Bob, a property owner, is talking to Sally, a woman he’s just hired as a gardener.
So, just keep the grass mowed and watered and the lawn edged and that’s it.
There’s usually cold cokes in the fridge, we’ve got a brand new lawn mower. If you get hot, feel free to use the patio or come inside for breaks.
“Hmm”, thinks Sally, “This guy’s not too bad.”
Sally gets to work. She’s mowing away when Bob comes out to get the mail. As he’s going past, he stops to ask how it’s going. As he leaves he points at a strip along the street and comments that they used to have tulips, and maybe they’ll put some in again.
Later on he runs a short errand. As he’s walking out to his car, he stops to compliment her on the nice looking job, and says some day he’d like to put in a hedge. He’s got a nosy neighbor.
When he comes back, he’s thoughtfully brought her a cold coke. He also stopped off and bought a couple hedges. Can Sally put those in for him?
The trip to the nursery gave him a great idea – what if they took out the gravel along the walk and put in a vegetable garden?
Sally finishes the mowing and is edging when Bob comes out with an article he’s printed out on xericulture. Maybe they should be ecological and replace the grass with….
But he’s talking to the wind. Sally’s long gone.
Scope creep. We all know about it, we’d never do that, right?
Bert’s CTO of XyzzyCo. He’s just sent out an all company email about how cool FizzyFoo is. Yesterday he sent out one about TireFyre. Last week he announced a plan to change the company database to NumNum. The week before that… Oh, I don’t remember.
Bert really enjoys learning about new technologies. He thinks it’s the most enjoyable part of his job – a real relief from the day to day grind. Speaking of which, he’s due in that meeting about why the team’s not advancing on it’s roadmap. Probably there’s going to be grumbling – there always is.
Maybe he can cheer them up by telling them about the new methodology he read about in a magazine and wants to implement.
Scope creep by enthusiasm. Insidious version of this.
This can particularly be a problem in AI-ish circles, where power often does come from several weaker systems. Instead of making what we have work, we add Tensorflow or Watson or Flubbycalc without actually thinking the problem through.
And ‘enthusiasts’ are often drawn to such things.
Additionally, it’s difficult to prove that we don’t need XYZWord in our NLP project. But wiring XYZWord into the project is probably not where we need to be. Improving testing and validation, and then improving ability to meet tests in a thoughtful manner that handles many cases is the road to salvation.
The problem isn’t choosing cool technologies, t’s changing them. If you choose new and “cool” technologies at the beginning of the project, fine, if you keep using them and not try to replace them with something “cooler”.
Methodologies can become ‘enthusiasms’.
A friend tells me he was working on a project where they had a comprehensive set of format rules, and enforced them strictly.
During the project, the team kept arguing about the rules and changing them. Whenever there was a change, they’d reformat (by hand!) all the code.
Code format and conventions were strongly emphasized over functionality.
Of course the project went nowhere, and eventually my friend left.
So it’s not just change. It’s also being more interested in the tech than in the problem.
In summary, some ways enthusiasm is dangerous:
- Scope Creep – seeing the cool project without seeing the cost
- Technology enthusiasm – being more interested in playing with the technology than in getting to ship.
- Faddishness – changing technology horses mid stream
- Artistic sensibility – being more interested in your api design, code conventions, methodology, cool language, theoretical advance, or vaunted expertise than in shipping useful stuff.
Thanks to RLa who provided a lot of thoughtful discussion of this article.