Game development a problem for PS3?

In large part there is no easy answer to this sort of thing. There are so many different possibilities, that it's impossible to say without some detailed insider information from both parties.

To answer your questions... I'd say it's too early to tell whether or not developers are having difficulties with the PS3. It's only been out for around six months, as opposed to about a year now for the 360 (or is it 2 years). It will take some time for developers to familiarize themselves with PS3 development.

As for why it's hard... That probably has to do with the way the PS3 is constructed. It uses IBM's Cell CPU, which is sort of like your average dual core chip for a desktop or laptop, only on some serious steroids.

Instead of just relying on raw clock speed, which has returns that get lower and lower as the clock speed goes higher, the Cell attempts to distribute the processing load across several cores. Well, rather that's the goal, but it's up to the developers to actually make it happen. So, you may have say 4 cores running at maybe 1GHz each, which isn't the same as a single core 4GHz CPU. A single core chip can only process one thing at a time, just at 4GHz in this example. A properly distributed app for the Cell could process 4 things at a time, at a rate of 1GHz each. Generally speaking, multiple CPU/core systems don't make things any faster, but they make things run more smoothly. On a game console, you can see the appeal to that. Different aspects of the game can be broken up and processed by the different cores simultaneously. As opposed to a long queue where a bunch of little things cause some of the bigger things to get backed up or vise versa.

The main problem with this, is that it makes the resulting code much more complex. The developer has to think about the best way to break up the code across the cores, and also have some way of dealing with the fact that data sent to Core 1 may not take as long to process as data sent to Core 2. So if say Core 2 is responsible for rendering the basic environment for a game, and Core 1 is handling the rendering of the characters, you need the data from Core 2 before you can do much of any real value with Core 1's data.

Up until the last few years, parallel programming (as this whole process is called) was largely the domain of super computers and very high end programs. So few systems ever had more than a single CPU, it was never really considered worth the effort to support SMP (Symmetrical Multi-Processing -- 2+ CPUs) in programs. It was mostly high end things, like server operating systems and large database programs like Oracle. Photoshop is probably about as low end as it gets for programs that supported SMP.

In kind of a vicious circle chain of events, since few companies found it to be economically worth while to support SMP in their programs, there was little demand for this skill in new hires. So, colleges didn't make it a required part of the computer science curriculum, maybe didn't even offer an elective covering the topic, making it hard for companies to find people who knew how to do this if they wanted to support SMP, so many would just abandon that idea.

I suspect that will start to change very quickly now that it seems multi-core chips are here to stay. Of course that doesn't do a whole lot of good right now.

While the 360 also has a 3 core CPU in it, the PS3 has something like 6. The 360 also uses an x86 CPU, which is a very old and very well understood platform. The Cell is a modified version of the PowerPC Power5 line. I also wouldn't be surprised if Microsoft did like they did with the original Xbox, and used a variant of DirectX for the main programming API. This may make it easier for game developers to put a game out on the 360, but will likely be rather limiting in the long run.

Another potential source of problems, is that companies like EA will, in an effort to reduce costs, try and create some kind of a common game engine for cross platform games. A lot of factors can come into play here. If they don't think that the PS3 market will be quite as big, they may not spend as much time on that end of things. Microsoft isn't above outright bribery either, not that Sony didn't do it to poach a bunch of long time Nintendo affiliates in the PSX days. They will offer subsidies or various other types of bribes to get game studios to make the 360 version better or even make the game exclusive to the 360.

About the only thing you can do is wait for things to improve, which I'm sure they will. A number of very new technologies are being employed in the latest round of consoles. It's going to take time for developers to get their heads around it all. By the time next year, I would expect things to be quite a bit different. And Sony plans to keep the PS3 around until around 2017 or so. Only time will tell what Microsoft does. They dropped the Xbox like it was nothing as soon as the 360 launched.

The long and short of all this is that it's a little unfair to compare games for a console that's been around for almost 2 years now, verses one that's only been out for around 6 months. If you compare some of the initial round of titles for the 360 to those for the PS3, you'll probably see that both sets suck. It takes a good year or two for the quality of the games to start improving.

Most of the people who say the PS3 is harder to develop for over the 360 have probably never written a single line of programming code in their life. They're just repeating something they heard someone else say. So far, John Carmack, of id Software is the only one I recall making such comments... And he got his start making PC games like Wolfeinstein 3D, DOOM, Quake, etc. If the 360 is more similar to PC development, it seem pretty logical he'd find that easier to develop for. So while I respect his opinion as someone who's actually done some programming -- far more than people who just repeat what he says without any kind of understanding of why he says what he says -- I think it's important to take into account influencing factors.

From a purely technical standpoint, the PS3 is better... But history is littered with technically superior products that failed in the marketplace. And once you get the basic idea of parallel programming down, it shouldn't be too difficult to apply that to 3 or 6 cores. It just takes some time and a lot of mistakes to really master those concepts. Of course for a lot of reasons, it would be very good for the console gaming market if Microsoft stopped making consoles. Having two competing consoles is kind of the sweet spot for that market, and having three just makes life harder for game developers.