Messages in this thread | | | Subject | Re: Ken Thompson interview in IEEE Computer magazine (fwd) | Date | Tue, 04 May 1999 14:39:54 -0700 | From | Ed Hall <> |
| |
Nat Lanza <magus@cs.cmu.edu> wrote: > . . . > Good god, have you LOOKED at some of the code in the kernel? Yes, > there's some very nice code, but there's also some crap. I don't > really see how you can deny that. After all, if it's as wondrous and > perfect as you claim, then why is there still kernel development? > . . .
Agreed!
The weekend after my university received the tape of Unix 7th Edition (to upgrade the 6th Edition system I ran there), I read the entire kernel source from beginning to end. If a program could be a thing of beauty, this was it; I've seen nothing quite like it in the 20 years since (and I've seen a lot of code, including some later versions of Unix). Yes, by today's standards the identifiers were short and the comments sparse, but the shear clarity of its construction made them unnecessary. It all fell together like a single, extended thought.
But in terms of features, Unix 7th Edition would simply suck compared to Linux. And parts of the Linux kernel (I don't claim to have read all of it) achieve a similar level of clarity. But hundreds of developers spread around the world can hardly produce the seamless code of two developers who spent long hours together at adjacent terminals. So we really can't expect Linux to have Research Unix's coherency and elegance. But we can strive to better what we have.
There is much strength in our numbers. Our many viewpoints brings us a larger space of solutions to bear on each problem. But we have to be conscious of our weaknesses. There is a genuine danger that the Linux kernel will become harder and harder to develop because of the many different (and not necessarily compatible) ideas behind it. It isn't that folks like Linus will no longer be able to understand and work with the kernel confidently, but it will be harder for the rest of us to "get it right." Yet "making the kernel easier for new folks to understand" isn't exactly a priority. This could be our Achilles Heel...
-Ed
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |