This is reminding me of a blog post I read yesterday (at highscalability.com via lin.ear.th.inking) about relational databases vs. the "big table" database Google uses. To attempt to summarize, RDBs rely on categorization/tables, the big table on redundancy and maintaining real-world connections. More importantly to the ongoing conversation here, they're totally different mindsets and both excellent choices in certain situations. Of course, one you can run off a single computer, the other you need a server farm... :)
Brad > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Thu, 26 Jun 2008 20:52:32 +0930 > From: stephen white <[EMAIL PROTECTED]> > Subject: Re: [Geowanking] Spatial analysis was Re: MapMaker > To: [email protected] > Message-ID: <[EMAIL PROTECTED]> > Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes > > On 26/06/2008, at 2:04 PM, Mike Liebhold wrote: >> And despite Stephen White's and Schuyler Erles earlier comments >> diminishing the importance of "red dots", besides 3D, another >> important frontier is learn true geospatial analysis; learning to form >> proper queries for the -specific- dots, lines or blobs from what is >> growing into an immense web of geocoded data. > > > I'm a bit surprised to find myself the poster (or whipping) boy for > wanting to go beyond map mash-ups. It just seems very obvious to me > that mash-ups of all kinds will always be specialised interfaces that > have, and suffer, from all the same problems as layers in GIS. > > When trying to find information, what is the difference between a > clumsy search field, a clumsy layers box, a clumsy bunch of red dots, > or basically any of the current approaches? They all end up at the > same original problem of being unable to specify what is wanted. > > There are two components to this problem. The first is being unable to > accurately, without bias, specify what is being searched for. The > second is being unable to accurately, wihtout bias, sort information > into searchable categories. > > That is the same valley of death that mash-ups dive into every time. > They always categorise information in the first place, then want the > information searched by category. Categories are the problem. > Categories are not the solution. > > Hence my push to look back towards the original data with its 3d, > time, location, and photographic capture by nature. Now given a bunch > of raw data, how CAN it be organised (not categorised) such that it > can be searched by the same means? > > I can only be the poster (or whipping) boy for one topic at a time. So > I won't expand further until the discussion is more amenable to this > specific avenue. > > -- > [EMAIL PROTECTED] > > > > > ------------------------------ > > Message: 2 > Date: Thu, 26 Jun 2008 07:47:38 -0400 > From: "Andrew Turner" <[EMAIL PROTECTED]> > Subject: Re: [Geowanking] Spatial analysis was Re: MapMaker > To: [EMAIL PROTECTED], [email protected] > Message-ID: > <[EMAIL PROTECTED]> > Content-Type: text/plain; charset=UTF-8 > > On Thu, Jun 26, 2008 at 12:48 AM, David G. Smith PE PLS > <[EMAIL PROTECTED]> wrote: >> Seems 'neo' comes full circle and returns to what the 'paleo' guys have >> already been doing for decades in desktop GIS... >> >> *ducking tomatoes and exiting stage left...* >> >> Seriously, the excitement factor of it is in bringing true, robust >> geoanalytical capability to this Mashup-oriented self-service Web x.0 >> paradigm in the form of easy-to-feed, easy-to-consume web services. >> Grab >> dataset 'a' on the fly from here, run your analysis on the fly there, >> and >> show the results in your [insert framework du jour here: Google Earth / >> Virtual Earth / WorldWind / OpenLayers / Widget / iPhone / whatever else >> we >> dream up ] >> > > Good points, but a little off as well. > > There is a mix in this discussion between Neogeography, GIS, and the > land that is betwixt the two. And in this mix is the desire by various > domains to have the utility of the others. > > The problem with GIS in the Web x.0 world hasn't been that the utility > wasn't desired, it's that it is couched in interfaces and > functionality that is targeted towards a different kind of user than > the one asking the question "find me a camping site". > > So what is interesting is that all this data we're begging to collect, > collate, and aggregate, and the innovative interfaces for visualizing > the results are that they will be potentially just as useful for > asking the question an 'average user' may ask "find me a good camp > site on my route to B" as the question a geospatial analyst may ask > "what is the density of campers along highways in the western US". > > However, the two are still distinct. There will be different engines > that power analyzing these questions and different pieces of interface > that provide exploration and querying depending on the type of answers > desired. > > Andrew > > > ------------------------------ > > _______________________________________________ > Geowanking mailing list > [email protected] > http://lists.burri.to/mailman/listinfo/geowanking > > End of Geowanking Digest, Vol 55, Issue 29 > ****************************************** > _______________________________________________ Geowanking mailing list [email protected] http://lists.burri.to/mailman/listinfo/geowanking
