Explaining software development methods by flying to Mars

Chris@feddit.uk to Programmer Humor@programming.dev – 746 points –
66

You are viewing a single comment

Someone shared this on Mastodon so I'll just repost my thoughts from there. (Bonus for Lemmy, I was forced to squeeze all my thoughts into 500 characters, so this is the most succinct I've been on this site!)


Pretty incredible how little people seem to understand these. For one thing, every method other than waterfall is a subtype of agile methodology. The major distinction is that waterfall has a series of phases from design through building, testing, and delivery that attempts to plan the whole project up front. Agile methods focus on smaller iteration cycles with frequent, partial deliverables.

Something like kanban is designed for continuous delivery: we want to go to mars weekly.

LEAN development is a scam though, that one is accurate.

The author is also hyping up waterfall too much. Agile was created because waterfall has its shortcomings (e.g. the team realizes too late that what they’re building isn’t what the customer wants).

But I also think it also represents how poorly implemented these ideas are. People say they do agile/kanban/scrum, but in reality they do some freak version of these.

So often it's waterfall planning and execution with agile names for roles and meetings.

It’s barely waterfall planning either. Often there’s no planning, at least no coordinated one.

Currently at my current workplace we lack coordinated planning between teams. It seems like everybody is working in their own directions and it can take months until we get feedback from other teams. Mostly a product management problem.

I agree, but agile and scrum are not meant to be followed to the letter no matter what. So people are doing it right if they notice some part of the process should be changed to make it work for them.

The amount of people who say they do agile/kanban/scrum but have never talked to a customer/end user, let alone released something, is frightening

LEAN from the web:

After each iteration, project managers discuss bottlenecks, identify waste and develop a plan to eliminate it.

1st iteration:

Project Manager A: Requiring approval of multiple Project Managers for the same thing is causing a bottleneck. So is having to wait for a specific manager for a specific topic.

Resolution: Let all managers approve everything and need only a single manager's approval.

2nd iteration:

Project Manager B: There are too many redundant managers. It's a waste of resources.

Resolution: Get rid of all mangers but one. Actually, let the engineers manage themselves.

3rd iteration:

Consensus: LEAN development is a scam though

My impression of management science, at this point, is that it's not. The good ones just do it.