*spin-off* Importance of Backward Compatibility Discussion

[ShadowRunner]

I don't built up a nice library of XBLA games that I'd be sad to see go.... would consider getting a BC peripheral box though if that was an option.

Would it cause you to jump ship though for that reason alone, considering you wont get BC for your XBLA games from Sony either? I guess this is the question the companies are asking themselves. Having BC could be a factor in preventing people jumping ship if a competitors box is more desirable, but lack of BC wont cause the a jump. Id also suspect those that have a big desire for BC are the more loyal fans of the systems who will be invested in certain exclusive franchises already, if even these people need a lot of encouragement not to jump ship then that system has much bigger problems than lack of BC. In short relying on BC to keep your userbase would be like applying a bandage to a gunshot wound.

Its a nice feature but people place way too much importance on it in.

 

[upnorthsox]

It's a good idea, but only a pittance of the PS3 library would ever get ported over, with many good, mostly less-successful games left out in the cold.

Seeing how long it took to update Ico, Shadow of the Colossus and the previous God of War installments I wouldn't hold my breath for even a handful of new updated games. Wth all the special sauce hoops you need to jump through to get code to run smoothly on cell you'd need to rewrite large tracts of game code to make it run on a regular CPU, which takes time and costs money. Priority for that is probably going to be low.

BC is good for it lets you start running games on your system right away. It'd be no good if you had to wait months, possibly upwards of a year after launch while games are getting ported, Sony would get so much criticism it'd almost be no point in claiming BC at all.

This.

It's why SPUs couldn't have been included in the ps4 design even if they were wanted.

Push heavily for multiplaform releases (ps3 and ps4) of new and recently released games and offer an add-on type device for BC.

DD only game devs will have to learn on their own the importance of BC to their customers and support accordingly.

 

[joker454]

Would it cause you to jump ship though for that reason alone, considering you wont get BC for your XBLA games from Sony either? I guess this is the question the companies are asking themselves.

See, I'd ask it a bit differently. Would it cause you to pick one platform over the other (for a new buyer) because one has b/c whereas the other doesn't? I mean it worked great for Colecovision back in the day, it was a new machine with the usual tiny library a launch machine has except that it could also play all Atari 2600 games. For someone evaluating the tiny libraries of the new machines, if one can play a large chunk of the existing XBLive/PSN library do you think that would sway them to buy that machine instead of the other?

 

[xexuxjy]

No, probably not , though it would probably delay any purchase decision for quite a while , if I knew my exisiting library would work then I'd pick one up early as a 'ooh shiny new toy' thing, otherwise there would have to be an exceptional launch game (like Kameo was for me on the 360) to pick one up at the start (and I guess that launch game could be on either of the platforms...)

 

[patsu]

Sony should know the benefits of B/C. The PSP library is incredibly useful for Vita, 'specially as digital games. I think if they could, Sony would deliver B/C in PS4, or repurpose PS3 into other devices.

 

[BadTB25]

See, I'd ask it a bit differently. Would it cause you to pick one platform over the other (for a new buyer) because one has b/c whereas the other doesn't? I mean it worked great for Colecovision back in the day, it was a new machine with the usual tiny library a launch machine has except that it could also play all Atari 2600 games. For someone evaluating the tiny libraries of the new machines, if one can play a large chunk of the existing XBLive/PSN library do you think that would sway them to buy that machine instead of the other?

If performance wise they were more or less equal, BC would be big for me. Truthfully, I have way more X360 games compared to about 20 or so PS3 games, so it would be more meaningful for me if the NextBox had BC.

I will more than likely pick up both eventually.

 

[ShadowRunner]

Wouldnt you most likely stick with the next xbox anyway though even without BC if they were pretty much equal?

Conversely if PS4 had BC and next xbox didnt would the content available on PSN currently make the PS4 your preference?

Personally i will be looking at the new games coming out for the next gen consoles at launch and into the future, and will most likely want to pick up the console with the exclusive franchises im most interested in.

I can see the appeal in the back catalog of either machine to those who havent owned a current gen console, but i suspect most people that pick up the new consoles within the first year would have.

 

[BadTB25]

Wouldnt you most likely stick with the next xbox anyway though even without BC if they were pretty much equal?

Conversely if PS4 had BC and next xbox didnt would the content available on PSN currently make the PS4 your preference?

