[Paridel]

And well-developed OSS projects such as Linux don't let amateurs contribute. They can submit, but their submissions don't have a chance in hell of getting into the production code.

That isn't exactly true... FreeBSD doesn't let amateurs commit, but you can get your patch in Linux fairly easily if there isn't a competing patch especially if it is for hardware that isn't widely used. I had a sound card that had a very poor driver submitted to ALSA, later incorporated into the 2.6 kernel. Enabling it would make my computer randomly restart when playing sound. Quite fun. But it wasn't a widely used driver so even thought it was very poor code it made it in.

To be fair it has since been fixed. However how do you think Linux stays ahead of FreeBSD on drivers? I'm sure it's partially because they have more people writing them, but they also have a less stringent acceptance policy.

I'd say don't take my word for it, download the kernel source code and look for yourself, but not everyone here understands C.

If you don't understand C just untar the source and grep for various bad words. That's quite an eye opener too, but Linux has long since spoken about how the f-word in the kernel comments should be preserved.

Another example of in my opinion code that shouldn't have made the cut is in drivers/net/sunhme.c:

static void happy_meal_tcvr_write(struct happy_meal *hp, unsigned long tregs, int reg, unsigned short value)

Hey! That's funny! So much so in fact that it makes reading the code painful because you have no clue what's going on. That's OK though, there are comments in the code to help with that!

/* Welcome to Sun Microsystems, can I take your order please? */ /* Would you like fries with that? */

I would have gotten canned so fast for including something like that in production code it would have made my head spin.

Now being less picky about the code does mean that Linux adopts new hardware fairly quick, and people are normally quicker to fix a bad driver than they would be to implement a new one from scratch. So you can argue about whether or not that is a good thing. If you want stability you can always go with an older kernel. But no, bad code and code from amateurs and first time submitters does make it in there.

-paridel

[antiRepublicrat]

That isn't exactly true... FreeBSD doesn't let amateurs commit, but you can get your patch in Linux fairly easily if there isn't a competing patch especially if it is for hardware that isn't widely used.

You're right, drivers are a good exception.

I would have gotten canned so fast for including something like that in production code it would have made my head spin.

Did you read any of the comments for the leaked Windows code? That was pretty funny too.