Why a question was put "On-Hold"?

Question

I saw this question: Fixing Raster Calculator Error 000539 ImportError: No module named numpy? ; and it was put on hold due to it being of topic, and the following reason:

"Questions describing a problem that can't be reproduced and seemingly went away on its own (or went away when a typo was fixed) are off-topic as they are unlikely to help future readers." – Chris W, John Barça, Erica, PolyGeo

I however think it is valueable for future readers, in particular new-home users of arcmap. I must say I flagged it as a duplicate to ArcGIS Raster Calculator error 000539 ImportError: No module named numpy?, which showed the same error. However, solutions presented are different. Therefore, I think that to mark it as a duplicate can link to different, and possibly good solutions, to one problem. I.E., either use map algebra, or install numPy

Answer 1

As the instigator of that particular vote and close reason, I thought I'd share my logic.

First, I did see the earlier duplicate vote but there wasn't as much info at the time and I skipped it. That particular question also hit three review queues at once, so prioritizing what it was being evaluated on was also a factor. I don't disagree that it can be considered a duplicate and marked as such, and I'm ok with that approach.

The reason message is fairly clear. It's a corrupt/improper install and not something that can be reproduced; per the asker's own answer, 'moving to another computer' isn't really a valid solution that will help most people. Nor is 'do a reinstall'. I want to point out something from the help files here on what is (or isn't) on topic:

We are not a GIS Software Support Site. We encourage you to seek official routes for support for specific issues you may have.

Now we don't adhere strictly to that. There are several questions that are of this nature where the bottom line answer is 'reinstall'. This could fall into that category, hence why duplicate is ok and a valid option. However since that's already covered at the other question, I just felt the not reproducible reasoning was more appropriate for that one rather than keeping both.

It's something of a fine line - is a question about why my SQL syntax isn't working off-topic because I've made a syntax error? We see a lot of those and they usually stay. All of those 'oops I did something the wrong way' can have value, especially since some mistakes are more common than others. It's really just up to individual interpretation/opinion and having watched the questions on the site for a while. I've seen a number of similar 'oh it was a bad setting/install that caused it' questions handled in the same manner, so that influenced what I went for.

Another factor is knowing something about what happens when questions are closed. Sometimes, they go away/get deleted eventually. I looked at the question and determined that if closed as it was a the time, it would disappear anyway after a while (regardless of whether it's off-topic or a duplicate, it's still closed). I've already mentioned that I felt it didn't really hold much value as a sign-post duplicate sticking around (I'll also mention that while workarounds help and can be upvoted and even accepted, they still don't actually answer the question).

Answer 2

As the fourth and final person to vote for this to be placed on hold, pending closure, I would have been happy to see it closed as a duplicate of the earlier question that you cited, or to have both those questions closed as being off-topic due to being non-reproducible.

In both cases a corrupt installation of NumPy appears to be what has led to the observed symptom, but no procedure has been provided to reproduce that corrupt installation. I've now add an answer to the earlier question saying that an uninstall/reinstall of ArcGIS for Desktop (with the Python, NumPy, etc that comes as part of it) is likely to be the expedient to restoring a working configuration, and made the later question a duplicate of it.

However, when I came to this question it already had three votes to close based on being non-reproducible so I opted to endorse rather than over-ride them with a duplicate vote.