LÖVE

Robin wrote:

tentus wrote:I'd be surprised if switching compression formats gives you even 1% reduction in final size.

That's a bit overstated. Each megafile might be optimally compressed, but when you .zip a lot of files, there's bound to be some sort of repetition across files compression formats can take advantage of, like tar.gz files. (Come to think of it, I don't think ZIP actually does that, but it could.)

Maybe. Before I wrote that post, I ran a quick test using Kurosuke, a 4.16 MB game. I made the following four files using their default settings in 7z/winrar. Note that the smallest is only 10% smaller than the largest, and this is with a game that is only 1/5 the size of my hypothetical game.

(I'm actually a little confused how the .tar managed to increase file size, rather than reduce it.)

.tar files aren't compressed, you use it with other compression software like gz, bz2 or xz.

thelinx wrote:.tar files aren't compressed, you use it with other compression software like gz, bz2 or xz.

Shows how much I know.

The more you compress something the more processing it takes to decompress. It would make distributing easier, but thats the only benefit. I'd rather have faster loading time, that and zip can still compress pretty well, you just need to tell your compression tool to do so.

Someone (I forget who, it might have been David Rosen) made an observation recently that zipped files can take a shorter amount of time to read, especially with modern hardware, because CPU speeds are so far ahead of HD read speeds.

slime wrote:Someone (I forget who, it might have been David Rosen) made an observation recently that zipped files can take a shorter amount of time to read, especially with modern hardware, because CPU speeds are so far ahead of HD read speeds.

Probably true on a bit of a sliding scale. The more, smaller files you have, the truer that will be (8 1kb files vs 1 32kb file, for example.)