8

This was originally posted as an answer in the How to suggest tag synonyms? question, but given my uncertainty in what synonym to suggest and the nature of their relationship, comment suggested it might be better as its own question.

I just noticed a new pair of candidates for tag synonyms that I'm not sure how to handle or if anything at all should be done:

and

One came first, one's American, one's Esri... Many search results list one first with the other in parentheses after. Different softwares call the tool that does the same thing (I think) one or the other. In terms of GIS and the questions, I'm not really seeing a difference between the two (please, correct me if I'm wrong) so it seems silly to have both separately. The wording in the Wikipedia article on Voronoi diagram would seem to suggest that Voronoi is more generic and Thiessen more specific to geophysics/spatial contexts - for whatever that's worth.

I'm wondering if maybe might be the answer to which both get made a synonym.

Then again, there's also the term Delaunay Triangulation, which currently has no tag.


Edit

A related technical discussion on GIS SE is the following:

What is the difference between Create Thiessen Polygons tool (ArcGIS) and Voronoi Polygons tool (QGIS)?

  • 1
    Is the following approach reasonable? If we ask 2 questions: 1- Are voronoi and thiessen polygons designed with the same technique? If yes, which researcher came up first with the idea? If both theorems were published independently at more or less the same time, I'd go with the tag thiessen-voronoi or voronoi-thiessen to give the proper credit. This tag could be used together with the tag 'polygon' so the former would not have a big name. If the answer to the first question is "no", then, they should be separate tags. I'm just brainstorming here, do not know the answer. – Andre Silva Jul 9 '14 at 23:37
  • 2
    @AndreSilva That would be a logical approach, but I can't really speak to validity/appropriateness as I'm not knowledgable about the history beyond what I see in a couple of textbooks and Wikipedia. There are actually other names as well, as far back as Descartes, and other names actually used the method before either of them formalized something. A lot of 'who was first' seems to be related to application of the method. The Wikipedia article delves into this in the History section, as well as a paragraph in the linked Thiessen article. – Chris W Jul 10 '14 at 0:20
  • 1
    Veronoi polygons are named after Georgy Voronoy, a Russian mathematician, who lived 1868 - 1908. Thiessen polygons are named after Alfred H. Thiessen, an American meteorologist, who lived 1872 - 1956. Both people created the same idea and method independently and were not aware of each others works. Their discovery didn't become important until decades after their demise... thus Voronoy - Thiessen is the same method... I support PolyGeos' approach of a dual tag as both names are just as important. Delaunay is what they are derived from (now) and is a totally different kettle of fish. – Michael Stimson Jul 17 '14 at 0:47
  • 1
    A bit more info on the history that I touched on in my answer: The original Voronoi paper was published in 1908, the original Thiessen paper in 1912. They were independent discoveries, as Michael notes. Though their usage dates back to Descartes, it appears that these two papers were the first formal published definitions of Voronoi/Thiessen polygons. – Conor Jul 22 '14 at 19:36
  • 1
    I just implemented the tag synonynms from the answer that you accepted, which seemed to echo broad (albeit not total) agreement in this Q&A. – PolyGeo Sep 16 '14 at 22:40
  • 1
    @PolyGeo Thanks for creating the new tag and synonyms. I note that according to this meta question that the old tags are supposed to be edited out of questions where used and replaced with the new one. While that can be done one at a time with a small number (as I recently did for the 9 'map-resolution' questions) in this case there's a total of 33 or so. The link above also mentions a tag merge tool which I believe you as a mod can use to change all those old tags into the new one at once without individual edits (and bumping). – Chris W Sep 18 '14 at 1:27
  • 1
    @ChrisW I contemplated a merge but that would "weld" both of the individual tags to the new composite one and there would be no way to recognize the original tags later if our policy (which had some dissent here) were to change. I normally prefer to do a merge in "synonymy cases" but in this instance I think the case for not doing it, and not retagging, is more compelling. – PolyGeo Sep 18 '14 at 1:38
7

I am revising my Answer after reading that of @Conor.

I now favour on the grounds that a Google search on either "voronoi" or "thiessen" would find all our Q&As on either.

I think should be the master of:

I think the creation of voronoi-thiessen-polygons or thiessen-voronoi-polygons would be too cumbersome and I do not think we need to use the (meta-tag?) with the proposed .

