>>>>> "krm" == Kenneth R Mort <ken@post3.mort.net> writes:

   krm> CF had a limited (although large) number of re-writes. . . .
   krm> I was wondering about i-nodes

That is why we were saying to use noatime and nodevmtime mount
options.  With noatime inodes supposedly don't need updating just
because you read a file to update the 'last accessed' time, but
'finger's mail-last-read feechur stops working.  nodevmtime prevents
'finger' from reporting idle times, but the mtime of ttys doesn't have
to be updated every time you type a letter.  though maybe /dev should
just be an mfs like on the sysinst root, that's not what i do yet.

maybe some day this will be an lfs vs. ffs issue.  especially since
there is no seek problem with CF so the ``data-reorganizing cleaner''
is unnecessary, and in cheaper pocket embedded devices with onboard
FLASH chips instead of CF, you just have raw flash blocks that could
correspond to LFS segments, no flash translation layer or
erase-balancing.  but maybe the industry as a whole, not to mention
lfs, isn't to that point yet.

-- 
NetBSD: The choice of hundreds worldwide
		-- Tom Harvey