I got my first ever incoming link, and it’s from RUSSIA. Yes, apparently someone in Russia thought that Roach Roundup was cool enough to write about.

According to the Babelfish supplied translation, Roach Roundup is a “small, but sufficiently amusing toy”. The translation of the plot, like anything babelfish generated, is fairly amusing to read as well. I’d link it, but babelfish doesn’t seem to allow that.

Many thanks to the reviewer, this really made my day :).

The review: Логические таракашки

Roach Roundup has been updated to include music and all the other little things I left off in the initial month period because I ran out of time. You can snag it from the “Games” section on the left.

I’ve heard it said that a true artisan never blames his tools for his shortcomings. I say that better tools let you get more work done faster and with less aggravation, and in that vein I can only say FINALLY to the release of Gimp 2.4.

A quick qualifier here. I am not a Photoshop hater. I’m just cheap. Moving on…

Gimp has finally fixed my largest beef with the software, and a cause of much consternation with my image editing, especially digital paintings. Previously, when zooming your canvas in and out, the Gimp would simply do a point sample on the image, causing all kinds of chunky artifacts and making it extremely difficult to do anything with a canvas scaled below 50%. No more. The new Gimp finally scales canvases smoothly when zoomed, something that I wish would have been implemented a long time ago.

Besides the ability to see your canvas at scales smaller 100%, a quick glance at the toolbar on the side reveals quite a few new tools. There are a couple of new selection modes, none of which I think I will use, and a Photoshop style healing brush. Neat.

But the real kicker for me in the tools department are the new transformation tools. Any selection region or layer can be squished, scaled, rotated, perspective distorted, etc, using easy drag tabs. Which I think is just awesome. The previous transformation tools were almost universally difficult to use, some of them even required you to put in numbers and kind of guess what the result would look like. WYSIWYG tools like this are a great addition to the program.

One thing that I’m really curious to test out is memory consumption. The previous Gimp did a fairly poor job of managing memory, with layers taking up RAM according to the number of pixels on the layer, instead only storing portions of the layer that weren’t empty. I hit a few walls on larger images that required me to cut some corners and scale back a bit before, hopefully those times are a thing of the past now.

I haven’t really had a chance to get any hands on time with it yet, over the weekend I plan to do some coloring work to get used to the interface changes and such. Check back Monday-ish or so, I should have some new color work in the gallery by then.

Here are the Linux and Windows builds for Roach Roundup. I finally managed to get the Windows version to link, and barely. I don’t think I’ve ever run into that kind of dependency hell before.

Only four “levels”, no sound, no music, and no gameover/success screen. Hardly an awe-inspiring release. Still I did manage to get a good bit done with the time I had over the past month, and I learned a simply ridiculous amount of useful knowledge, both from doing research during development as well as from my myriad mistakes. Next time, things will be better. It may take two or three games, but these will get fun and polished eventually, I’m promising myself that. And hey, I’ll probably revisit this in a few months when my generalized engine is done and redo it the “right” way.

As for what’s next, I’m not sure. First thing is going to be a good long nap, I’m kinda burned out right at the moment. But not too long. I’m think I’m going to start extending SplashMap into a more robust game editor in a few days after I catch my breath. Currently, it only does maps, but now that I’ve discovered Lua, I realize that with very little effort I can extend it into a full visual game editor, ala gamemaker, but one that uses OpenGL transformed sprites for everything and is tuned to do exactly what I want it to do, and very, very quickly. Most of the engine code is now in place (thanks to Roach Roundup for filling in a lot of the holes), exposing that to the mapper should be fairly easy. I realize after doing Roach that personally, it really helps me to be able to see my work as I’m doing it, bouncing back and forth between code and game design more than necessary just adds frustration to the process unnecessarily. It also makes making changes to the game design hard. Towards the end, I realized that having rooms of varying shapes would have made Roach Roundup 200% more interesting, but because I foolishly took the route of just hardcoding the room shape in instead of finishing the C++ rendering pipe for SplashMap maps, I couldn’t make the change without a prohibitively expensive amount of effort. Live and learn.

Oh, and here are the files in question. The Windows version should work fine on pretty much any 32-bit windows (95 might be a little dicey), the Linux version requires that you have the SDL and SDL_mixer libraries for whatever distribution you happen to use installed. Oh, and OpenGL drivers or else Mesa.

Roach Roundup for Windows
Roach Roundup for Linux

And that single error message, dear readers, is the only thing between you and roach killing nirvana. This is driving me crazy, I’ve got the engine done just in time and I can’t compile it for Windows because of a linker hiccup. Wonderful. It’s way too late for me to continue beating my head against it, though, so I’ll get it compiled up as soon as I have the time and can figure out why ming is choking.

All told, though, I don’t know that it’s that big of a loss if it’s a day late. There still aren’t any sound effects, or music, and I really doubt there will be. The month is up, and for as much as I think it would be cool to extend this project as I talked about doing a few posts ago, it’s tempered by even parts of wanting to get back to work on Muse. I’ve learned a lot while working on this project. Here are just a few of the highlights:

1. Don’t attempt to design your game by typing coordinates into the source. You should use some sort of visual editing tool, preferably one that lets you edit scripts right into it. The less you have to touch the engine source, the better.
2. Use those scripts. C++ is not a good choice for the game logic most of the time, keep it to engine logic.
3. Create either your assets first, and then your code, or create your engine first, and then fill it with assets. Doing both at the same time is tricky.
4. Don’t put off verifying that compiling to windows works until the night of your deadline!
5. Don’t get stressed out about going halfways on some items of a one-month prototype.

So, yeah. In some ways, this whole experiment went better than I expected, and in some ways, it went worse. I would have liked to have gotten some more polish into the game, and some sound, but considering how badly I allowed it to billow into having menus and an intro story and everything, I’m not sure I would have had time to do both anyway.

So that’s that. I’ll post builds for Linux and Windows in a few days when I have the chance to get ming working on windows, and then it’s full speed ahead on getting enough of Muse done to put out a demo of that. Hopefully a few months at most.