3

This question is triggered by comments on the question Measuring tool shows different measurements for same building in two different map documents?; and is:

Why is it important to have the software (often version too) details included when asking most (but not all) questions at GIS SE?

In your answers to the above question, please make clear whether you agree (it is important) or disagree (it is NOT important).

The case I will make below is that I think it is very important.

My reasoning is underpinned by this answer to Framing (asking) good Questions for GIS Stack Exchange?, and is purely related to what goes into the body of a question - titles and tags are two ways to summarize the body and I think it will be better for them not to be drawn into this discussion, and thereby risk muddying it.

I am a GIS Professional without high school, undergraduate or postgraduate qualifications in GIS or Geography. However, as someone with undergraduate and postgraduate qualifications in Zoology, I am extremely interested in GIS as a tool. Conversely, my interest in formal learning of Geography (and Surveying) topics is minimal. Nevertheless, by getting involved in Q&As on Main, I often learn about aspects of them in bite size pieces, which I like to do.

Irrespective of whether I am reading a question with a view to answering it as a GIS Professional proficient in its subject matter, or as someone hoping to try and understand the question (and possibly attempting an answer), my approach is a "hands-on" one, and I start to think about how I can test what is being asked.

If I am going to test, then knowing the software used by the asker means that I do not have to guess that detail, and thus risk wasting my time and the dissatisfaction of downvotes, by writing an answer to an unspecified requirement.

Knowing the version is not as important, but GIS software evolves quickly, so that detail often provides clues as to the constraints the asker is working within, and any constraints I need to place on how I can answer.

For example, using ArcGIS for Desktop, because it is the GIS product with which I am more familiar:

  • we have had over 2,000 questions tagged and at ArcGIS 10.1 for Desktop many new ArcPy capabilities were introduced (arcpy.da, Python toolboxes, Python AddIns, etc) so a pre-10.1 ArcPy question is often answered very differently to one at 10.1+.
  • we have had over 6,000 questions tagged arcgis-desktop or arcmap and at ArcGIS Desktop 10.0 the ArcMap GUI was overhauled so an ArcMap 9.x question is often answered very differently to one at 10.0+.
  • we have had over 625 questions tagged and at ArcGIS Desktop 10.0 many new ModelBuilder capabilities were introduced (Model Only tools, Iterators) so a pre-10.0 ModelBuilder question is often answered very differently to one at 10.0+.

There are a significant number of questions for which the GIS software being used appears irrelevant to me, and there are many other questions which I have no interest in answering, so for those I do NOT seek clarification of software being used via a comment.

However, when I am attracted to a question, the first detail I take in before deciding whether to try and answer it is whether I can test. If I am going to test I need to know whether the question uses software I have access to, and it is often also helpful for me to know the version of that software.

  • 4
    I do not think that anyone would disagree with your final conclusions that when a question originates from a software problem, it is helpful to know what software it is. What I have objected to in some of your comments is the apparent insistence that software be referenced even when the problem clearly is not one of software usage but of understanding the principles of GIS. In that case it would be misleading for the OP to name any kind of software, as if their concern were about running the software. Let the questioner decide whether they are asking about the software or not. – whuber Sep 19 '14 at 18:27
  • What may be clear to you in regard to the principles of GIS is not necessarily so to me. The decision on whether a questioner is asking about software is, of course, theirs, which is why I use comments to clarify if that is the case when I am unsure. Perhaps we can use a gis-principle tag to try and recognize that question class in a similar way to how we use the algorithm tag. – PolyGeo Sep 19 '14 at 22:15
  • 1
    FWIW I agree with @whuber in this particular case - tagging that question with software could imply that it was an issue with the software, when in fact it's a GIS fundamental. But taking your point that you'd approach this experimentally, you could attempt to replicate the problem using whatever GIS was at hand. Doing so would lead to the conclusion that it was software-independent. – Stephen Lead Sep 22 '14 at 3:44
  • @StephenLead Yeah - on that particular question I was trying to point out why I asked - when I was still coming to grips with coordinate systems it was a mixture of the theory and hands on that seemed to make it click. – PolyGeo Sep 22 '14 at 4:00
  • I just posted a question where I've deliberately left off the software tags, since the answer should be the same regardless of the software used.... – Stephen Lead Sep 26 '14 at 4:14
  • @StephenLead I'll be interested to see if any potential answerer asks you to clarify whether you just forgot to say what you were using, and whether having no software tags appears to narrow or expand the pool of potential answerers. – PolyGeo Sep 26 '14 at 4:35
  • @PolyGeo me, too! – Stephen Lead Sep 26 '14 at 4:39
  • @StephenLead I just came across your comment before the one above. I notice that there is only one unaccepted answer on that question so am curious about your thoughts on software-agnostic questions nowadays? – PolyGeo Sep 9 '17 at 4:21
5

I agree that when a question involves software, it is critical that the version numbers of said software are included. As with any technical support procedure, this is among the first and most important pieces of information needed to accurately answer such questions, which is why I suggested including it in my response the aforementioned 'How to frame...' question.

The main reason for this is features/bugs which exist at one version but not another, but also smaller changes like different UI layouts and options can result in confusion if they are different between versions.

