It's more a legal thing which you will see with any commercially released codebase. The iD hackers have to jump through innumerable hoops with their publishers' legal department to get this code out to us at all, so it kinda sounds ungrateful to moan about the licenses.

Also, different parts of the code are covered by many different licenses and the copyright on them belongs to many different people, so it is important to be very explicit about everything.

When it's an OSS project owned and operated by a small group of people, there's little motivation to worry about this stuff. Also when you choose a license like the MIT one which basically says "I really don't give a fuck what you do unless you sue me," you have a lot less incentive to plaster it over everything, because you're one step away from being in the public domain anyway.

When you're releasing the doom3 code, things are very different.

EDIT: Deleted my GP post to stem conflict.

What I meant (probably poorly phrased) is that for readability for personal study (like gcv was asking about), I think it's worth stripping huge license preambles from source files. There's obviously good legal reason for them being present to start with.

You're a quick couple of shell commands away from stripping them off yourself, if it's that big a deal.