I'd rather be precise than accurate

I was listening to a podcast from the Agile Conference 2005 - a little old now but still very relevant. It was Arlo Belshee who mentioned that, in relation to project management and project execution, specifically software development estimates, he is more interested in precision than accuracy.
Sounds odd at first, but think about it. What this boiled down to was that if you are precise, that is you can do / estimate precisely the same way each time, you can make that more accurate as you go.
The analogy to a rifle range was made. If you are shooting at a target, and put five shots in the zone, just outside the bulleye, all one inch either side, then you're pretty accurate right? If you put five shots through precisely the same hole, a whole 12 inches away, you're precise but not as accurate. But at least now, you can recalibrate (or make improvements/allowances), one way only, and all of a sudden you're accurate as well. If you go again, and this time only four shots go through the same hole, and one is left by an inch, something else has changed - you're accurate (probably more so) but now, you're not as precise and you'll want to know, what happened.
My take on this is that if you are going to compare results from one timeframe to another, or across projects, it is not that relevant what the unit of measure is, or how close to some arbituary target it turned out to be (how accurate it was), but more so, how it changed since last time. What impact has changes in your environment had to your performance.
This has just had an impact on me as I attempt to work out some costs verses benefits, verses customer satifaction verses any other measure I choose to come up with for some business intelligence anaylsis I've decided, on the back of this article on Gantthead, I, as an IT Manager, should have a bit of a handle on. There are so many costs to take into account - and while of course I want to be as accurate as possible, what is more important than the overall cost figure and it's accuracy, is the precision of my measurements each week.
Accuracy, will no doubt come into the picture come budget season - another story. No good being precise and running out of money in December I guess.

More Broken Windows...

Check out a ripping post, here, on theagileblog. It's about Rudy Giuliani, fighting crime and managing NYC Agile-style!

It's only been a week, and already I've saved at least 2...

About a week ago, I inherited the maintenance of a small ASP.Net system. Some people I used to work for have had a system they are intending to market, built for them but are unhappy with the usability of some of the screens and want some changes made. Not confident with the previous developer and lacking a little faith in what they are being told I suspect, they came to me to get my opinion.

The changes are fairly superficial on the whole, but after a whirl-wind tour of the current system and the changes they require I ventured off (source code in hand) to attempt to scope out exactly what was being asked for, and how I proposed to tackle the solution...

...three revisions of a scoping document later, I think we have nailed a solution that they will be happy with, and that I am happy is technically sound - enough so at least that I will be delivering something that is maintainable down the track. More so than the current system anyway.

Now there has been a lot of back and forth between myself and my "end customer". Sometimes I suspect she has been a little frustrated that I hadn't yet jumped in and started "making the changes" she requires.

I can happily say thoug, that the time we have spent splitting hairs in this spec, (about a week) has saved us probably two or three, a month from now. I would've delivered something she didn't want the first time, and the second major review was when she highlighted a couple of points I'd missed, which I would've been doing in the end anyway, but undoubtedly not charging for...so now it's all out in the open, expectations are set and in a day or so, we'll be off and coding - the right thing this time.

This is a scenario I've encountered many times and one that I think is common particularly in small development shops and projects. "You are working if you're not coding". I have experienced on many occassions now, the benefits - in both time saved and expectations met - of being particularly pedantic about ensuring everyone understands the ins and outs of what is being request, and repeating that constantly. Hidden "gotchas" or unspoken "must haves" almost alway come out of the woodwork at some stage.

.Net 3.0

I see where Microsoft have announced .Net 3.0. It appears though, we cannot expect too much new functionality from the framework, but really, the encapsulation of all the Foundations - which in itself if probably a good thing.

Here's a Fawcette article on the matter.