As to whether this information should be included in the tags vs. the body of the question I think I am agreeing with you when I say that should be a separate question (e.g. Using version tags?), and one which we seem to have already decided to handle differently based on the software suite (e.g. QGIS vs. ArcGIS).

  • 2
    I agree with your reasoning. But the main thrust of this thread isn't whether version numbers should or should not be included, but whether the software should be mentioned at all. – whuber Sep 19 '14 at 18:22
5

For most questions on GIS.SE I would say that naming the software is important/critical but the version is not.

In almost any case where a question is 'how do I...', the answer will vary (depending on how explicit you are) with software being used. SE strives for specific questions and answers. Any question that doesn't name software can only elicit general or speculative answers. For example - 'I have two shapefiles and I want to compare x spatial relationship.' Answer: 'You need to do an overlay/intersect' vs. 'you need to do overlay/intersect by doing x y z'. I would say that most SE questions are about how to do/fix something in or with access to a specific software. A smaller fraction are either 'what software can do x' or more general, concept-based questions that are unrelated to software.

Your example of ArcPy opens the door to a broad exception category, but from what I've seen so far most questions are not impacted by software version. However, since we're specifically discussing the body of the question and not tags or title, I see no harm in including it. In fact, as blah238 points out it's often one of the first things needed or asked in those questions which already are naming a specific software. As we've discussed elsewhere, it's often difficult to know whether version is relevant to a question or not, so better to include it in the body when asking than not.


If this question were rephrased a bit to ask should both be included in the question body whenever possible, I would say yes and I agree. Importance is both subjective and varies on a very case-by-case basis in terms of questions. A general guideline to just include them, however, is a lot more rational. There's no downside I can think of (space constraints, a version in title implying answers will be specific to that version even if the question isn't) to including both. And it's such a small bit of additional info to include in the body, while it might not be relevant/necessary/important that's less an issue if it's there and doesn't need to be than the other way around.

And just for the record, my opinion does differ for tags and titles, but that's a different question.

  • I have just watered down the version:product ratio in my question title and bolding but added some more examples of when knowing the version has been necessary to answer appropriately. We also have another one of these "phase transitions" coming when 10.3 is released soon. – PolyGeo Sep 18 '14 at 4:17
  • 1
    You write as if "general ... answers" were somehow inferior or not acceptable. The OP points out, though, that "GIS software evolves quickly." There seems to be an inherent contradiction here: if the software will soon be out of date, then why shouldn't we be interested in only general answers? In which case, naming specific software--either in the body or the tags--would usually be counterproductive because it would falsely suggest that the question (and therefore, likely, the answer) would have only temporally limited interest. – whuber Sep 19 '14 at 18:21
  • 1
    @whuber No, the answer should match the question. I think we are all making generalizations and/or have different perspectives about types/quantities of Q's, and perhaps referring to "most" clouds the discussion. As quoted I was talking about 'how do I' Q's where general answers often get few or no upvotes, downvoted, flagged for brevity, or followed up with 'yes but how do I do that' comments. Are you suggesting all answers be software agnostic and let the asker figure out how to apply the concepts? I think fast evolution and temporal limits more relate to version than software. – Chris W Sep 19 '14 at 22:05
  • It is probably also relevant to mention that questions will still be asked on versions earlier than the current one for many years to come. Some of those I can answer without testing if I know the version being used. – PolyGeo Sep 20 '14 at 1:01
1

It is only important when the OP wants to accomplish some kind of analysis in a specific software.

On the other hand, there are questions (and they are not just "some", but many) which the software is not important and the OP is opened to suggestions, because it'll be easier to get answers.

For example, the tag and many questions under it just asks for ways to get things done, without necessarily demanding a specific software.

TL;DR

If it's obvious a question is demanding for a specific platform solution (and/or version), then it should be tagged with it.

Otherwise, asking the OP to demand a software is counterproductive, because having an answer accepted is not so important as having a full and diverse repository of answers involving multi-platforms.

  • I'm not sure that lidar is a good example. I just reviewed its last 10 questions, none of which have been edited since they were originally asked, and counted two as having no software constraint, two as not being overly specific (Open Source and Python), and the others specifying software (2xQGIS,LAStools,SAGA,QTM,ArcGIS for Desktop). My advocacy is to learn whether there is an unspecified software requirement on questions I am contemplating providing an answer to, and is not to demand that there be one on every question. – PolyGeo Sep 18 '14 at 22:18
  • 3
    @PolyGeo, thanks for the edit. Some questions that are tagged with the software do not necessarily mean the answer has to be on it. Many OPs express they are open to other suggestions, meaning the software tags will not be the most important. What you advocate is correct. The key point of my answer is that I think we are already doing a good job with tagging, flagging, asking the OP to clarify, etc. – Andre Silva Sep 19 '14 at 1:48
  • 1
    I could not agree more with the last sentence of your last comment. Poorly defined questions are coming in but are being quickly triaged via the community review queue, flagging, tagging, commenting, etc. – PolyGeo Sep 19 '14 at 8:42

You must log in to answer this question.

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