Re: ANNOUNCE: grapevine 0.1 "The flying sheep" released

[Nathan Hurst <njh hawthorn csse monash edu au>]

On Sat, 5 May 2001, George wrote:

> On Sat, May 05, 2001 at 12:25:24PM +1000, Nathan Hurst wrote:
> > > On Sat, May 05, 2001 at 09:31:48AM +1000, Nathan Hurst wrote:
> > > > Are there any _fundamental_ reasons why we can't use syslog 
> > > > for grapevine too?
> > > I asked Georg this on IRC yesterday, but I didn't get a clear answer;
> > > certainly it is better to build on what has gone before.  If syslogd
> > > needs to be extended to support user-level notification better, that
> > > would be a good project to work on.
> > 
> > That was my point entirely (hence the '_fundamental_' :).  Syslogd can do
> > things like log to remote systems, write to console, and it has a known
> > configuration system.  It also means that existing log analysers can be
> > used and extended with gnome apps.
> 
> Well grapevine is more "per user" while syslog is more per system.  Though it
> think it would be fairly easy to make grapevine watch syslog and send selected
> entries to the user.  In fact this was always planned, it just isn't implemented
> yet.

Could syslogd be enhanced to become more 'per user' for the reasons
outlined in emails?  Before creating another standard, see if the existing
standard can already do it, and if not, whether it can be extended.
Otherwise we are covering the same territory again, repeatedly, ad
nauseaum, over and over.

Put another way: write a list of advantages of grapevine over syslogd, and
of syslogd over grapevine, determine overlap and existing working code.
Then convince us...

> > > But however it is implemented, I think grapevine is very interesting.
> > 
> > Absolutely, I was thinking about getting a student here to work on the
> > same idea...
> 
> Get a student to help me with grapevine then :)

I'll have a chat.

njh the grump.