• Hanzito (unregistered)

    Which makes me wonder: excluding bugs covered by already reported bugs, is there an upper limit to the number of bugs?

  • (nodebb) in reply to Hanzito 
    if (inv.inv_defect_number == 0) {
	out.println("<option selected value=\"0\">Select</option>");
	out.println("<option value=\"1\">Defect 1</option>");
	out.println("<option value=\"2\"> Defect 2</option>");
} else if (inv.inv_defect_number == 1) {
	out.println("<option selected value=\"1\"> Defect 1</option>");
	out.println("<option value=\"2\"> Defect 2</option>");
} else {
	out.println("<option value=\"1\"> Defect 1</option>");
	out.println("<option selected value=\"2\"> Defect 2</option>");
}
  • (nodebb) in reply to Hanzito

    I heard a story from a guy who worked at a sub-contractor for Hitachi years ago. He said that Hitachi had this policy that any code which is developed, tested and found to be fulfilling the requirement, and checked in, then doesn't go anywhere - not even to QA - until the author himself finds a minimum of 3 bugs in his own code. And, according to my acquaintance who told me this story, everybody always found 3 bugs in their own code - AFTER they developed it and it "worked."

    In my experience, I find bugs in my own code all the time, whether I can find 3 in every "module", however it's defined, is difficult to say, mostly because I don't have the resources of Hitachi to continuously re-check everything. Having said that, I find that story to be very believable.

    Therefore, to answer your question, you can say that any system out there has at least N * 3 bugs, where N is the number of feature changes.

  • (nodebb)

    @Hanzito ref

    Which makes me wonder: excluding bugs covered by already reported bugs, is there an upper limit to the number of bugs?

    Like most measures of complexity, I'd say we can only give a rough upper bound.

    But my vote is that the maximum number of possible bugs is O(number of characters contained in all source code and similar quasi-executable artifacts)

  • my name (unregistered)

    old code? yeah, at least 3 weeks

  • Sauron (unregistered) in reply to Hanzito

    Yes there is, because the amount of code is finite.

    Now of course, we could imagine that a whimsical manager changes his minds every day about what the program behaviour should be, and opens new bug tickets. But even that artificial source of bug is finite, because the manager's lifespan is finite.

    Now of course, we could imagine that the company keeps recruiting such whimsical managers. But even that source of managers is finite, because the world's population is finite (and can't grow forever, because the Earth's resources are finite).

    Additionally, the resources needed to track the bug tickets are finite (whether it is Jira or "Manual Jira" https://thedailywtf.com/articles/Manual-JIRA- ).

  • (nodebb) in reply to Mr. TA

    Sound like a case of playable metric.

    The idea sounds nice, but it is most easily handled by including 3 intentional bugs. Just a little off-by-one here and there,BAM, three cheal bugs caught.

  • (nodebb) in reply to R3D3

    Yes I thought about that too, and most likely in a corporation like that, there was a good amount of what you are describing going on. Having said that, I'm sure with managerial oversight doing outrageously obvious bugs is difficult.

  • Duston (unregistered) in reply to Hanzito

    "If you have an infinite number of monkeys typing code..."

  • Dr, Pepper (unregistered)

    Who knows why this code was written. A normal process involving code reviews should have flagged this. But maybe the important thing was to get the code into production before the deadline; with the assumption that later on someone will come back and fix it. Of course we know that never happens; so the poor code hangs around forever (and ends up here).

    It's broken code, for sure, but maybe it allowed the developer to reach the actual goal of having something running in prod by the promised date.

  • Steve (unregistered) in reply to Sauron

    "Yes there is, because the amount of code is finite." I remember reading a long time ago (possibly in "The Psychology of Computer Programming", but don't quote me!) that when IBM would release an update to the OS for the 360 they would fix N (where N on the order of a few thousand) bugs - each and every time they did a release...

  • Steve (unregistered) in reply to Mr. TA

    "until the author himself finds a minimum of 3 bugs in his own code. " So if I force 3 bugs into the code I'm home free.

  • (nodebb) in reply to Sauron

    Now of course, we could imagine that a whimsical manager changes his minds every day about what the program behaviour should be, and opens new bug tickets

    Re-categorize the tickets to "request" or "story". A bug must be something the application does that violates its specification. If you have an application without any specification at all, then you have a bug free application - what it does it what it does and it does it flawlessly.

  • (nodebb) in reply to Jaime

    "Works as coded"

  • (nodebb) in reply to emurphy

    Yes, I know it's not the best situation.

    However, you really have two choices:

    1. End the madness now and force the stakeholders to actually write down how they want a bit of the application to work. Now you have the beginnings of a specification.

    or

    1. Continue to react to "bugs" that might be actually software defects where the original request was clear. They might be obvious bugs where the application does nothing productive. Or, they might be whimsical re-thinks of the functionality. This is the train the dysfunctional team is currently on, staying on it isn't going to produce better results "this time".
  • löchlein deluxe (unregistered) in reply to Dr, Pepper

    What do you mean, code review? Pay people to look at code when they could be writing code instead? Bah humbug!

    (I would hope there's an antfile somewhere – anybody remember ant, the thing that was in between make and maven? – that generates that code.)

  • Paulie (unregistered)

    Pro tip: If you burn a citronella candle next to your computer, it helps keep bugs away. Try it!

Leave a comment on “Location Chooser”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article