I'm writing about a question I asked that is similar to another question, but not identical, and possibly very different.

It was closed under claims of "already asked", "already answered", and "exact duplicate" but none of these three claims are actually true.

I asked the opposite question (SHP->KML) and the original question asked KML->SHP. While someone did answer how to do SHP->KML question under KML->SHP question, it was answered in a comment to an answer, not a direct answer. As such, the answer is at a 3rd tier level (Question, Answer, Comment) of a question which is the exact opposite of my question.

Based on this, it does not seem reasonable to close my question. An analogy for me is someone asking a question on randomization and the question being closed because a question on sorting has already been asked.

I think it is also relevant that Google does distinguish on word order, so someone searching for "How to I convert shapefiles to KML" might not find the question on converting KML to shapefiles, and even if they did, they might not notice the answer since it is, essentially, hidden in a comment.


I agree with you and do feel that it was closed in haste.

Your Question may sound very similar to that other question, and those who use ogr2ogr know that the answers are similar, but Others might not know that.

When somebody will come from google searching for SHP->KML, they would not necessarily read all the comments on a a question which deals with KML->SHP.

Some might feel that we are increasing noise by allowing 'inverse questions', but I think that since this site is for 'one question one answer' and not for discussions, we should allow these inverse questions.


Inverse questions should be evaluated on a case by case basis by the community. If you feel the inverse workflow is significantly different than the original, please clarify that in the question itself.

In this case, I agree with the closure. For shapefile-->KML-->shapefile operations, typical workflows are the same in both directions. For this question, separate questions would be redundant and splinter any novel answers (and their discovery) away from an existing thread.

  • I agree with this. There are likely cases where the questions have considerably different answers, but I really don't like the idea of having two "List of X" questions with a high degree of overlap. Unless there are references between them, I believe that this is doing a disservice to future searchers. Ideally, the original question could be modified to "How do I convert between SHP and KML", and any unique answers merged.
    – JasonBirch
    Aug 29 '10 at 23:21
  • 1
    I'd feel much more comfortable about this if the original question were modified to be more generic "to/from SHP and KML", and if the SHP->KML answer by @glennon were promoted from a comment to an actual answer.
    – mankoff
    Aug 29 '10 at 23:34
  • +1 for the importance of describing why Question X is not the same as Question Y. Aug 30 '10 at 22:12

From When is a duplicate question not a duplicate?, "Rule of thumb: If you ask a question similar to another question and it is likely to get the exact same answer, you have yourself a duplicate question.".

And some fine tuning in At what point does a question become a duplicate?, namely: that dupe-or-not? a bit touchy feely and don't forget that "There’s often benefit to having multiple subtle variants of a question around, as people tend to ask and search using completely different words, and the better our coverage, the better odds our fellow programmers can find the answer they’re looking for."

In my opinion this case is a dupe, and those searching with different key words (e.g. shape to kml) may land on this Q but will find the real answer in short order; the comment describing shp to kml in the related question is above the fold.

If using ogr2ogr is not a solution for you, then please expand your question so people can understand why you need a different answer, and we'll re-open the question.


In this case it is a conversion between two known, open formats that represent equivalent data.

Conversely, if you were converting from SHP to PDF, or from TIF to MrSID, or from flour to bread, then an inverse question does belong as a standalone, since it's not a transitive conversion. Performing the reverse "conversion" (transformation) is either a different workflow, or may not even be possible (bread to flour).

I echo glennon's sentiment - if you find that an existing question / answer does not help your situation, say so! Links between questions are always helpful. Personally, I would rather invest in a little research (clicking through questions) than expect that I get the right answer, first time, every time.


Maybe we should add a "data-conversion" tag. That way, with 3 tags ( shapefile, kml, and data-conversion) both shapefile->kml and kml->shapefile would show up.

More generally "how do I convert data from x to y?" would follow the same 3 tag pattern.

Given how much time GIS professionals spend converting data between different formats, I think this is an important issue.


You must log in to answer this question.

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