This metapost discusses the use of python tags for Shapely libraries and the top-voted answer suggests that removing the python tags limits the potential audience as those who favourite the python tag may be missing out on un-tagged questions. I do not think this is being applied, even though it is the highest up-voted answer.

Specifically, in regards to this question where both my arcpy and python tags were rejected. If a user specifies that they are open to a pythonic solution, then shouldn't we be tagging this as per the meta? The mod comments on that post suggest that minimal code needs to be provided for these tags to be used and no python code was provided by the OP. I cannot find a meta post that suggests that this is required - especially when a question is more theoretical in nature.

Further, there are posts that contain the python or arcpy tags without any given code in the question as well 1, 2, 3.


  1. When is it appropriate to use these tags when the meta isn't being followed?

  2. If no distinction is made by the user for python solutions or arcpy solutions which if assuming 1) holds true, is an appropriate tag?

I am not interested in discussing minimal reproducible code and closing those questions, I am more interested in when the tags should be applied and why this isn't consistent.

  • 1
    arcpy is specific to arcgis production functionality "performing geographic information system (GIS) functions available in ArcGIS" pro.arcgis.com/en/pro-app/arcpy/get-started/what-is-arcpy-.htm
    – Mapperz Mod
    Apr 5 '18 at 17:04
  • 3
    As expressed in comments beneath Luke's answer, I agree allowing using [python] and [arcpy] if OP wants to (and strongly disagree when someone goes on a campaign, against a meta post, editing out [python] from every question tagged [arcpy] (don't forget to vote on that post, if you haven't voted already)). continues... Apr 5 '18 at 18:32
  • Providing code in questions when they are abou programming is another subject which was also discussed in Meta. See: gis.meta.stackexchange.com/questions/3994/…. Apr 5 '18 at 18:33
  • @AndreSilva thanks for those links. I linked the 2nd one in my q already - it seems there is an upvoted answer but it is not really followed through at all. The question asked for python solutions, I added a python tag it was rejected, that goes against the meta answer. Am I missing out on the correct use of these tag(s)?
    – GISHuman
    Apr 5 '18 at 20:44

I am going to first focus on your closing summary:

I am more interested in when the tags should be applied and why this isn't consistent

because I believe consistent interpretation of the site's guidelines is very important.

This site's guidelines, like those of any Stack Exchange site, are numerous and are therefore absorbed over time rather than on the first day someone uses the site. I think of them as a hierarchy starting with those we want people to become aware of and try to adhere to first:

  1. Those that are in the Tour
  2. Those that are in the Close reasons
  3. Those that are in the Help Center
  4. Those that are tagged at Meta GIS SE (and Meta SE)
  5. Those that are being fleshed out in Meta GIS SE (and Meta SE)
  6. Those that are yet to be fleshed out but are evolving as precedents in the way the community processes its Q&As

You cited Why is the Python tag inappropriate for a question about Shapely? and expressed concerns that the guidelines present within it were not being followed. That is a Meta GIS SE Q&A that I am acutely aware of because it is one where I modified my workflow to no longer remove from questions about shapely and a plethora of other spatial Python libraries. That was not the reason that I removed the tag from that question, and why I think your question here is not a duplicate of that.

The reason was to try and reduce the scope of the question to explore the GUI solution option first. The asker seems to make clear that this is their most important question and that a Python solution is a fallback. That fallback can always be asked about in a new or revised (if no answers) question that demonstrates a code attempt.

To elaborate some more, in terms of curating content on this site, like the question Creating polygon boundaries to group points based on common attributes using ArcGIS Desktop? which triggered this Meta Q&A, my thought process was as follows:

  1. The version that I first edited was https://gis.stackexchange.com/revisions/278270/2 and it included the key phrase:

I am using ArcMap 10.6 but also open to a Python script solution.

which I interpreted to mean that its asker was hoping for an answer from the ArcMap GUI of ArcGIS Desktop but that they would also be open to someone writing Python code for them. In effect this could be interpreted as asking two questions: How to do it using the GUI? How to do it using coding? With the emphasis being placed on the first question.

I chose not to edit out the second question, so that the keyword of Python remained there for anyone searching question bodies using "python" to find it and decide whether they wanted to write or look for Python code (and provide an answer). However, I removed the tag because various guidelines from the Help and Meta GIS SE Q&As have indicated that many/most of those helping people with coding questions want to always see a code attempt first. This is consistent with our close reason below and is very similar to how Stack Overflow operates. We model a lot of our guidelines (especially those related to programming) on theirs so that moving between the two sites is not too jarring. Had I left the tag then, to be consistent, I would have voted to place the question on hold.

Questions seeking help to debug/write/improve code must include the desired behavior, a specific problem or error and the shortest code necessary to reproduce it in the question itself. Providing a clear problem statement and evidence of a code attempt will help others to help you.

I think a number of us have been commenting for quite some time that if we see coding tags like , , , etc we expect to see a code snippet first so that we volunteers are not viewed as a free coding service. However, I am not sure that "no code => no coding tag" has been said anywhere, partly because there is some grey in its interpretation. Occasionally, a coding question does not need code first but I think such questions should be rare.

  1. You added a tag for
  2. I rolled back your addition of that tag partly because the question did not mention ArcPy (just the broader Python programming language) and mostly because it contained no code. I also commented:

@GISKid for an tag to be applicable there needs to be a code attempt in the question.

and apologize for writing that a little tersely, but I thought you would get the connection I was making to our "no code" close reason without me having to expand.

My overall view is:

  • I think a question with no code snippet, that has not yet been answered, should very rarely be tagged with a coding tag like , , etc
  • I think it is fine for someone to find a question with coding terms in its question body and answer it using code. I encourage such answers with upvotes.
  • Only when a Q&A, or just a question, or just an answer, has code in it, does it become appropriate for a coding tag like , , etc to be applied. I often add these when I see questions with code answers that lack the tag to alert the coding sub-community to them.
  • By doing the above we consistently present to site visitors that if they click on a coding tag they will almost always find code snippets that they can use and learn from.
  • 3
    You cited 'Why is the Python tag inappropriate for a question about Shapely?' and expressed concerns that the guidelines present within it were not being followed. That is a Meta GIS SE Q&A that I am acutely aware of because it is one where I modified my workflow to no longer remove python from questions about shapely and a plethora of other spatial Python libraries. With all respect, that question was about [shapely], but Luke's answer was about all python libraries. Your answers did not surpass Luke's, but you still kept removing [python] from [arcpy] and [pyqgis] questions. Apr 6 '18 at 13:56
  • My previous comment was a side note about an excerpt of this answer which I believe was not accurate. However, I agree with you that OP in the main site wanted a GUI solution (that way, the question could remain open). If OP wanted a scripting solution the most, then, a snippet/reproducible example would be required and Q would be tagged accordingly. Moreover, if someone had provided an scriping solution to the original version of Q, a code tag could then be added later. This is a gray area, but the answer's reasoning makes sense to me. Apr 7 '18 at 14:49
  • I chose compromise over popularity but I believe my answer here is entirely accurate. That choice was based partly on some since deleted comments in support of compromise and so appears weaker than it was when made. Thank you for seeing past that obfuscating issue which is discussed elsewhere and already linked to.
    – PolyGeo Mod
    Apr 7 '18 at 22:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .