As has already been noted, there are plenty of GIS questions on SO, particularly in . Some of these questions clearly belong on SO, in that they are asking for programming advice that just happens to relate to geographical data.

More frequently, however, they are asking if functionality that exists in current GIS systems can be performed in R. The most frequent such topics seem to be:

These I believe clearly belong on GIS.SE, although most still remain on SO.

Yet others fall in a grey area. For instance, this question ( http://stackoverflow.com/questions/10212956/subset-spatialpolygonsdataframe ), wherein common GIS functionality is asked for, but the answer uses base R concepts to implement a custom solution. Similarly, this question about cropping required some custom code and arguably some knowledge of S4 methods, but ultimately the question was one about fundamental GIS capabilities.

I hope with this thread we can come to a consensus on what belongs on SO versus what belongs on GIS.SE .

How much programming is required to make it a programming (and thus SO) question? How much geography is required to make it a GIS.SE question?

share

4 Answers 4

up vote 12 down vote accepted

There's a surprisingly clear distinction between pure programming questions and GIS related questions. Any question that references a GIS platform in an essential way should stay here--it would like be viewed as too specialized on SO. When I see the occasional question here (typically with Python or R code, sometimes VB) that is obviously just about programming, and could be answered by any programmer completely ignorant of GIS, I migrate it to SO.

(I have tended not to migrate Field Calculator questions that are obviously from users struggling with rudiments of Python or VB, on the theory that those users are unlikely to communicate effectively with respondents on SO and that SO readers might not appreciate the distinctions between ESRI/QGIS/etc. field calculations and full-blown code.)

The most troublesome decisions recently have concerned R code. There, I usually have to read through the code to determine whether the trouble is with R itself or, alternatively, with GIS concepts or the use of a GIS-oriented R package. Because SO has a strong R community (they have fielded about 22K questions to date and include some of R's development team), I am inclined to migrate the former questions and leave the latter. The main problem with this policy is that it can take time and effort to make the migration decision. But I think it's worth it, because I see great value in accumulating a corpus of R-based GIS solutions on this site, due to R's potential to provide an open-source GIS analysis platform that is more powerful and flexible than anything else out there.

In response to another answer here, I need to emphasize that you should not cross-post questions! SE policy strongly discourages that. It creates confusing duplication and is time-consuming to rectify, requiring communications among moderators of multiple sites, migration, editing, and merging.

share
    
Marking this as the answer for now, but I'm still interested in a clearer line in distinguishing one from the other. Great answer, though. –  Ari B. Friedman Dec 18 '12 at 22:44

I Don't think that this problem can be solved or whether we can reach a 'consensus on what belongs on SO versus what belongs on GIS.SE'.

The problem of overlap amongst many of the Stack exchange sites is an old complaint, and has been discussed hundreds of times on meta.so

It has to do a lot with the background of the person asking the question. For example, many web developers started using the Google Maps API without having any background in GIS. And naturally they asked questions on SO. There are more questions tagged Google Maps API on SO, than GIS.se.

Does that mean that these questions belong here? I am not so sure of that. Most of the members of GIS.se are experienced with ESRI and OSGeo software, and not specifically with the intricacies of the Google Maps API. But you'll find lots of answers on this topic on SO.

I think that at the end of the day, we need to ask for each question, whether there is a better home for it, which has members experienced in the topic.

It also has to do with the community already present as well as the attitude with which the question was asked. For example, there are many questions tagged with esri both stackoverflow and serverfault. If you look at them, you will see that most of them have one or two answers and very few views. If those same questions were asked her, I'm quite sure that you would get a better response on most of them.

Then there are question, which are best suited to our sister sites, even if they have to do with gis software. For example this one

So In short, there will always be an overlap, and you can always cross post the question, provided you have tailored it to the audience of that particular stack-exchange site.

share
1  
I Don't think that this problem can be solved or whether we can reach a 'consensus on what belongs on SO versus what belongs on GIS.SE'. I think we should still try though. The goal should be to keep the overlaps to a minimum. Otherwise, we might as well be part of SO. You might want to change your example, too btw. Your post links to a question about ESXi ( a virtualization product) not ESRI. –  R.K. Dec 15 '12 at 7:41
    
@R.K. Thanks for catching the incorrect link. I have corrected it to the one I wanted to link to. –  Devdatta Tengshe Dec 16 '12 at 5:16
    
If we post more GIS based programming questions here, it would draw more programmers with GIS expertise to this site, which I think is a good thing. –  awe Jan 14 '13 at 9:11

The existing answers are excellent. I would only add that, some time ago now, I flagged for migration and commented a number of questions on SO which I thought would be better served here and received pretty much a non-response; none migrated anyway.

