On 5/12/07, Mike Kupfer <[EMAIL PROTECTED]> wrote:
>>>>> "Ian" == Ian Murdock <[EMAIL PROTECTED]> writes:
fvdl> I'm not sure what you mean here. Are you talking about making -h the
fvdl> default? I doubt that would be standards compliant..
[...]
Ian> What exactly would break if -h *were* the default behavior?
[...]
Ian> (unless there's a whole collection of scripts that parse the output
Ian> of df that I don't know about..)
One of the things system administrators often do is monitor their
systems for signs of trouble, which includes partitions that are getting
close to full. I would not be surprised if there were a lot of
homegrown monitoring scripts that parse df output.
Ian> (And, I say to say, the fact that we're arguing that we must
Ian> preserve non-human-readability because that's just the way we do
Ian> things kinda highlights a larger point...)
The phrase "that's just the way we do things" seems like a straw
argument to me. Has anyone actually said that? (Perhaps I missed it in
the flood of mail... )
The reason for maintaining the default format is to avoid breaking
users' software.
The reason for maintaining standards compliance is to make it easier for
users to write portable code.
Now, we can have a discussion about whether POSIX is still a relevant
standard. Or we can have a discussion about whether it's worth forcing
users to change their scripts. Or both. :-)
My point is, there are often good reasons for the way things are in
Solaris. It will be difficult to hold a meaningful discussion about
changes if we gloss over or (appear to) trivialize those reasons.
Very fair point.
My larger point: There's a big difference between ABI
breakage (something Linux does regularly, and you're absolutely right to
criticize it for doing that--but that's OPPORTUNITY) and tweaking
the default output of a utility that may be used in a handful of scripts.
-ian
--
Ian Murdock
650-331-9324
http://ianmurdock.com/
"Don't look back--something might be gaining on you." --Satchel Paige
_______________________________________________
opensolaris-discuss mailing list
[email protected]