If a delaunay tag pops up at a later date then I think it should be discussed separately to determine any appropriate synonymy. Although related to voronoi it seems more closely to .

  • 1
    To clarify, you're ok with either becoming master but you do think one should, and you lean to thiessen? I agree the three words is cumbersome (maybe drop polygon?) but I'm not seeing how even that specifically goes against anything in the naming conventions page you linked to. It would be counter to the concept of synonyms (if they truly mean the same thing), where to follow that we should pick one or the other. I'm just concerned about those who might know one but not the other (a concern for which there are many arguments against/to alleviate, I grant you). – Chris W Jul 9 '14 at 22:29
  • 1
    Gotcha. It does seem if I see both written at once, it seems to be by someone who knows people might call it either and wants to cover all bases. This also seems to relate a bit to the command specific tag question I asked. In QGIS it's Voronoi; in ArcGIS it's Thiessen. Much like Define Projection / Assign and Project / Transform (Arc/Post). – Chris W Jul 9 '14 at 23:48
  • 1
    @ChrisW My answer has been heavily revised based on comments and Conor's answer – PolyGeo Jul 11 '14 at 8:05
  • 1
    I like your suggestion and your comments on leaving "polygons" off of the end of the tag. With voronoi/thiessen, it's implied that the end result is a set of polygons and as such there's no need to include polygon in the tag. Provided google indexing works in the way you are implying I support this tag merger as master suggestion. I would also suggest that if we adopted this method that we clean up the 30 or so existing questions to follow this same paradigm. – Conor Jul 11 '14 at 17:10
  • 1
    If it were possible to use a fore or back slash between the two I would suggest that, because it more clearly indicates they are two separate but alternative terms (rather than hyphen and implying a single, two-word term). But slash isn't and wouldn't fit convention anyway, so... And yeah, that polygon tag has always bugged me too - it doesn't really tell you anything by itself imho. – Chris W Jul 11 '14 at 19:16
  • 1
    I'm no authority on tag synonyms except for what i think seems to be "usual" behavior when two existing tags have same/similar meaning: someone with appropriate privileges makes them synonyms, and that's that. Nobody makes another composite tag via hyphens out of synonyms! – Martin F Jul 11 '14 at 21:57
  • 1
    Tag creation has few "rules" - see gis.stackexchange.com/help/privileges/create-tags - I am only supporting a "composite" one here because I think a good case has been made for us favouring one over the other being likely to lose a significant proportion of incoming searches in their space. – PolyGeo Jul 11 '14 at 23:13
  • 1
    @martinf You're correct, as I said above. But as we're discussing this particular case isn't very clear-cut (at least to me). I tried to find a similar example, but was unable to do so. Most synonyms correct spelling, acronyms, pluralization, etc. Very few actually change words, like 'extents' and 'bounding-box' or 'internationalization' and 'locale'. Here we've got two proper nouns for what amounts to the same thing, with no clear evidence one is more widespread or well-known (as relates to GIS) than the other. – Chris W Jul 12 '14 at 0:11
  • 1
    @martinf The "someone with appropriate privileges makes them synonyms" = Community Moderators (we have five). Those privileges do not extend to trusted user (>20K level). Synonyms can be voted for but achieving synonynmy via voting is very hard - see the Q part of meta.gis.stackexchange.com/questions/3077/… – PolyGeo Jul 12 '14 at 0:18
5

We should tag these questions with both Voronoi and Thiessen. We should not make these tags master/synonyms or combine them for the following reasons:

  1. Which of the two terms is more ubiquitous is highly dependent upon the field - In hydrology/meteorology, Thiessen is almost exclusively used (Alfred Thiessen was a meteorologist). In other fields such as mine (computational geometry) they are almost exclusively referred to as Voronoi Diagrams and Voronoi Polygons.

  2. It wouldn't make sense to combine them, as they are never really referred to as Voronoi/Thiessen polygons. As it is, people who are looking for help on GIS.SE may refer to them as one or the other depending not only on what program they are using (QGIS vs ArcGIS) but also their field of work.

  3. By keeping the tags separate we can allow outside users from search engines to use either/or term and still pull up results from either question that may be applicable to all forms of GIS, as the tagged question does not always specifically refer to the "Voronoi Polygons" QGIS tool or the "Create Thiessen Polygons" ArcGIS tool. If it was possible to "link" the two as synonyms without creating a master tag I would advocate for this.


