Re: A linux patch for a typo

["Ronald F. Guilmette" <rfg@monkeys.com>]

In message <Pine.LNX.3.91.980528234527.8276A-100000@marx.ljm.wownet.net>, 
Lin Zhe Min <ljm@ljm.wownet.net> wrote:

>...announcements (or better called `specification-update'?) are done, by 
>which all system programme authors can be noticed and then react. I 
>suppose this kind of announcement can be done in linux-kernel list, or
>(if possible) in another list (linux-kernel-announce?), or even at a
>web site or in a well formatted file accompanying kernel distributions ...

Personally, I don't see the need for such a level of documentation and/or
formality.

As I have said before, the C library (and _all_ system header files) _are_
my interface to the kernel.  If _any_ aspect of the application=>kernel
interface changes, however small, I believe that that change should be
reflected in a corresponding change to _some_ one of the ``kernel''
include files (which the normal system include files are often including
indirectly) just as is happening now/already.

The disagreement which began this thread was not about a lack of adequate
documentation of changes.  It was about different ideas for how best to
achieve interoperability between various different kernel releases and
various different library releases.  As I have said, I believe that in
practice (and especially given the limited manpower available for develop-
ment and interoperability testing) that goal, while highly desirable, is
likely to remain elusive and (at least to some degree) unachievable.


-- Ron Guilmette, Roseville, California ---------- E-Scrub Technologies, Inc.
-- Deadbolt(tm) Personal E-Mail Filter demo: http://www.e-scrub.com/deadbolt/
-- Wpoison (web harvester poisoning) - demo: http://www.e-scrub.com/wpoison/