Graeme

I've got a big pile of old gaming stuff that's looking for a new home, as it's cluttering up my parents' house. There is a catch: It's located in South Cumbria (nearish to Kendal), and I can't deliver. I can post, but you'd have to cover the costs. Spectrum Games I've got a couple of knackered ZX Spectrums and a pile of games. I need to check what's still there, but I used to have things like Avalon, Astroclone, Fairlight, Lord of the Rings, Mercenary, Elite, Dun Darach, Alien 8. I suspect that there are forum members younger than some of the above. PC Games Various things from the mid-nineties. Daggerfall, Frontier + First Encounters, Starlord (that's a real obscure one), Carmageddon, Wing Commander 3, Privateer (and the expansion pack - 3 1/2" disk!), Privateer 2 and Sub Culture. Edge Magazine I've got a few old issues of this too. Mid to late nineties. I can't be asked to break things up beyond the three categories above. If I had the time to do that, I'd probably eBay it all. If someone wants all of the above then that would make my life really easy.

I tested the PocketPC version of Tomb Raider (were you on that too Spaffy?), which was very effective at wrecking iPaq screens. Regardless of how bad Tomb Raider was, Who Wants To Be A Millionaire (iPaq / Palm) was far, far worse. Graeme

My guess is that the alien invasion is down to the exploitation of "Imulsion", which is the source of the original war in the backstory posted at Shacknews. I also suspect that the COG will be proven to be the villains of the piece, on the grounds that: (a) They're an oppressive government; and ( they're socialists. I'm with Lailoken on the fact that this title has Generic Game Aesthetics: Big Guns, bigger aliens, rusty machinery, gothic looking architecture? It's all there. Graeme

I suspect that the PS3's multiple threads won't be that useful for AI (though I'll probably be proved horribly wrong ), as I think that the SPEs are really big vector engines. That's one area where I think that the Xbox 360 could win in terms of processor design, though it's a case of wait and see. The specs for the nVidia 7800 are pretty close to the RSX, albeit with a lower clock speed (450 to 480 MHz vs. 550 MHz). I'd guess that both chips will have a similar architecture, and I think that they have the same number of transistors, so a 7800 SLI rig should pwn the PS3. I would expect to see a slow down in GPU performance increases - nVidia claimed to be way ahead of Moore's law at one point - simply because the current generation of chips have 300 million + transistors, making them hard to fab. One of the big limits on chip design in general is having the staff to actually lay out the gates, but I think that graphics hardware tends to use a lot of identical, repeated elements. (OK, the one I was briefly involved in did, but that was its USP.) The PC gaming market seems to be less important than it was five years ago, and I don't think that it will drive a new generation much faster. Graeme

I should have said in the above that the Xbox 360 and PS3 processors won't be lacking in power, but rather it will be harder to utilise efficiently. Graeme

The reason that the Xbox 360 and PS3 are considered to have poor processing power for some tasks is that they don't do out-of-order execution. Out of order execution means that a CPU keeps a pool of instructions to process, and it can reorder them to some extent to allow for better efficiency, provided that the commands don't depend on earlier results. So, for example you might have a piece of code like: a = b * (c - d) x = y + z In an in-order CPU, like the new consoles have, then the instructions would be processed in order. i.e.: c - d b * result of (c - d) y + z An out of order CPU is capable of rescheduling instructions, so that it can use resources more efficiently. So if the CPU was waiting for the multiplication hardware to become free, the instructions would be reorganised to give: y + z c - d b * result of (c - d) Which would run faster than the code in its original order. The big advantage of this is that you don't have to tune the program to the same extent to get it to run fast, and less deterministic, branching code will also be faster if branch prediction is employed. The comments made in the GDC "Burning Down the House" panel are an upshot of this. I haven't seen the Gamespot article, but I suspect that it says the same thing. Basically, it makes the CPU less good at being a general purpose unit. So, why do it? The reason is that it takes up a lot of the chip's die space to implement an out-of-order design, and it looks like IBM, Microsoft and Sony have decided that more cores are more useful. With regards to the Cell, it always looked as though the SPE were pretty limited purpose units anyway. (They always struck me as being big vector engines.) I heard a rumour that Apple didn't go with the Cell CPU for the Mac on the grounds that it didn't work as a general purpose CPU. My guess would be that the Xbox 360 has a better general purpose CPU, whilst the PS3 will be better for doing large numbers of floating point calculations. Like Anandtech, I think that the consoles will be similar in terms of their overall performace, but will have quite different strengths and weaknesses. There's been no official word about the Revolution as yet, but rumours suggest a dual core CPU at 2.8 GHz. It's possible that Nintendo have gone for an out-of-order execution part (fewer cores than the Xbox's, so possibly some free die space), but we'll have to wait and see. The other thing to bear in mind is that previous consoles has offered graphical performance that's on a par with a contemporary top end gaming PC. I don't expect that latest consoles to be any different. Graeme

