Open-source advocates can sometimes be a bit sensitive of criticism of open-source software. And they can sometimes overstate the benefits of open-source over closed-source programs, either free or paid. I use lots of open-source software, and am glad it’s there, but recognize that closed source programs have their own benefits as well. So when I saw this blog post on PerryGeo complaining about an article about free GIS programs in the latest issue of American Surveyor magazine, I guessed it wasn’t as bad as he thought it was. Well, I read it, and it isn’t as bad – it’s much, much worse. I’d go so far as to call it a classic example of the Dunning-Kruger effect in action.
The article purports to be a representative survey of “simple to use” GIS, “easy to find on the Internet, readily available, easy to download and easy to install”. The software needed to be able to load in georeferenced vector and raster imagery in standard GIS formats, let you edit/create layers, do some basic analysis, and let you export the map in some reasonable format. From the many dozens of free GIS software packages available online that meet these requirements, he could only come up with four:
The selection of Google Earth for this article is a joke. I love Google Earth, but it’s not even close to a true GIS, and right from the start it should have been evident that it couldn’t perform most of the functions he wanted; evaluating it here is a waste of time and space. ArcGIS Explorer is marginally better, but for the intended uses, it doesn’t cut it. That leaves DIVA-GIS and Quantum GIS as the only two legitimate contenders. I’ve played with DIVA-GIS a bit, and it’s a good free GIS program, even beyond its primary function for analyzing geographic and environmental factors for species distribution. But to consider this program as a prominent example of what’s available in the free GIS world is nonsense.
But it’s what he does to Quantum GIS that’s the real crime. He dismisses it as “too complicated to use right out of the box”, and says that based on the documentation the program won’t do what he wants it to. The author claims to be head of GIS and mapping for a consulting company, and yet he can’t even find the “Add Vector Layer” and “Add Raster Layer” buttons prominent in QGIS’s toolbar? Can’t change the layer properties, which isn’t that much different from DIVA-GIS’s approach? Can’t check out the toolbar buttons that provide virtually all the functionality that he says he’s looking for? In the table that compares functionality between different programs, he puts question marks in most of Quantum GIS’s column, even though it wouldn’t take a “GIS expert” very long to figure out that Quantum GIS could perform most of those. DIVA-GIS is a nice piece of software, but Quantum GIS is clearly superior. Condemning Quantum GIS without a fair evaluation doesn’t serve the article’s readers very well.
The author tries to hide behind some weasel clauses. He claims that this isn’t meant to be a comprehensive review, as his rationale for covering only a few programs. I call BS on this; if you don’t have the basic background necessary to write such an article and can’t be bothered to do the research needed, then pass it on to someone who has the expertise or time to do a good job. He also claims that if someone unfamiliar with these topics can’t use the program “right out of the box”, without spending some time learning how to do things, that’s an indication that the software is too complex. Total BS; that’s like telling someone who’s used to Microsoft’s primitive word processor WordPad that Microsoft Word isn’t worth learning, because you’ll need to spend some time learning some of the more advanced functions.
What’s more, the data he uses for his example requires some level of user sophistication. They have to know the difference between raster, vector and DEM data, and understand the concepts of georeferencing for all those data types. The concepts behind the functionality he wants aren’t trivial, either; he wants software that can do buffering and network analysis, as well as information searches. This isn’t rocket science, but it isn’t intuitively obvious, either; anyone who understands these concepts has to be reasonably well-informed about GIS issues. The author’s implication is that his readers aren’t smart enough to be able to figure out a program unless it’s trivially easy to use, which is really an insult to them. But if they understand the data concepts, they’re smart enough to be able to spend a little time figuring out the program.
I’m not surprised PerryGeo is pissed off; years of work has gone into making Quantum GIS a first-rate free GIS program, but the author of this article blows it off completely without a fair evaluation. And he does his readers a disservice not just by denigrating Quantum GIS, but also by implying that he’s done a representative evaluation of free GIS options, something the article doesn’t even come close to doing. What an embarrassment.
End of rant.