English Language & Usage: Multi-Layered Discourse Room

12:11 AM

Oh, is that it.

Good to know!

12:29 AM

@Cerberus No, I'm serious. I've interviewed many, many programmers in my time. You talk with someone, give them a coding problem and ask them how they'd approach it, show them some function or method and ask them what it will return (and do the same with some code that has an error and ask them what's wrong with it), things like that.

You can pretty quickly assess someone's competence.

That doesn't tell you what they're like on the job, however. Some people can't produce under pressure, and it's pretty much all about pressure.

Cerberus

12:45 AM

@Robusto Sure, that makes sense.

@Robusto Why is it all about pressure?

Robusto

@Cerberus Because there are always deadlines.

Some places are worse than others, but if someone hires you for a job they want to be sure you can solve problems and get the job done on time.

Cerberus

@Robusto Hmm but have those deadlines been drawn correctly, then, if they cause so much pressure?

I mean, it would not seem optimal.

Deadlines do not always give one the most reliable results.

Robusto

Planning has guesswork as one of its components.

There are always things you don't know that you don't know.

And, as we find over and over again, nature always sides with the hidden flaw.

That's not to say that in a well-run shop it's all guesswork. Most of the time, with competent people, you are able to guess pretty accurately how long something will take.

tchrist

> In the mouth of the whale: lobster diver swallowed by humpback lives to tell the tale
The Cape Cod fisherman estimates he was in the beast’s mouth for 30 seconds; experts say the encounter was a fluke

Robusto

But then you will also find the vice president will overrule your team lead and say,"No, we need it by the beginning of Q2."

@tchrist Har har har.

A friend of mine and his wife are good examples. Both are programmers. He's brilliant, but obsessive. He'll focus on a single problem and solve it perfectly, brilliantly. But that takes time, and he frequently gets way behind in his work. His wife is a MacGyver: she triages, patches here and there, pulls in tools that help her get done, and she makes her deadlines always. She finds what will suffice and moves on.

tchrist

12:58 AM

@Cerberus The problem is that business plans run on schedules, and you set those schedules only after you've "planned everything". This never, ever works. That's because during the planning phase you see the matter as from a great distance, bereft of fine detail and unknown entanglements. Only once you have actually prototyped out the rough draft of a WORKING solution can you begin to estimate how long it will take to productize it. But this is always far too late.

So there is always incredible pressure to do what you said you would do before you knew diddlyfookingsquat about what "doing it" would actually require.

Cerberus

@Robusto Right, so less pressure in a well run shop?

Robusto

@Cerberus Yes. Pressure is worse in a poorly run shop.

Cerberus

@tchrist Then either don't make such promises early, or promise twice the time you think you will need?

tchrist

Thrice.

The rule of thumb is thrice.

Cerberus

Of course I make the same mistake for my own work.

Right, thrice or force.

Robusto

1:01 AM

Under-promise and over-deliver. That's the way of the world.

Cerberus

But I'd feel very bad about myself if I made the promise on behalf of others and forced them to work under pressure.

tchrist

You don't. You get them to agree that they'll do what they don't yet know how to do in that amount of time.

That's the problem.

Or you have them tell you how long. It always takes thrice as long as that even when you take this rule into account.

It's a bit like how projections, like progress reports, always become rosier each level up the reporting chain you go. Do that a few times and there's nothing left of reality to those.

Of course this can be much worse at "old style" companies than at those following more modern management practices and expectations.

A good example of this which you'll be able to immediately understand is the practice of creating a formal Incident Report every time something goes haywire that affects your company's customers in a disruptive way.

Companies that can produce those WITHOUT penalizing whoever messed up can learn from their mistakes and not repeat them. But those who penalize the people who messed up create a negative atmosphere so that people won't come forward and there's no good analysis to learn from, and the problems recur and exacerbate.

These are complex situations, and many parties are involved. You have to be brutally honest about exactly what happened when, who did what and who noticed what, and what was done to fix it.

There's a timeline portion of the Incident, then various insights and actions to take to prevent a recurrence.

Robusto

Most of the places I've worked have adopted the Agile methodology, which includes planning estimation. Only one ever did it right, and that's the company that got bought out by a megacorp.

tchrist

Yes, it's "the" thing.

Robusto

1:17 AM

When it's done poorly it's an abomination.

Transcript for

English Language & Usage: Multi-Layer…