Recently, I was thinking about how long it takes to write software and how hard it must be to estimate for bigger projects.
Because I look at something like, Andy, you've often talked about how you, when you're bored or when you want to learn a new language, you'll make Snake in that language, which is a pretty simple, straightforward game.
And I assume doesn't take you more than a few hours, right? But then if I said to you, write an operating system from scratch, like that would take you years probably.
And, you know, what about a fairly complicated application? Well, that might be I don't know, three or four months, maybe more.
How do you even go about estimating how long things are going to take?
Yeah, well, by the way, how long it takes to write Snake massively depends on what environment you're using.
If I'm really familiar with it, then I could probably do it in a few hours, but everything takes longer than you think.
That's rule one of estimation.
Normally like three times as long as you think.
And if you take that into account when you estimate, then it just takes three times longer than that.
But then you have to say three times longer than that.
It just keeps compounding.
It's always longer than you think.
But people have been doing this for a long time, right? And it's not guesswork.
There are techniques.
A lot of people don't use those techniques.
But there are techniques, so allow me to rant slightly.
The key factor in whether your estimate is accurate or not is how small the bits you break it up into are.
So the more detail in your estimate, the longer it is, and the more accurate it is.
Now obviously, you've got to balance that against doing so-called waterfall development, where you design everything up front, and you say it's going to take us 18 months, and then you come back in 18 months.
and it's like 10% done or whatever.
But the more detail you have in your estimates, the more accurate they are.
But here's another golden rule for you.
Don't estimate in time.
Humans are unbelievably bad at saying how long something's going to take.
Recently, I was thinking about how long it takes to write software and how hard it must be to estimate for bigger projects.
Because I look at something like, Andy, you've often talked about how you, when you're bored or when you want to learn a new language, you'll make Snake in that language, which is a pretty simple, straightforward game.
And I assume doesn't take you more than a few hours, right? But then if I said to you, write an operating system from scratch, like that would take you years probably.
And, you know, what about a fairly complicated application? Well, that might be I don't know, three or four months, maybe more.
How do you even go about estimating how long things are going to take?
Yeah, well, by the way, how long it takes to write Snake massively depends on what environment you're using.
If I'm really familiar with it, then I could probably do it in a few hours, but everything takes longer than you think.
That's rule one of estimation.
Normally like three times as long as you think.
And if you take that into account when you estimate, then it just takes three times longer than that.
But then you have to say three times longer than that.
It just keeps compounding.
It's always longer than you think.
But people have been doing this for a long time, right? And it's not guesswork.
There are techniques.
A lot of people don't use those techniques.
But there are techniques, so allow me to rant slightly.
The key factor in whether your estimate is accurate or not is how small the bits you break it up into are.
So the more detail in your estimate, the longer it is, and the more accurate it is.
Now obviously, you've got to balance that against doing so-called waterfall development, where you design everything up front, and you say it's going to take us 18 months, and then you come back in 18 months.
and it's like 10% done or whatever.
But the more detail you have in your estimates, the more accurate they are.
But here's another golden rule for you.
Don't estimate in time.
Humans are unbelievably bad at saying how long something's going to take.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.