Matthew Gallant writes a reply to Chris Hecker's rant at GDC a few months back.

At the time when I first read/heard Chris's rant it felt very good, since I've been thinking along those lines myself. But after reading Matthew's reply, I realize that feeling was in part a way of deceiving myself. I haven't been more productive by focusing on one thing, what I have done is drill myself down in new, deeper holes only tangentially related to the games I want to make. Just recently I added OS X joystick support to GLFW, which in itself was a pretty big feat requiring a lot of my skill, but it wasn't getting my game projects further along.

In reality, all these side projects keep me busy and makes sure I don't ever confront the fact that games need to be finished and released for me to get better at it. Now go read Matthew Gallant's post (and follow links and watch videos for more inspiration).

This maddening news really got to me this morning. Just can't let it pass without comment.

When people found out the PS3 Slim would not have the "Install other OS"-feature the question was raised if Sony was about to remove it entirely. They assured they wouldn't.  One could probably argue that Sony shifted its marketing message away from the (presumably) tech-savvy early adopters to target a wider base of customers more interested in the entertainment value. This feature is simply not a selling feature anymore. (Was it ever, really?)

Apparently, the reasoning behind not including the feature in the Slim was the cost of maintaining the hypervisor drivers. I suspect that now, after the geohot "hack" surfaced, Sony was faced with having to update the hypervisor anyway. Instead they opted to just scrap it entirely. Sony must have maintained the possibility of a hypervisor breach as a calculated risk and have had plans on how to mitigate this without crippling end users. Where are these plans now?

How many people actually use this feature? Not many I imagine. Out of the few people that actually installed linux on their PS3s it seems like most did it out of curiosity; installing full desktops and toyed around with emulators. It's unfortunate that these are the most prolific users, posting videos on youtube of their sub-par GNOME UIs. I think (or at least hope) that Sony's original motivation was to spread architecture knowledge by letting would-be PS3 programmers get their feet wet. Think the BASIC interpreter present in the Commodore 64 and how that kickstarted many would be programmers in their young years, myself included. In a similar way, Linux on the PS3 provides unique access to a real parallel architecture, where programmers can learn and get experience no doubt needed in the future as things get more and more parallel.

Some professional PS3-devs thinks this is a non-issue, that it changes nothing. Of course it changes nothing in the limited worldview of Sony Santa Monica, with full hardware access and (by now) vast platform experience. I am trying to create my own little corner in this business, I was hoping to be able to get at least some idea of how my code fares on a modern gaming console. I don't have the money and.. heck, I probably don't even fulfill all the requirements to even get the opportunity to buy devkits. The reasoning of people like Christer Ericson of Sony Santa Monica keeps me away from Sony development, away from learning. There is just no valid reason to ever keep people away from learning. Treating PS3 programming as a dark art and trying to keep outsiders ignorant, what a stunning way to make more developers choose your platform.

There's also the whole argument of the popularity of using networked PS3s as clustering hardware, providing computing power to solve research problems, but screw those guys! Right, Sony?

Totally worth it.

.. and I had no way of finishing my entry in time. There is a thread over at the tigsource forums where I describe my idea briefly. Please whine at me there or in the comments here if I don't finish it anyway.

The reason is a bit more complex than just me being sick. I had a massive amount of technical issues as I tried to make my code better and more coherent. The code has been worked on on and off for several months and variants of the same codebase has been used in many little experiments of mine. It was a cobbled together thing with little organization, so not setting a scope and starting to reorganize made time go by without much happening on the visible front. The codebase is well on its way to become better.

Then there's that 3D model format thing.

When I started to make some 3D models in Wings3D for this project I ran into some limitations I didn't like, so I decided to scrap my OBJ format converter and try to move that code into a custom exporter for blender. This turned into an interesting but time consuming task as I played around in the blender python API (which I'll use a lot from now on). The exporter is not 100% done, but what I need is there.

This is also something I see as a reason I don't finish stuff. Talk is cheap.

The theme for GAMMA 4 was announced yesterday along with rules. As a way to achive my goal for more openness, I have decided to take part in some events like this just to make myself release things publicly. I don't think of the GAMMA 4 projects as something huge since they put pretty strict limits on play-time among other things. The idea of GAMMA is not huge productions, it's more about being creative within the constraints given. Some awesome games have seen the light in previous GAMMA competitions.

I spent some time last night and today thinking about possible game ideas and mechanics. I have two ideas I want to prototype, but I'll share more on that later.

Unity is now available for free. That's really awesome news for everyone that wants to get started making games. Or at least the creative types. Unity is known for its content creation pipeline (amongst other features, of course), and how easy it is[linked text in swedish] going from art to prototype.

So most artists can't code. The free release of Unity will surely help these people. The other side of the coin is that we programmers often are bad artists, but I have yet to find tools of the same calibre for us.

So what would these tools be? The best candidates are probably tools to procedurally generate graphics assets. Cities and architecture can be procedurally generated, as can foliage, textures and terrain. I bet large studios like Bethesda use their character creation code to randomize NPC appearance, and possibly even some parameters of their behaviour. These are the tools I, as a coder, would enjoy having around. As it stands, tools like these are developed in house at larger studios and, if they are good enough, split off and licensed for lots of money to other studios (see SpeedTree and CityEngine for example).

Give a game programmer a problem and he'll kill it with math. I guess it's good news that the problem is solvable, but we should be able to share tools like this just as well as engine code.

Wow.. two updates in two days!

Wanted to drop a line for anyone who might want to check out what's happening. This weekend is completely full of moving around furniture and cleaning the house. That's what you get from juggling full-time work, family life and your own company at the same time; sometimes it all collides and priorities need to be shifted.

Hopefully I'll get some time to hack though..

.. just not visibly so.

This is very frustrating, because I'd really like to show what's happening behind the scenes. I'm working on core technology to make a couple of things easier and faster that's been taking too much time for me before. At the moment I'm in the finishing stages of reworking 3D object loading and a simple file format I save pre-processed 3D data into. The challenges has been making my "simple format" flexible enough, but I'm very nearly there.

Here's a screenshot of what it looked like when I debugged the loader functions..

Apart from the model I tested my toolset with having a few broken faces, it turned out alright.

After being told many, many times by people that I should start sharing what I'm working on instead of never showing anything I've come to the conclusion that they were right. I've been keeping too quiet, and sometimes that's the best way to make sure nothing you're working on ever sees the light of day.