I'll second this. I built a hackintosh 2 years ago that ran 10.5 beautifully (leopard), with quad core, 4GB RAM, 750GB hard drive, for only $800. The problem is that now there is no driver support to run snow leopard, so I'm stuck running an OS from 2 years ago that probably won't be supported much longer, and I can't even install point releases very easily without risk of breaking my system.
So, 2 years later I've got a system that runs Windows or Linux beautifully but isn't really a functional Mac. This can be great for a short time, but just keep in mind that anything you build now will probably not run 10.7 when it comes out next year.
But if you're a hacker, and you like doing it (like I do), you can bring new OS support. For instance, I wrote patches to get my system up to speed on 10.6 the week of its release. Lots of other people now use my code. That's rewarding.
Like anything, it's a tradeoff. Do you want to learn about deep osx internals? Or do you want to go back to your day job writing webapps? Both can be rewarding; it's a function of your free time and how much you want things to "just work" vs how much you want to learn.
Not that I necessarily recommend the hackintosh route, but if you had 10.5 running, 10.6 should run as well. I had my system moved to 10.6 within a week of it's release.
I had an ATI 3870 which was supported in 10.5 but ceased to be in 10.6 - a shame, since it outperformed the nvidia cards in core graphics by a large margin.
So, 2 years later I've got a system that runs Windows or Linux beautifully but isn't really a functional Mac. This can be great for a short time, but just keep in mind that anything you build now will probably not run 10.7 when it comes out next year.