Re: 2.2.1-ac1 oopsing

Peter T. Breuer (ptb@it.uc3m.es)
Fri, 29 Jan 1999 20:15:32 +0100 (MET)


"A month of sundays ago Alan Cox wrote:"
>
> Does 2.2.1 vanilla oops in the same way ?

You have found the advantage of a (minor) code split. Easier testing.

I usually split my code onto two platforms and work first one then the
other, porting continuously. The minor differences that arise often
allow faster bug hunting. It certainly also keeps the code better (less
specialized, fewer hidden assumptions ...).

It's akin to genetic programming. One deliberately mutates the code
base _horizontally_ to find slight advantages or extremes of behaviour.
Fruitflies are irradiated that way sometimes too :-). Same idea.

Linux has hitherto kept to a vertical development model. People do
experiment horizontally with patches, but the mutations are always
merged back or discarded. They are never deliberately kept around.

Debugging in Linux has depended more on mutations in the userbase (my setup is
different from yours) to get the genetic variability needed for infomation
gain. But mutating the code base is also valid. I suggest you keep
maintaining .acX as deliberate deviants. You and Linus should port back
from each other, not just one way traffic.

It _helps_ debugging. It helps development.

Linux and bsd* show the same mutual (genetic) interchange. Successful
code is cloned. Unsuccessful code is tried out and pruned in both.

If only I were a generic algorithmist ..

> Alan

Peter

-
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/