation, and personally I think the move to abstracted hardware makes more sense than the fixed box paradigm. And as a result, getting better results warrants smaller, more frequent hardware upgrades rather than the 5+ year huge leap. But it'd be wrong to underplay the benefit of targeted boxes versus an ever moving target. The latest, greatest hardware in these case will never be ideally used. It'll basically be current gen +1, and it'll only play the new, exciting technologies once it's become worn in and a little slow by the latest standards.