1

I'm writing a technical document that uses checked run-time error and unchecked run-time error as terms of art. I need to talk about what happens when a computer program does something bad and causes one of these errors, but I'm not sure what verb to use.

  • In English, we make a "mistake," but I'm pretty sure we don't "make" an error.

  • Do we commit an error?

  • Do we cause an error?

Ideas appreciated. Sources, example usage, and authorities even more appreciated.

Sample phrase:

If client code verbs an unchecked run-time error, the implementation provides no guarantees; anything can happen.

  • Comments are not for extended discussion; this conversation has been moved to chat. – Kit Z. Fox Dec 2 '15 at 19:02
  • The error was injected due to erroneous coding or hardware design, and it was detected or sensed when runtime checks in hardware or software, uh, detected or sensed it. It is latent if it has not been detected. – Hot Licks Dec 2 '15 at 19:05
  • But note that an unchecked runtime error is one that occurs without being detected. This causes the error to propagate into later stages of processing where it's effects may be detected, or may be ignored entirely, causing the error to, eg, cause a check for $1,234,567 dollars to be mailed, vs the correct amount of $1.23. But the terminology around this can get pretty detailed, and you really need to find a good tutorial on the terms and concepts involved. – Hot Licks Dec 2 '15 at 19:10
  • 2
    In this context, I always use the verb trigger: Client code triggers an unchecked run-time error. – Graffito Dec 2 '15 at 20:06
  • 1
    @NormanRamsey The comments can be upvoted, too. But they can not be downvoted. And there are [insert favorite expletive] out there who don't bother to explain with a comment why an answer is wrong, simply downvoting and moving on. I have seen it several times on different answers from other users which I have found useful, or neutral. It once happened with my answer, too. From that time on, I also prefer commenting to answering. – Thinkeye Dec 2 '15 at 20:53
0

Make an error seems to be far more broadly used than cause an error and commit an error as the below Ngram Viewer shows:

enter image description here

However, if you search "code makes an error", "code causes an error" and "code commits error", only code causes an error appears. I would advise you to use cause before an unchecked run-time error.

enter image description here

  • @HotLicks Can you elaborate on your comment? I am at a loss. – user140086 Dec 2 '15 at 19:58
  • The Ngram Viewer looks like a really nice tool; I wasn't aware of it. Big crossover between "commit" and "cause" from 19th to 20th centuries. And "trigger and error" comes on only since 1990. Very interesting; thanks! – Norman Ramsey Dec 2 '15 at 21:47
1

Personally I think I might use any of make, commit, or cause depending on what I was talking about.

I think my likely choice would be in that order, i.e. make, commit, cause.

If the error was due to something I'd done that was of an everyday nature, such as written an address wrongly, I think I would say Oh dear, it looks as if I've made an error.

If it was something very big and important it might be commit. E.g. A large minority of people in Britain believe the House of Commons is committing a major error in agreeing to authorise the Government's extension of its bombing campaign to Syria, as well as Iraq

If it was a discussion about how an error had occurred, of if it involved a system or procedure of some kind I think I would say The error appears to have been caused by a programming fault or There will now be a big argument as to who was responsible for causing the error.

  • Basically I agree with WS2. However I have a comment concerning the word "error". I think an "error" is something that happens, whereas a "mistake" is something that is committed by somebody. So "make a mistake" is correct, but "commit a mistake" is more lofty style, which I would prefer. – user9 Dec 2 '15 at 19:37
  • Keep in mind that "commit" has a specific meaning in computer science. – Hot Licks Dec 2 '15 at 19:42
  • When an error "happens" to stop a hardware peripheral from working, like for instance a paper jam in a printer, the device may send a signal that it has a problem. It could make a red light turn on, meaning "I need help from a human!" That's what I associate with the language "raise an error". That's not a mistake -- these things happen. – Greg Lee Dec 2 '15 at 19:48
1

Java, .NET, and probably other modern languages use the terms (which are also keywords) throw and catch to describe error handling. If your code generates an unintentional error, then it is said to 'throw' an error, also known as 'exception', and there may or may not be exception handling logic that 'catches' the exception. When you 'catch' the exception, your code may take some type of action such as logging it, and then possibly 're-throw' the exception back to the calling code so it knows something's wrong.

Here's probably more info than you ever needed to know https://msdn.microsoft.com/en-us/library/fk6t46tz.aspx on this

  • Thanks, but by its very nature, an unchecked run-time error is neither detected nor handled. – Norman Ramsey Dec 2 '15 at 20:28
  • Right. So in your example, I would just say "If client code throws an unchecked run-time error, the implementation provides no guarantees; anything can happen. – public wireless Dec 2 '15 at 20:36
  • 1
    There is no throwing action here. An unchecked run-time error is silently undetected. – Norman Ramsey Dec 2 '15 at 21:36
  • To err is human. To throw an exception takes a third generation programming language. – user662852 Dec 2 '15 at 22:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.