On Wed, May 28, 2008 at 10:19:38AM -0500, Mike Gerdts wrote: > On Sun, May 25, 2008 at 11:26 PM, Richard Elling <Richard.Elling at sun.com> > wrote: > > Seems to me that it should use iCalendar (RFC2445) format > > http://en.wikipedia.org/wiki/ICalendar > > no sense in inventing yet another calendar format... > > -- richard > > That is a wonderful idea. If the storage mechanism also allows for > attachments, this could be a good move forward too. With the current > cron implementation, I have run into problems with the following:
I thought so too, but, looking at it more carefully I see that iCalendar is geared around tasks to be performed by humans (and non-task items too), and as such solves one problem (time/recurrence specification) that we need solved here, and others that we don't (e.g., free/busy time). > Problem 2: Where to put the shell script? Right, iCalendar provides for attachments, which is where a script/program can live (or it could live in a comment property, or we could extend iCalendar if need be). Of course, some sort of convention will be necessary. And we risk pursuing a full calendaring solution here (now _that_ would be creeping featurism). Plus we'll want to consider using CalDAV as a protocol so we can get existing UIs for free :) (just kidding, mostly!). > As others have said - I think the value is in a decent programmatic > interface rather than in SMF itself. I'm not so much for or against > using SMF as I am for modernization of cron. Amen. Nico --
