-----Original Message----- From: Cees de Groot [mailto:cg@cdegroot.com] Sent: Friday, February 28, 2003 6:31 PM To: The general-purpose Squeak developers list Subject: Re: Possible goals for future releases (was: Re: Let's Release 3.4)
You set goals or you set dates. As long as this is a volunteer effort where people wildly fluctuate in the amount of time they can contribute, you cannot set both.
I'm in favor of setting dates.
I think you need to propose targets for BOTH goals and dates, with an understanding that either the goals or the dates are the higher priority.
Some goals may be high enough in importance that you are willing to move the date, and some goals may not be that important.
Setting dates provides a motivation for people to choose to spend some of their time on things that they want included.
If you only set dates and no goals, Squeak is less likely to get assigned a priority in people's daily lives, and the date may arrive with nothing new to release.
I think these ideas are fairly basic principles of project management, but feel free to disabuse me of my pre-conceptions.
-Dean Swan
On Sat, 2003-03-01 at 00:43, Swan, Dean wrote:
If you only set dates and no goals, Squeak is less likely to get assigned a priority in people's daily lives, and the date may arrive with nothing new to release.
Well, if you define 'goal' as something you want to accomplish with Squeak, as opposed as 'target' - I define it here as a goal you want to have completed by a certain date - that's fine.
A goal is to modularize Squeak. It will very likely not be accomplished in a single release. Another goal is to document it. Goals typically shift, in reaction to the environment.
And I don't believe that a date will arrive with nothing new to release if you don't set explicit targets for a release. In a quarterly release cycle, postponing your work (missing the current deadline) will mean almost half a year delay, immediately.
Furthermore, I'm not saying that we should stick with this regime forever - see it as a 'therapy' against: - slipping release dates and thereby uncertainty with outsiders what is happening with Squeak; - a rusty release process - with a strict quarterly schedule, you get into a rythm and the whole process can be smoothed out quickly; - to a lesser extent, protection against 'too big' releases that have the associated risk of breakage. As I said, my experience is that it helps tremendously; my experience is usually with weekly or fortnightly release cycles, but that's clearly to short for Squeak.
squeak-dev@lists.squeakfoundation.org