Work related post ahead.
Nice to see one of my favorite co-workers at FM talking about something our engineering team believes so strongly in.
It is not easy explaining these principles to people who want to see a detailed six month GANTT chart that takes six weeks to create and six days to undo. If you spend six weeks actually doing the work, you'll be six weeks closer to the goal and six weeks wiser.
The dude with a GANTT chart just has a GANTT chart.
One of our investors has a great analogy about choosing between being a bus or a heat-seeking missile. The bus has a route, a destination, and an itinerary that it must stick to. The heat-seeking missile is built to adjust. You fire it in the general direction then iterate and evaluate as you move forward. Hit the thing that's hot, even if it moves (and it will move).
I'll leave you with a Winer-ism I like a lot from 1995 called We Make Shitty Software…With Bugs! It's just the first four paragraphs (pre-permalink era?) so don't worry about seeing the phrase "Indigo Girls" on that page.
daveadams
pre-permalink era?
Well, sorta. It was pre-permalinks, but also pre-weblogs. DaveNet was an email newsletter, archived on the web. I think Dave was running scripting.com in a weblog-ish format by then, but DaveNet was a separate product.
tuesday
january, 6 2009
Andre Torrez
Ahh, okay, I just noticed there are perma-links to the other items, just not the lead. I just didn't want people to think they had to read that whole damn thing.
tuesday
january, 6 2009
usernameguy
I really go back and forth on this particular issue. I've seen the other side of this one, up-close, way too often to err on the side of Just Doin' Stuff. Small companies habitually lash out at whatever problem is closest so much that it becomes a mindset. And that mindset, taken too far, leads to some nasty tactical problems. It's like a greedy algorithm - sometimes they work great, sometimes they work awfully.
That said, I've seen about as many projects, and companies, fail from over-planning. God, those awful claw-out-your-eyes week-long planning meetings that tell you everything you already know.
Anyhow, GANTT charts have a place, I think. Just not every single project. When you're like, "wtf is going on here, I don't even know what to do first, we could be in trouble if we do the wrong thing first" - that's the time to whip up a GANTT chart. But that's like, maybe, 1 project in 6. They're only really useful on like, BIG projects. Big, big projects. And only ones with a lot of weird dependencies you can't break.
Having someone paid to be master-of-GANTT-charts is basically asking them to be useless most of the time.
thursday
january, 22 2009