[lkml]   [1997]   [Jul]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Subjectdocumenting the kernel

Albert D. Cahalan writes:
> There will be more successful kernel development in the
> future if adequate documentation exists. Specifically,
> the best time to document the dcache/inode stuff is now.

If anyone starts the initiative, please tell me. I'll try to help as

I've got quite an experience with a design and implementation of large
systems, but it's completely different with an O/S kernel, mostly
because of very tight links between hardware and software and because
the kernel code is not properly commented. It is sad, that some great
guys in the Linux kernel development think of people, asking for the
code to be documented, as of a bunch of whiners, who just want to
bother them. Instead, the better documented the cod is, the more
people will be able to fix bugs they encounter without asking others,
so the more time the gurus will spend on introducing new features
instead of fixing the old bugs.

It's not fun just to beg for bugs to be fixed without contributing
oneself. But you gurus, ples realize that it *is* impossible to take a
stock 2.0.x kernel and to start reading right away. A good document
comparing a modern Linux to the hypothetical UNIX described in the
Bach's book (listing all improvements of Bach's algorithms and data
structures) would be great. As soon as I finish reading Bach, I'll try
to start writing something like that.


Alexander L. Belikoff
Berger Financial Research Ltd.

 \ /
  Last update: 2005-03-22 13:39    [W:0.037 / U:1.688 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site