Some other things to consider:

  • You are correct regarding the generality of the two usages. The Voronoi Diagrams is a more general term that was described in abstract terms by the mathematician Georgy Voronoi. In Thiessen's paper, the diagrams refer only to point-within-catchment areas. However since then usage of the term Thiessen Polygon/Diagram has evolved so that the terms are pretty much synonymous.

  • Some may argue that it is extra maintenance on our part to add the other tag to Voronoi/Thiessen for every one of these questions that comes up but seeing as how there are only 33 total tagged questions of either at the time of this post it seems a moot point.

  • A Delunay Triangulation is dual to Voronoi/Thiessen polygons but is not exactly the same concept - therefore it should be awarded its own non-synonymous tag when questions surface.

  • Voronoi/Thiessen Diagrams are also known as Dirilecht Domains but this usage is not as common in GIS due to the choice of nomenclature of the tools as mentioned above. Thus we probably should avoid making a tag of this as its usage in GIS is probably too low to warrant one at this time.

  • 2
    +1 I agree that any ArcGIS/QGIS preferences should be irrelevant to the choice of tag name here – PolyGeo Jul 11 '14 at 8:06
  • 2
    Regarding 3, search engines don't rely (at least not solely, if at all) on tags. Either term returns questions that contain that term, regardless of tag. If tags are used in searchs the merged tag might generate more results, since it will pull answers using the other term which may not be present in question or answer, as you point out. Separate tags would solve this, but with a limit of 5, using 2 tags describing the same thing seems inefficient. I agree they are synonyms but a special case where neither seems right as master - hence the question and merged tag suggestion. – Chris W Jul 11 '14 at 19:09
  • Isn't the purpose of the tag synonym mechanism supposed obviate using both synonyms for a question? – Martin F Jul 11 '14 at 22:01
  • 1
    @martinf unfortunately with the current system you have to choose a master tag when choosing a synonym tag. Since neither of these two terms should logically be the master (they are pretty much equally ubiquitous depending on field, as referenced in the post), it doesn't make sense to "promote" one of the tags by making the other the synonym. This is why PolyGeo has proposed the combined tag. It would be the master while the other two would be synonyms, and they would defer to the master tag when they are attached to questions. – Conor Jul 17 '14 at 1:49
3

Delaunay is different from Voronoi/Thiessen. It is based on triangles only. In fact, Delaunay is more closely related to TIN.

  • I was about to say that you are right when I found en.wikipedia.org/wiki/… - in any event I am going to adjust my answer to reflect that Delaunay may be better handled in a different discussion. – PolyGeo Jul 11 '14 at 23:04
0

I think a committee should make a decision on this, which contains both QGIS and ArcGIS (Geomedia, Mapinfo etc.) specialists.

Or an admin should create a poll here for all the gis.stackexchange members who reached a given reputation (1k for example).

Probably Voronoi would win, which I think the winner should be, since Voronoi published it earlier:

His best known work (1911) dealt with the description of weather prediction with a geometric method for dividing land areas, that although known from Dirichlet Tessellation (1850) and the Voronoi Diagram (1908), apparently had never been used in meteorology for interpolation of measurements.

I really don't know why ArcGIS uses this Thiessen Polygon.

-2

I know this is not the answer for this question but if there is anyone looking for an answer to "What is the difference?" question, I believe that person should get a simple answer here.

Thiessen Polygons and Voronoi Diagrams are pretty much the same for calculation in a GIS application. Voronoi lines are drawn by using perpendicular lines from all three edges of Delaunay Triangles. The intersection of three lines of a triangle will be a new node. In the end, these nodes create polygon shapes which are Voronoi Polygons. In order to be called as Thiessen Polygons, the value of the Delaunay node itself (for example precipitation observation station as a node and precipitation as its observed value) assumed as valid at every location inside of its own Voronoi Polygon.

Explaining this issue with a picture will be much more understandable I think. In the picture below, the red lines show Voronoi and black lines show Delaunay Triangles:

Difference between Voronoi and Delaunay Triangulation

You can get additional information from:

http://www.geom.uiuc.edu/~samuelp/del_project.html

  • 1
    This answer would have value if the question was about the differences, but that question would appear on Main rather than Meta. The question here is about "one tag or two" which this answer does not address. – PolyGeo Mar 17 '16 at 11:22

You must log in to answer this question.

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