how did it get this far?

https://www.theatlantic.com/technology/ ... de/540393/

Lincoln's and Churchills quotes illustrate the core problem for software.

For the most part, axes and trees have very well known and relatively homogeneous properties. Variance of productivity is very low.

But coders, software tools, and software problems have unknown and heterogeneous properties. Variance of productivity is extremely high.

Indeed, software is a craft, not science and not (yet) professional engineering.

Quite true!

Here's a question: is math a craft or is it like professional engineering? If I give you a new postulate, what is the accuracy of the estimate of the length of the proof and the time required to produce it?

Maths is more like Philosophy, pure reasoning and consistency.

OK. So is philosophy a craft or is it like professional engineering?

Do philosophers & mathematicians have predictable rates of production?

I see the computer scientist took the longest time to produce an answer!

Still, what's the Venn diagram for disciplines in terms of predictable cost and predictable product quality? Which are craft and which are engineering? Which disciplines are like turning a predictable crank to get a predictable outcome (ax+tree=logs) and which ones involve irreducible complexity or novelty?

Do philosophers & mathematicians have predictable rates of production?

No.

I see the computer scientist took the longest time to produce an answer!

Still, what's the Venn diagram for disciplines in terms of predictable cost and predictable product quality?  Which are craft and which are engineering? Which disciplines are like turning a predictable crank to get a predictable outcome (ax+tree=logs) and which ones involve irreducible complexity or novelty?

CS calls it an edge case?

Do philosophers & mathematicians have predictable rates of production?

No.

Perhaps CS is the same and can never have the kinds of reliable production methods found in other types of engineering.

Do philosophers & mathematicians have predictable rates of production?

No.

Perhaps CS is the same and can never have the kinds of reliable production methods found in other types of engineering.

I see CS in its current form as a branch of (discrete) mathematics, for a large part the study of ADTs and operations on ADTs.


The crafting/engineering fun begins when we apply CS to problems.

//
Fred Brooks

rule of thumb for scheduling a software task:

1/3 planning

1/6 coding

1/4 component test and early system test

1/4 system test, all components in hand.