But it does need Hulk Hogan and Mr T. See: http://www.emogame.com/bushgame.html Graeme

In response to some of the points made in previous posts: Publishers are evil It depends on the publisher: I've worked with one who I heard allegedly never pays small developers without legal action, yet has been up for various business awards. This is not atypical for the industry, and some of the business practices are terrible. Developers should just worry about making good games Development is a business, and the developers should be concerned with making money. This is not a bad thing, contrary to some of the postings in this thread. Developers should sell their games direct This is a risky strategy. I'm not sure how much AAA titles cost to make these days, but I'd estimate £5 million to £10 million. Assuming that the return is £20 per copy (probably on the high side), then the developer would have to ship 250K to 500K to break even. As has already been mentioned, about 10% of games break even. The publisher's role is to manage risk, and that's why they take a big chunk of the earnings. There are other business models for sharing the risk - theatrical shows will often rely on a network of investors to buy shares in the production - which could work. Downloading hurts publishers, not developers Most developers never see any royalties, and in a lot of cases are running projects at a loss. (The publisher will fund the game to completion, not the initial prototyping phase.) So lost sales are not likely to impact them directly. However, if the publisher isn't making money then they won't be funding developers, who probably won't survive. And risky (i.e. original, third party) projects will be the first to go. And, Spaff, I don't know if you're still working for that Hove based outsourcing company, but traditionally external and contract work is the first to go in bad times. If you are, then going for warez could put yourself out of work. Professional software is too expensive Try the open source alternatives. Some are very good (Eclipse as an IDE, OpenOffice), and can compete with their commercial counterparts. Others are a bit rough around the edges, but work. Graeme

I'm sure that after all the games that the Thumbs staff play, you'd be able to defeat Viacom with your L337 skillz.

The Nokia room was after my time at that Company, but applying for an N-Gage dev kit requires to have secure premises that Nokia can inspect at any time, and you have to be willing to indemnify them for several hundred thousand euros if any of their IP leaks. Ironically, given Nokia's protection of their own IP, they ran a competition a while back for new N-Gage titles. Amongst the rules was a clause giving them sole ownership of all the IP for all entries, including any patents that might be used. That put me off even thinking about entering. Graeme

Hey, I only wrote the above because I had nothing to do at work this afternoon. Plus I'm tired of people just accepting some pretty extreme claims without questioning them. If you're interested in the ATI chip in the Xbox, take a look at: http://www.hardocp.com/article.html?art=Nzcx I don't think that the article is that strong technically, but there are two interesting points: (1) The embedded memory on the chip contains a bunch of processor elements. This is what allows the Xbox to do anti-aliasing for free, and sounds like a Pixel Planes architecture. (SIMD graphics approach taken by the University of North Carolina - try Google or Foley and Van Dam for more.) (2) ATI claim twice the performance of their top PC GPU for the Xbox GPU, so similar to the nVidia RSX in the PS3. However, Heise suggest that it might be half this based on MS's numbers. So there's another gap there. Graeme

