Well, this issue is hoot,smooooooking to be precise. Well thats plagiarised from Jeff Koinange. For those not in the know, this is a local presenter here in Kenya with his own shenanigans(hope thats the spelling).
Well on to the matter at hand. Regarding on which side i am on, of course. I am on the NoGIS side. First, for those not in the know what the talk has been about regarding NoGIS, well it is largely borrowed from NoSQL. And NoSQL being the use of non-relational datastores rather than the typical relational. Now most of the databases we currently use are relational in nature. And thats why database engineers are crucial to any project. Database architecture. Well i do remeber the properties of relational databases to have the ACID properties. Now the main draw back or relational databases was scaling. In terms of scaling the queries take longer. Sure? Can i prove this? Well this is material for my next blog post perhaps? On how geodatasets scale on relational databases vis-a vis- non relational.?
Well this is an issue i will be looking at closely in series. I think the name could be the reason behind all the acrimony. I try to imagine having a database with features of 2,000,000. I think the database querying capabilities will slow down. Here is where NoGIS wins. Though there are different line’s of thought on this, where some say NoGIS is more to deal with place while GIS is more to deal with location. So this leaves me puzzled too and still trying to get my head round this. But pleased to know that NoGIS started as a joke. On this issue there are two main schools of thought, NeoGeography and Paleogeography. The main issue being non relational databases cannot be used to represent traditional maps. Well this is true. Spatial is no longer special. And now we are seeing more platforms integrate location information. This kind of renders the traditional geographers(the kind that insist on raster data pun intended) that can and should not be referred to as GIS. The problem should be how do we solve traditional GIS demands which are very heavy without having to sacrifice some core GIS functionality?
Having a look at non-relational libraries. MongoDB has a spatial index that one has to add. On the couchdb front there is a library. There is also non-relational hosting on the cloud by the name iris. Forget the url will post here later. Anyway, i believe location information should work hand in hand with traditional GIS. They should be able to supplement one another but i am of the opinion that NoGIS or neogeography is more relevant today than ever.
Back to NoGIS related geekery…