Weekly Links #47

Hello, everyone. We had a long weekend in Romania, courtesy of December 1st falling on a Monday, and I spent it meeting with friends. In exchange for the newsletter being late, I give you a new version of RogueBot:

Yes, after going in the wrong direction for weeks, I completely redesigned the gameplay, and it's off to a good start this time. Even with just the absolute basics in place, it feels like a game. It's frantic. It's challenging. (If you want an easy game, try Buzz Grid.) It requires both dexterity and planning. And it feels like there's room for improvement, both on the player's and the developer's side.

In other words, a success.

In other news, on a recent visit to PyGame yelded a couple of intriguing projects. One is Pyjsdl, which allows compiling Pygame apps to HTML5 — not such a bad idea, considering Python is noticeably higher level than Javascript. On the other hand, you're losing out both on performance (due to an extra layer of calls) and features, since the HTML5 canvas is a lot smarter than the one in SDL, even with all the stuff PyGame adds on top. If you're really that tired of typing JS boilerplate (and I am, often enough, even as a fanboy) why not just use CoffeeScript?

Now, PyJ2D goes sort of in the opposite direction, implementing the Pygame API as a Java library, so as to allow games to be run on the JVM via Jython. Which on the one hand makes even more sense, since Python is a lot higher-level than Java. But on the other hand it's likely to be even slower, if for no other reason that all calls have to go through the Java reflection API. Also, the AWT canvas wasn't designed with games in mind. (But it still has functionality not present in Pygame.)

Of course, I haven't made any tests yet, so I could be wrong in both cases.

I'm nearing the end of this newsletter (short one this week... sorry), so I'll just conclude by pointing at this article about eliminating grind in RPGs. A topic I care about since, you know, I've made a couple of roguelikes, with plans for more. So far I managed by making combat avoidable — which includes tying progress to other aspects of the game. Specifically, exploration: you gain a level every time you go down to a new level of the dungeon. An uncommon approach these days (I've only seen it in Lost Labyrinth), which is too bad, because that's pretty much how the original D&D worked, allowing each player to tackle the game the way they deemed fun. But just try telling that to rockstar game designers who think they know what's best for everyone.

How would you do it?