• 0 Posts
  • 6 Comments
Joined 2 years ago
cake
Cake day: June 14th, 2023

help-circle


  • Well now. A few things, here:

    • there are not 9 × 9 × 9 × 9 × … possible ways to play. After the first move, 8 squares remain, and so on, so there’s at most 9 × 8 × 7 × … = 9! = 362880 ways that the game can be played, ignoring the fact that most of those can be eliminated as reflections and rotations, or as win positions before you fill the whole board.

    • we don’t care how we got there. Each square can either be blank, a cross, or a nought, so 3^9 combos = 19683, and most of those are illegal, as only the boards where there’s (one or zero) more crosses than noughts are good. And you don’t need to store ‘the computer’s move’, just jump directly to letting the player go again. Let’s guess we need at most a quarter of that.

    • we could have created a single web page with 5k anchor elements on it back in the HTML 1.0 days, ignoring the fact that it would have taken a while to download on our 28.8K modems. That wouldn’t have been 170 Mb of unnecessary tagging, even with the ‘lay it out with tables’ style we had at the time.

    Google do seem to have a predilection for reinventing the past, poorly. I hear that their bonuses are based on inventing ‘new’ things, though, so it’s in their interest to pass it off?


  • The side channel resistance includes such matters as ensuring that the cypher takes the same amount of time, regardless of the key, but also such super-sneaky insights as the amount of power used to run the cypher, which can be measured from the CPU temperature. Every bit of the cypher that you can be sure of makes it easier to guess the rest. And even if you coded this algorithm in assembly, the CPU will interpret it as microcode and run that, potentially leaving you vulnerable - this is not straightforward stuff.

    Like vzq says, implementing this properly is for a cross-disciplinary team of experts in their fields.


  • Just been replaying Castlevania: Lords of Shadow, but I very much doubt the sequel is what they’re working on.

    First game - amazing art, very imaginative, solid gameplay, great engine, pretty decent plot, good level design with a reasonable balance between fighting, exploration, and puzzling. A bit flawed - gameplay-wise, the camera hates you, you’re a liability in the air, and it takes far too long to get given your basic equipment - and it’s also far too QTE and exposition heavy, assumes that you’ve an attention span of about four seconds before you need a reminder. But it was the style at the time, and it’s a great game, especially in the second half.

    Sequel sorts out your bad air controls and gives you your weapons up-front. Retains the awesome art and design in the Castlevania sections, but contrasts them with the boring-as-fuck awful ‘city’ levels where you spend most of your time. Includes far too many lengthy instant-fail sections where you’re sneaking about as rats, or hiding behind pillars from some dudes - I’m Dracula, for fuck’s sake. Voice actors all sound bored to be there. Still assumes that you’re a goldfish that loves doing QTEs. Could also make a great game by only including half of it, but unfortunately the good bits are mixed in with the shite bits, rather than having a slow start and a strong finish.

    Really interested to see what they’re going to do - rarely have I seen a developer that can achieve such great things, but execute on them so badly, and fail to recognise what needs to be left on the cutting room floor.


  • Doom3 is one of the landmark games for GPU-based shading - it made a lot of use of GPU stencilling, multiple texture targets, computation in shaders; it massively advanced the state-of-the art in forward rendering. So much so that any modern GPU is very well optimised for ‘the things that Doom 3 wants to do’, because that’s what every game that wants advanced lighting wants to do. The problem then for using it for any kind of benchmarking is that basically any modern card will output Doom3 at 200fps at 4k and still be in power saving mode. It would be like trying to stress-test a CPU with Wolfenstein 3D - the state of the art has long moved past that, you can’t use the results of that to tell anything apart any more.

    Trying to get Doom3 to render in 16:9 resolutions tho, rather than 4:3? Now that’s stressful…