Thread by @svembu

1/ One of my early lessons in sofware and R & D is that complex projects cannot be precisely estimated in terms of schedule.

I still design and code and my engineer-self cannot give a good estimate to my CEO-self!

It is a very fundamental issue: iism.org/article/drivin…

2/ Not being able to come up with a precise schedule is true for any complex R and D project, not just software.

The "trains run on time" school of project management should ask "How long does it take to lay a track through unknown terrain?" because that's a better analogy.

3/ I once came up with a mathematical (!) explanation of why effort estimating complex software or R and D projects is essentially impossible but fortunately my being CEO meant that I could not get fired for the blasphemy!

So how do you track such projects?

4/ I track a complex software or R and D project by assessing the overall sense of progress. Are we making headway? Are we gaining knowledge? Even if the knowledge is "this approach won't work" because that is valuable knowledge too.

This assessment is necessarily subjective

5/ I used that dreaded "subjective" word! The sense the progress in a complex R and D effort is necessarily subjective and involves an aesthetic judgment, not only objective criteria.

In that sense engineering involves conviction, taste and judgement, not just numbers!

6/ The Japanese have epitomized this asethetic judgment in engineering - there is the notion of "avoid waste" or "avoid ugliness" which involves matters of taste and conviction.

So we need physics and metaphysics in engineering and we need numbers and good sense in business.

7/ So much of modern business is in thrall to "scientific" or "metrics driven" or "data driven" management techniques and they end up making organizations souless wastelands.

That is my fundamental disagreement with business schools.