Really Important Software

  • The IEEE article makes a lot of sense.

    Particularly the need for recoverable 'graceful failure.' I remember reading that traffic signal controllers have a separate supervisor system independent of the normal timing and light switching circuits. It has exactly one job, to look for inconsistencies (double green light for example) and immediately shut down the system into caution signal mode. The beauty of this is that the supervisor has a very simple job, and  'knows' nothing about the normal day to day operation. Bugs or errors in the timing circuits will not affect the operation of the supervisor.

    • This reply was modified 3 years ago by  jay-h.
    • This reply was modified 3 years ago by  jay-h.


    -- FORTRAN manual for Xerox Computers --

  • Very interesting article, Steve. I especially enjoyed the analysis on the Vox website.

    Here at work we're taking baby steps towards introducing Agile project management. Good for us. Waterfall has been done here for many decades, so it's slow going. Unfortunately, my view is more from the outside, looking in. Being a large IT department, there's people (BA's) who get to know what's going on, whereas the rest of us are basically kept in the dark. However, I think, looking from the outside, in, that the temptation to know everything, such as you would try to do with a Waterfall approach, is still very prevalent. It appears to me, again looking from the outside, that there's a big temptation to get involved in "analysis paralysis", rather than, as you said in your article, adopt "... the idea that we don't know everything about how to build our software when we start."

    Kindest Regards, Rod Connect with me on LinkedIn.

  • Thanks for this op-ed and for the links. I have zero engineering training, but when I heard the news of a second 737 Max crash that sounded so similar to the first one, I immediately suspected a connection, and turns out there was. Ever since I read The Checklist Manifesto by Atul Gawande, I have been aware of the exemplary job that the air industry has done with regard to flight safety. Yes, there are many other things to complain about regarding airline travel, but it has been true for decades now that flight safety was treated as a top priority, and the safety records showed that.

    The scariest thing to me about the 737 Max crashes is the information coming out that business decisions could have trumped safety or routine checklist-style decisions. That would be a terrible change in process priority and has to be reversed at once. As you called it "Really Important Software" has to be held to a far more stringent safety standard that gets vetted for attempts to sneak business interests ahead of safety before an airplane is sent out to production.

    • This reply was modified 3 years ago by  webrunner. Reason: Add book link
    • This reply was modified 3 years ago by  webrunner.

    A SQL query walks into a bar and sees two tables. He walks up to them and asks, "Can I join you?"

Viewing 4 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic. Login to reply