Now my strategy is that whenever I bump into a GIS question elsewhere in the network (not really that often) I look for an existing pertinent answer here to link back to. The knowledge base is big enough now there is often something.

share
    
There is now a formal process to migrate older questions (more than 60 days) en masse, but it requires mods on both sites assembling a list, getting together, and reviewing them. Sometimes (for newer questions) it works to flag a GIS-related question and leave a note for the mods gently suggesting migration to here. At the time a mod fields that flag, it's almost as easy for them to migrate the question as to keep it, so you don't have the same barrier to migration. –  whuber Dec 19 '12 at 18:41
    
that's good to know @whuber, thanks. –  matt wilkie Dec 19 '12 at 19:38

Firstly you need to define Geography - that alone is a wide field. Geography typically encompasses Geomatics http://en.wikipedia.org/wiki/Geomatics. Here GIS SE accepts many wide-ranging questions including programming like python, some questions are more about the process or code that the GIS factor.

An Example is How to improve Python script processing speed (performance)?

Technically this is not a pure GIS.SE question - but there are many in our community that can answer this question. But when answered it does provide value to this community.

share
    
But then the site is only GIS SE so strictly speaking, we only need to define GIS. –  R.K. Dec 13 '12 at 18:51
    
Google Says "Geographic information system, a system for storing and manipulating geographical information on computer." –  Mapperz Dec 13 '12 at 18:59
    
I meant that the site is only GIS SE. Won't it be better to change the name if we want to be more inclusive? Something like Geospatial SE or something. We already cover lots of things aside from GIS anyway. –  R.K. Dec 15 '12 at 7:34
    
Non-GIS geomatic questions are closest to GIS.SE now, but there is an area51 proposal for them. –  Pavel V. Dec 15 '12 at 9:53
3  
I don't see the point of that Area51 proposal, in light of some sample questions that are squarely on topic here (re map projections and coordinate systems). All it would do is split people away from GIS.SE and further confuse everyone about where to post GIS-related questions. –  whuber Dec 18 '12 at 20:51
    
@whuber Would it be possible to change the site's name to Geospatial or something more inclusive? GIS is but a small part of the questions we answer on the site. –  R.K. Dec 19 '12 at 11:16
    
@R.K. I'd rather keep gis.se than have geospatial.se because GIS is what people will be looking for. Our FAQ should encourage those who really wonder if general geospatial questions are welcome here: "The Geographic Information Systems Stack Exchange is for questions concerning geographic information systems and science. We welcome cartographers, database administrators, geographers, programmers, and anyone interested in or using GIS." –  underdark Dec 19 '12 at 12:07
    
@underdark But a lot of questions aren't about GIS or GIScience and we consider them on-topic. I just fear that we are conflating GIS with geospatial. –  R.K. Dec 19 '12 at 12:56
2  
@R.K. That conflation was made long ago: Web searches for both terms, "geospatial" and "GIS," indicate they are intimately linked. One cannot do geospatial analysis to any great degree without GIS; and GIS ultimately is intended for some form of geospatial analysis, broadly conceived. –  whuber Dec 19 '12 at 16:21
    
@whuber They may be intimately linked but that doesn't mean that they're the same, does it? Otherwise, it would be like conflating the tool with the field. Equating spreadsheets to data analysis. –  R.K. Dec 20 '12 at 4:19
    
I don't think anybody was claiming they are the same, @R.K. To continue your analogy: if you are using a spreadsheet to do data analysis, your question is welcome on the stats site. However, the analogy quickly breaks down: spreadsheets are used in myriad settings, but it's extremely rare that a GIS is used for something other than managing, analyzing, or displaying geospatial data. My impression about the mix of this site's questions differs from yours, too: I would be hard pressed to find any questions that do not relate in some way to a GIS or to GIS-like processing. –  whuber Dec 20 '12 at 4:24
    
I guess my discomfort stems from the fact that we are using the term GIS as a catchall term for all things geospatial (e.g. remote sensing, geomatics, geodesy, cartography, GPS, to name a few). The field is so much larger than GIS. Or can I take it to mean that the site really is about about GIS but we welcome those other questions as we are the most related Stack Exchange site? –  R.K. Dec 20 '12 at 5:32
2  
@R.K. That last remark makes a lot of sense to me. Bear in mind that SE started as a programming site (StackOverflow), so all the derivative sites (like ours) bear the stamp of a programming orientation. You might be interested in an earlier discussion we had about this very subject at meta.gis.stackexchange.com/questions/383/…. –  whuber Dec 20 '12 at 14:44
    
Still a bit comfortable but I can live with that. Thanks for the patience :) –  R.K. Dec 21 '12 at 3:50

You must log in to answer this question.

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