- If the PS4 had BC and the X-Next didn't, I'd get a PS4 first.
- If they both have it, I'd get an X-Next.
- If neither had it and they were more or less equal, I would still likely get the X-Next first, but could be swayed by what launch games are of interest of me.

 

[Shifty Geezer]

See, I'd ask it a bit differently. Would it cause you to pick one platform over the other (for a new buyer) because one has b/c whereas the other doesn't? I mean it worked great for Colecovision back in the day, it was a new machine with the usual tiny library a launch machine has except that it could also play all Atari 2600 games. For someone evaluating the tiny libraries of the new machines, if one can play a large chunk of the existing XBLive/PSN library do you think that would sway them to buy that machine instead of the other?

For new consumers, I think the response to that scenario would be irrelevant. No-one buy s a new, expensive console except to play new games, so all the other features - price, performance, services, fancy controls - are what people are basing their purchasing decision on. "Oh, it plays some old games too," is typically a last thought in the decision making tree. Although it may have some influence of a tiny proportion of potential buyers, BC comes at such a cost the economy of its inclusion as standard in every box just isn't there. Make it an optional extra at cost to those who want BC, and everyone's served in most economically viable fashion.

 

[TheChefO]

So ... Back when this whole discussion kicked off, the argument was framed in the context of what would have to be sacrificed in order to maintain BC by sticking with old architecture ...

Now that we know what's in the new boxes, I'd say we would have been much better off with Cell2.0 and a 12-core xenon.

Performance wise, these new jaguar cpus wouldn't even be able to emulate the old dogs ...

Broken compatibility for slightly better(?) performance seems rather foolish.

 

[function]

A Jaguar core is likely to be a lot faster than a Xenon core though (something like 3+ times going by what I've read here), and at lower power consumption than a Xenon core on the equivalent process.

Losing BC will be a bit disappointing, but the cost of staying with Xenon was far too high. I was expecting a new and much more sophisticated Power PC core if MS were going to go for full BC.

 

[TheChefO]

A Jaguar core is likely to be a lot faster than a Xenon core though (something like 3+ times going by what I've read here), and at lower power consumption than a Xenon core on the equivalent process.

Losing BC will be a bit disappointing, but the cost of staying with Xenon was far too high. I was expecting a new and much more sophisticated Power PC core if MS were going to go for full BC.

In general terms I'm sure the jaguar will be faster, but from what I understand, there are still good chunks of gaming/graphics/physics code (not general purpose) where xenon isn't outclassed and in fact may be faster?

I'd think a smarter way about doing xb720/durango would have been a new ppc core or at the bare min (simplest) bump the cache size to avoid misses and idling... or I dunno, throw a large scratch pad cache close by with maybe some mechanism which could then intelligently move that data to and from the fast memory pool thus avoiding cache misses and idling which were the main causes of performance not hitting peak ...

Point is, yes they needed performance, but there were ways and methods to achieve higher performance without breaking BC and without breaking the bank either.

 

[3dilettante]

What are your criteria for "smart"?
Smart in the eyes of a gamer?

Why assume that a totally new PowerPC core could be designed as cheaply as using an existing x86 core as the basis?
What about how GCN is designed to be compatible with x86 memory paging, not PowerPC?

 

[TheChefO]

What are your criteria for "smart"?
Smart in the eyes of a gamer?

Why assume that a totally new PowerPC core could be designed as cheaply as using an existing x86 core as the basis?
What about how GCN is designed to be compatible with x86 memory paging, not PowerPC?

Smart as in bang for the buck and as in not pissing off the customer base by nullifying their downloaded purchases.

Some people here still bang on about "oh well, people will just repurchase the same old hd remakes yet again and the plethora of other download titles."

I contend that such an attitude toward the customer will lead to modified purchasing patterns in the future which drift AWAY from digital and toward hard media.

Another side affect in this fiasco of different architecture just for the sake of different (I'm not convinced these jaguar cpus will be head and shoulders above the old dogs), is that if BC is broken (likely) then my old catalog is not keeping me tied to the platform...

And with the pasture looking far greener in Sony land, it makes it that much easier to jump ship.

 

[3dilettante]

Smart as in bang for the buck and as in not pissing off the customer base by nullifying their downloaded purchases.

Either Microsoft hopes to mitigate this in some fashion, or it believes the backlash from the loss of compatibility is outweighed by other factors.
Or maybe IBM didn't want to play this round.

