Why Software Engineering Sucks
nerdlife.net
nerdlife.net
He's a student with no real world experience in putting together a system. Because of that, it's easy to conclude that drones create software from reams of paperwork and ninjas wing it and hack it out in a few days with Coca-cola and Ramen.
The real reality is that neither extremes is useful or done exclusively. Sitting down and putting at least some of the architecture down on paper can be useful. If you have a team larger than two people, it might be necessary to document some of the process so that people know what part they are playing.
I wish him the best of luck -- he will be tempered to neither extreme after he's spent some time working in various places.
Given this scenario it is more beneficial to have
- Well defined initial requirements - Every item speced out - Proper class diagrams - Functional requirements - So much detail that even X can code it (where X is a current/future employee with a variable technical and communication skill and interest level).
He additionally cites the case where he has personally observed "100 person companies" that manages to hire Group 1's. I am more interested in hearing about this case rather than read about his personal principles.
"My personal principle is that if you’re not smart enough to be able to communicate well and produce good code without a strict formal process, I don’t want to be working with you."
I am still trying to figure out how somebody could use a strict formal process to overcome the disability to produce good code.
This pile of process and paper is not a necessary evil to them, but something they want to actively encourage. If they could spend anywhere between 3 and 24 months having meetings, doing design specs, fleshing out requirements, and going through 2^n revisions of the same, they'll push for 24 months regardless. They don't do this to be actively malicious to the company or the product, but because they don't care about the company or the product. To them, it is just another job, and the point then becomes to keep it and get paid for doing it as long as possible.
His first example sounds very close to the team I currently work with if you add in constantly touching base with the origional stake holders. But we are still engineering, we still make controlled changes that we, as a team, determine are the best fit for the problem.