On Wed, 2009-02-04 at 21:07 +1100, Ken McDonell wrote:
> Independent of my deep egotistical reasons for voting to maintain the
> "km" prefix, I think Max has a very valid point here.
> On Wed, 2009-02-04 at 20:04 +1100, Max Matveev wrote:
> > On Tue, 03 Feb 2009 13:10:45 +1100, Nathan Scott wrote:
> > nscott> - Renaming kmchart/kmtime -> pmchart/pmtime is fine with me. I
> > nscott> will keep symlinks for backward compatibility, etc.
> > Unless you plan to provide bug-to-bug compatibility between kmchart
> > and pmchart I think that renaming it will be a mistake: people who use
> > pmchart now (not a lot of them but they do exists) will expect the new
> > pmchart to behave the same way the old one behaves so you will need to
> > support the same command line options, same behaviours etc. It would
> > be easier to keep kmchart name and just let pmchart die.
kmchart provides pretty much the same command line options (it
is platform independent, so deep-voodoo Xresource stuff isn't
going to work) - but the options used were all based on either
standard PCP options or the online pmchart man pages - so these
match up as best possible. It supports all the pmchart formats
that any current user of pmchart would be using, so old configs
should just work (yes, if they don't that'd be a bug).
The same users who we might be "helping" by not using a pmchart
name, will just get left out in the no-pmchart-for-you cold,
when SGI stops shipping the old one.
So, I'm comfortable having a single pmchart binary, and intend
to keep a kmchart->pmchart symlink to it for existing kmchart
users - and for Kens ego, of course. ;) (isn't "k" for KDE? :)
I think the advantages of having consistent naming conventions
far outweighs any potential downside (of which I don't really
see any, in practice).
If you really really feel strongly about this, lets discuss on
IRC ... going to take a pretty convincing argument to budge me
on this one though, and I don't see one so far.