Alan DuBoff wrote:
On Wed, 16 May 2007, Hugh McIntyre wrote:
Ian Murdock wrote:
> As I'll say till I'm blue in the face, I'm as obsessed with
> compatibility as anyone here. But I have to ask: What exactly
> would break if -h *were* the default behavior?
All of the (admittedly limited) versions of Linux I've tried provide
"df -k" as the default, not "df -h". Reporting disk usage in KB
rather than 512B as "df -k" does (and SunOS 4.x used to do) is the
only true way. of course :)
No matter, I think that Ian makes a good point. I use -h most of the
time when I use many of the utilities, df most certainly.
Why would we not want to make that the default?
I understand that it could break backward compatibility, if people are
parsing the output, but in some cases change in inevitable.
I personally think it makes sense to present the user (a.k.a. a human)
with human comprehendable data whenever possible.
I don't know if it has come up this yet or not, but with commands like
df where the default output for a human should be more useful - make df
change the output only when it is being used in interactive mode. A bit
like the man command activates the pager. Any scripts would still get
the old output.
IMO, a command without an argument should only be used by humans. i.e.
An argument should be provided to return machine formated data.
Doug
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
