1. 16

  2. 6

    In my opinion the biggest problem with bugs is that they make software development unpredictable. As a manager, you should insist that developers write well tested code otherwise all your planning becomes completely useless.

    The obvious problem is that it works initially when complexity is low. No tests but predictions are still accurate. Later when the first deadlines are missed, the panic sets in and developers get pressured to work faster. Now it becomes a downward spiral. Software development is a marathon. If you start with an unsustainable pace, you will not make it.

    Unfortunately, there is some truth in the hero problem for managers as well. If you do your job and deliver smoothly nobody notices. If you mismanage the project, you get attention from higher management and in the end you can be the hero and get promoted.

    1. 4

      It turns out that developers were creating more and more bugs, only to fix them afterward and get the prize.

      I’m skeptical this ever really happened. Certainly not nearly as many times as I’ve heard the story.

      Anyway, people are not immediately praised for not creating bugs, but that doesn’t mean there’s no recognition long term. Write a program in a space that has a common class of bugs. Wait for competitions to have those bugs. Not have those bugs. Receive praise.

      1. 2

        I think you are right, OpenBSD is based around security, which is really long term bug prevention - and they get praise for it.

        1. 4

          The vast majority of unix users are using Linux. Praise of BSDs is rather like a pat on the head.

          and they get praise for it.

          “praise” in that sense is quite cheap. If users are praising BSD while using Linux, it means nothing. If praise is your standard for success then it’s easy to DDOS your capacity.

          1. 0

            Well, the openbsd foundation also seem to get a decent amount of donations.

            1. 3

              Compared to what? SUSE, one of the non-Red Hat companies, was pulling in tens of millions for Linux. Red Hat and IBM were putting in even more. Then there’s whatever the Linux Foundation gets. Then that core infrastructure fund or whatever it was probably invests in lots of Linux projects. OpenBSD gets pennies in comparison.

              It’s also funny you use OpenBSD as an example given this.

              1. 0

                You just listed a 2006 article, the year now is 2019 - Fairly certain things have changed. Obviously it isn’t in the ten’s of millions, but OpenBSD foundation don’t offer commercial services either, so it seems like apples an oranges.

                1. 2

                  In 2008, they got a total of $500,000 in funding per the website. That’s five developers in Silicon Valley. Maybe ten in an area with better cost of living. That’s not much money or praise for a large, high-quality platform quite a few companies depend on and more should probably depend on if it’s security-critical software. Meanwhile, this toy picked at random just got $6 million in funding pledges. The Pebble thing got $20+ million.

                  “but OpenBSD foundation don’t offer commercial services either”

                  I’ve always thought that was a mistake. It’s definitely good to not be beholden to companies that might demand cruft to be in there. That could be addressed with a layered or parallel offering. At the least, they could be charging for support, some enterprise features, and so on to generate funding for the project. Maybe some vacation money for the developers, too. :)

                  The counterexample is OpenVMS. There’s cultural similarities between OpenVMS and OpenBSD teams despite BSD folks’ disdain for it. It was a system built by engineers for engineers with a focus on quality and security. They had dedicated weeks to finding bugs instead of adding features. Instead of free, they actually sold it. Result was a marketable, ultra-reliable system built by engineers that actually got paid. If the developers don’t, some company should try to build something marketable to governments and big companies on OpenBSD doing something similar. They can donate some percentage of their revenues back to the project.