On 02/28/2010 09:27 PM, Thomas Dalton wrote: > On 28 February 2010 21:05, William Pietri<[email protected]> wrote: > >> As to who I'm responsible to, that was Erik Moeller and is becoming >> Danese Cooper. We of course have a plan, which is publicly posted, and >> which I'm glad to answer questions on. Elsewhere in this thread (and in >> the blog post) I've explained why I haven't just made up an arbitrary >> deadline, but am instead trying to measure productivity and project a >> date. If you have further questions on this, let me know. >> > You've been working on it for months. Surely you and your team have > produced something in that time. Look at how much it is, compare it to > how much you think needs to be done (working out what needed to be > done was the first thing you did, yes?), do a bit of multiplication, > and give us your projected finish date. You shouldn't be "trying to > measure productivity", you should just be measuring it. >
That's an entirely reasonable approach, but there are two wrinkles. One, I underestimated the difficulty of releasing to a production-like environment. And until we have done that, we can't tell the difference between the things we hope are done and the things that are actually done. I intend to only measure the latter; measuring the former as if they were done is chancy. I am pressing vigorously for us to be able to do that soon, but there's only so much pressing you can do without long-term harm. Two, most software projects are inevitably exploratory. The difference between what we think we need and what we actually ended up needing is often large. So I could project dates based on all of the needs that we have discovered, and then somebody in the community will look at the software and say, "Hey, what about X?" And X will be some entirely reasonable thing that it is now obvious that we need. So I think it's better to release early and often and be open about the fact that it won't be really done until everybody (or, y'know, enough of everybody) agrees that we're now really done, or at least feel comfortable projecting that we're done. But if you'd like to make your own projections, all the data for the development work is exportable from Pivotal Tracker. If I thought I could take that data, or any other data, and give people a real date, one that they could have confidence in, I would be ecstatic to do so. But I can't, and I won't just give a BS date to get everybody off my back. William _______________________________________________ foundation-l mailing list [email protected] Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
