Comment by lanstin

Part of the problem is, to borrow an analogy, the frame has gotten really big and the space for the painting is smaller. We recently finished a two or four sprint to ... deploy yellow world. It has bitbucket pieces, Jenkins pieces, EKS terraform, et.c etc., we were all super happy that it worked, but it doesn’t do as much as you can do with twenty minutes in basic in a 1990 PC. And most of the people that wired all this stuff together can’t even write code to draw parabolas or calculate or whatever. It used to be I would develop for a few months then spend a week or two getting stuff in production, for new projects, or few days adding features/bug fixing form existing code, then get it shipped in a few days. There was thorough testing thru code all along, and evil network layers, and so on., and most of the head space was in the code. Now it seems like teams end up spending months trying to get a docker image with all their dependencies and yaml files and so on. And the code is just taking some SQL that ran in a database and running it on some new cloud platform that is still solving acid problems Oracle mastered twenty years ago. Half the code is just orchestration run this hello world Spark job that moves data from here to there. It is cobol level Scala.