Software Management 0.5

Just this morning, it appeared to me, and there the solution was to all our problems. It had been there, so directly in front of my face, that I haven’t seen it, but it is so clear now.

We need to separate testing from the act of programming.

Wow. That’s a statement. But I’m serious. It has never worked and the large amounts of failing teams with Scram or Krumban, or whatever they call it, got it wrong. Yeah, Agile got it wrong. Collaboration is for the weak. After having spent over fifteen years with this crap, we need to get our bricks for the silos out from the closet again, and build up walls between those teams. Give testers different offices, on different floors, in different buildings, heck, what am I saying, give them different planets to be on, so they communicate mainly over the bug tracking system.

And I want to see a test plan, with every written detailed test up-front. Now. Show me. And I want to see Gantt chart based progress report, every week, and every day or even twice a day if there is an escalation. And I don’t want to spent time on re-planning. The initial plan must hold. Test design documents are fixed after being created initially. Yeah, that’s what we’re going to do.

While we’re at it, is there a way to get that waterfall model more heavy? How much celebration may we add to it? Just that? Come on, give me more than just the usual bug metrics and that crap. I want three additional testing cycles in the end. And I want QA to approve every delivery we make. Sure, they have to sign it. On paper, three copies to each major division head. Exactly.

If you haven’t figured yet, this is a rant about a personal Black Swan that a friend of mine just told me about his replaced management. Exactly, they’re separating the collaboration between programmers and testers, dividing them, right now, in this century, nearly ten years after the Agile manifesto and it’s focus on team-values, and cross-functional teams. This manager made the experience that developers and testers agree upon a delivery in a dysfunctional way. Therefore he wants to separate them again. And every major change project is failed if it is not implemented after 90 days. I’m not that good as a clairvoyance, so, what would you suggest my friend to do next?

  • Print
  • Twitter
  • LinkedIn
  • Google Bookmarks

5 thoughts on “Software Management 0.5”

  1. > And I want to see a test plan, with every written detailed test up-front. Now. Show me.

    The manager is doing it wrong. A test plan should never include the actual tests! :-)

    Seriously, tell your friend to start running. Now! Before those hastily built silo walls start thumbling down on top of him!

    — Zeger

    1. Well I used something I learned from Jerry Weinberg in Quality Software Management Vol. 4:

      When the measurement system starts breaking down, STOP!
      When the insults begin, STOP!
      When people start writing notes to the file, or speaking for the record, STOP!
      When somebody says, “Thinking is a luxury we can no longer afford,” STOP!
      When morale turns irreversibly bad, STOP!
      When the rats start leaving, the ship is sinking, so STOP!

      (page 345, Chapter 19.3 Ways to Know When a Project Is Failing)

      As I heard, the morale was already bad before the latest announcement. Wondering what it now turns to.

  2. Markus, I started reading this and thought you’d fallen out of the sarcasm tree and hit all the funny branches on the way down :)

    Yes, there are management types like the one you mention out there, unfortunately. They are the “trial-and-horror” types from the medieval age of software development.

    What to do?

    If you’re friend can’t engineer change from within (maybe some ways to find angles for leading-by-example) and hasn’t got the the stomach for the prolonged effort that it would take, then I’d suggest to start looking for a new job…

    Sometimes you have to know when to change and put your energy to better use.

  3. Markus, again a nice posting written by you. I recognize the situation as I saw it happen or discussed about more often. Of course there are lots of other experts which can come up with good advises. I’m sure we all are able to prepare the “I told you so moment”. We might able to provide well meant advises. Sometimes it is better to do nothing. A former chief of mine told me once: “Don’t try to sell Elastoplasts to persons who doesn’t noticed the pain yet.” Sometimes it is good that people run into a wall and hurt themselves. If you are aware of this to happen, be a good friend and be ready to help him when he start needing the help.

    Good advices can be found everywhere, I even provided one unsolicited. I think the strength lies in how to act based on own knowledge instead on which source advice can be copied from. Copied sources are often not common sources and therefore harder to use.


  4. Told my colleague to start a personal journal about his experiences in this time. Maybe at some later point I’m able to spread it for him, but I don’t promise anything at this point. At least it should help my friend to get some relief on the most immediate pain.

Leave a Reply

Your email address will not be published. Required fields are marked *