Reporting and Fixing bugs (was: Re: Wolfram Alpha claims to be a primary source.)

I do other things than CAS stuff, and had occasion to use some programs from Microsoft. I found a bug, identified the cause of it down to a fairly narrow routine. Microsoft (at least at that time, maybe now?) provided no mechanism to report it other than suggesting that you join a newsgroup, and post it there. Which I did. I subsequently got a handful of suggestions that ranged from unhelpful/pointless like "did you try rebooting your system" to moronic like "reformat your disk and reload Windows, .net, ...". I am hardly an expert on these things but I (a) knew more than these morons, (b) knew how to read documentation and point out an inconsistency between behavior and documentation, and (c) displayed the courtesy to find a minimal demonstration of the problem / plus even a work-around.

Eventually I got an acknowledgment from someone -- apparently not a Microsoft employee -- but someone who was designated as some kind of expert by Microsoft -- that I had found a bug.

So compared to that behavior, almost any bug-report mechanism is an improvement.


What some people in this thread fail to recognize is that most bug reports are not particularly useful. The percentages will undoubted vary from product to product, but I think some average might be that about 1 in 10 bug reports is useful. What about the 90% remaining?
They might fall into one or more of these categories.
1. The user failed to read or understand the specifications, and the behavior is not a bug. It might be implicitly a request to improve the specification or the documentation, but it is not a bug. A technical response might be RTFM, but a more cautious response, like ignoring it, might be better.
2. The failure has been previously reported.
3. The user is reporting on a bug already fixed in the latest version of the software.
4. The report is useless because insufficient information is provided to reproduce it.
5. The failure is of such a minor nature that a correction is not cost effective.

It would seem to be the height of arrogance to totally ignore "user bug reports" and assume that internal quality control is sufficient to find all bugs. But it might not be a bad strategy to concentrate on fixing bugs reported by important customers whose business requires that software work as expected. It might even be a good strategy to add features as requested by important customers at a higher priority than fixing bugs that are on the periphery of the product. The primary goal of a business is to make money for its stockholders, which may or may not require making Mr. Cranky-Pants happy.

While people may be outraged by certain bugs they have found, it reminds me of an old joke where a man goes to the doctor and complains that it hurts when he does this {assumes a contorted posture twisting his neck and extending his arms, lifting one leg ...}. The doctor says, "So don't do that!"

How much money did you lose because you couldn't integrate a hypergeometric function.

The question of whether WRI is worse than other software vendors or worse than other CAS vendors, or whether Mathematica or Alpha is better or worse than something else, could presumably be studied. But I haven't seen anything but isolated anecdotes here. By contrast, it is pretty easy to object to " Alpha is a primary source..." whether or not WRI ignores bug reports.

Jon Harrop is an inveterate troller; he has posted some 5845 messages in 94 newsgroups, according to Google groups. His average over 2897 ratings is one star (out of 5). Statistics as of last month sometime.
It's best to ignore him.

RJF


.

Relevant Pages

  • Re: Filing bug reports in Debian (was Re: Debian Stole My Name!)
    ... I've been working with Christian for years on the aptitude package; ... user in response to a bug report. ... What there is, terms of a response, is, "It's not ... I deal with bug reports. ...
    (Debian-User)
  • Re: Debian Love
    ... ever got Debian to work, ... Unfortunately, the developers either did not get the message, ... what is the bug ID number of the bug you reported? ... I feel fine about that: all bug reports go directly ...
    (Debian-User)
  • Re: How to improve the quality of the kernel?
    ... >>> 20 other patches in the tree depending on it causing a regression, ... a big infrastructure patch exposing a latent old bug in some ... I think that we can handle bug reports like we handle modifications of code. ... these people can choose to use the bugzilla or any ...
    (Linux-Kernel)
  • What is a bug, what is a bug report?
    ... It is an almost universal observation from people who receive bug reports professionally or as a courtesy as part of maintaining free software, that the vast majority of so-called bugs are simply "user error" and have almost no informational value whatsoever. ... Microsoft deals with bug reports by telling people to subscribe to some newsgroup, and hope that some other person not paid by Microsoft will help "fix" the bug, or more likely point out the error in the complainer's way. ...
    (sci.math.symbolic)
  • Re: Known problem with LinearGradientBrush?
    ... Has anyone filed the bug with Microsoft? ... I am glad to hear that the WPF team is on top ... However I have filed about 10 bug reports in the past, ...
    (microsoft.public.dotnet.languages.csharp)