Another side affect in this fiasco of different architecture just for the sake of different (I'm not convinced these jaguar cpus will be head and shoulders above the old dogs), is that if BC is broken (likely) then my old catalog is not keeping me tied to the platform...

It is just my opinion, but I believe it is backed up by a vast amount of evidence, that a multibillion dollar company trying to usher a project worth hundreds of millions of dollars isn't changing things just to be different.
Also, there are very few metrics at this time by which Xenon isn't beaten soundly by a core seven years more recent.
One of the factors was probably that the consumer technology world has changed, and with integration, IBM's existing product range has a very limited appeal.
Either pay IBM for a totally new and untested Xenon successor and get nothing in terms of SOC elements and graphics, or pay for an SOC whose basic elements have been in the pipeline and validated by AMD.

For what it's worth, if Jaguar and GCN are the underpinnings of Durango, it's a sign of greater investment or flexibility on the part of the system designer. IBM didn't pursue Xenon or Xenon-like cores as a serious product in its own right: these were not the crown jewels of PowerPC. IBM built the chip and found it so compelling a design that it forgot about it.
If the rumors are true, Microsoft is getting a design where it knows its partner has already bet its continued existence on those technologies.

 

[trainplane]

It's all conjecture of course, but I agree that both MS and Sony went with x86 because it is already mass produced and is set really nicely for future compatibility. The problem is, if AMD goes bankrupt that could complicate things because there would be legal issues with the x86 license transfer to whoever picks them up. MS and Sony must already have contingency plans in case this happens. To me, it would have been either x86 or ARM A15 as the only real CPU choices. But an A15 wouldn't be able to use GCN on SoC as seamlessly.

 

[Shifty Geezer]

And with the pasture looking far greener in Sony land, it makes it that much easier to jump ship.

If the grass really is looking that much greener in Sony land, is BC going to be enough for you to buy the inferior console that'll play next-gen games worse than PS4?

 

[TheChefO]

Either Microsoft hopes to mitigate this in some fashion, or it believes the backlash from the loss of compatibility is outweighed by other factors.
Or maybe IBM didn't want to play this round.


It is just my opinion, but I believe it is backed up by a vast amount of evidence, that a multibillion dollar company trying to usher a project worth hundreds of millions of dollars isn't changing things just to be different.
Also, there are very few metrics at this time by which Xenon isn't beaten soundly by a core seven years more recent.
One of the factors was probably that the consumer technology world has changed, and with integration, IBM's existing product range has a very limited appeal.
Either pay IBM for a totally new and untested Xenon successor and get nothing in terms of SOC elements and graphics, or pay for an SOC whose basic elements have been in the pipeline and validated by AMD.

For what it's worth, if Jaguar and GCN are the underpinnings of Durango, it's a sign of greater investment or flexibility on the part of the system designer. IBM didn't pursue Xenon or Xenon-like cores as a serious product in its own right: these were not the crown jewels of PowerPC. IBM built the chip and found it so compelling a design that it forgot about it.
If the rumors are true, Microsoft is getting a design where it knows its partner has already bet its continued existence on those technologies.

I think in both cases, x86 was chosen as a cheap way forward for BC and in the case of MS, this is a standard tech which can accept an unaltered windows8+ install which will save them cash in the long run and offer a more compelling ecosystem.

Having said that, the underlying technology is not a world-beater.

The reasons for choosing jaguar had little to do with the architecture/performance itself and more for future growth/roadmap. Which is sad.

Had they brought AMD in early to customize an x86 chip which was designed for gaming scenarios (graphics, physics, floating point heavy perf), then this would be a much different scenario. As is, this seems to have been the low hanging fruit to accommodate their future roadmaps while spending little to get there.

It's almost as if MS wasn't making ~$1B /yr on gaming

@ shifty

It's one more thing to consider as many people will be doing when the machines are available.

No BC = freedom to choose the more attractive option

 

[Arwin]

Should either of them wish to move to GaiKai or similar streaming tech in the future, then using x86 is a pretty important first step anyway.

 

[patsu]

Should either of them wish to move to GaiKai or similar streaming tech in the future, then using x86 is a pretty important first step anyway.

Can't Gaikai run on ARM clients ? I thought they demo'ed on ARM before ?

Oh, Google native client:
http://www.google.com/url?sa=t&rct=...CPRP4qJRkPE_l6Wgukb7F1g&bvm=bv.41867550,d.cGE