Firstly, I misread Lailoken's post (and the Heise.de article) and thought that they were saying that the Xbox 360 has an X700 based GPU. My bad. Secondly, Heise.de's journalists don't question Sony and Microsoft's numbers too much, and are basically performing analysis on what is, in effect, PR. They make a comment about the Top 500 supercomputers list, but, honestly, do you really think that a consumer level console is going to be powerful enough to match a 128 CPU monster? Thirdly, no-one has pointed out the limitations of a SIMD architecture. They're great for mathematical calculations, so the PS3 is going to be a monster for processing physics and graphics. But they're terrible for any branch based code. (For the non-programmers, anytime the program has to make a choice between option A or option B.) So the PS3 will be less efficient than a PC or Xbox 360 at something like AI. Briefly, SIMD means Single Instruction Multiple Devices, and is a form of parallelism whereby a CPU can execute a number of identical instructions at the same time. The reason that it's bad at branching is that if you get a statement like: if ( x == 0 ) Do one thing... else Do something else... Then it's possible that the SIMD processor elements could find themselves on separate execution paths, thereby violating the single instruction thing. The way around this is to execute both paths (if necessary), i.e.: Activate PE if x == 0 Do one thing... Activate PE if x != 0 Do something else... Basically a SIMD architecture can be very inefficient, but picks up speed from being hugely parallel. There's a good chance that you won't see anything close to the peak performance in real world applications. Edit: I'm a software engineer with an interest in computer graphics, and I've worked with SIMD architectures. Heise is staffed by journalists, who may or may not have an engineering background, and probably don't have a huge amount of practical experience in the industry. Graeme

I took a look at the heise.de article last night via the Google language tools(I've not studied German for 15 years), and I think that they're barking up the wrong tree. The Xbox 360 graphics core (from the speculation that I've seen) is based on a totally different architecture to the X700 - many more pipelines (48 versus 8 on the X700) and a unified shader architecture. (Vertex and pixel shaders are part of the same pipeline, rather than separate on most other cards.) With regards to the performance, I think that Sony are being hugely misleading when they claim that the PS3 is twice as fast as the Xbox 360. For a start, no comparative benchmarking has been carried out on both systems, and these are theoretical numbers. My suspicion is that both use massively parallel SIMD architectures in their chips, and these can be inefficient. (In effect, the way you implement an algorithm on one of these chips means that a large part of it can be idle - you win on there being a large number of processor elements on the go.) Secondly, take a look at the specs of the current generation of PC graphics hardware. The current ATI X800 has 160 million transistors, the GeForce 6800 has 220 million, both are clocked at around 500MHz. Roughly speaking, the performance of a graphics chip is proportional to its transistor count, though the X800 and 6800 seem to be (more-or-less) on a par, suggesting that ATI have a more efficient design. The RSX is based on nVidia's latest generation GPU, which is claimed to be more efficient than the 6800 series, so it's probably somewhere around 1.5 to 2 times the performance of the 6800, as claimed. If heise.de are correct then we can assume that Sony's machine will be anything up to twice as fast as Microsoft's. But I don't think that they are right. My guess is that the Xbox 360 and PS3 are going to end up in the same ballpark performance wise, but they'll have particular strengths and weaknesses depending on the kind of software running on them. Graeme

The Gamecube's graphic chip was designed by ArtX, which became a part of ATi, and was originally an offshoot of SGI. Historically, SGI hardware never had that good peak performance, and I think that in some cases consumer level cards posted bigger numbers. The big difference was that SGI's kit could sustain performance close to their peak in real world applications, whilst the nVidia and 3dfx parts couldn't. I've got a feeling that I heard that the Gamecube chip (Flipper?) was designed along these lines. It'll be interesting to see whether the same is true for the Revolution. Graeme

I like Darwinia's cover art. Graeme