> geographic information" get any airtime? Why does the average video game > toss around 100k polygons phong shaded lit polygons at 60 fps while most GIS > clients struggle to show even 10k lines at 1 fps?
*Shrug* Game software sacrifices portability for performance. Map software often sacrifices performance for portability. Also, helping people be useful is less lucrative than helping people blow time, which effects how hacker resources are allocated. So there's that. I see your point, though. Why do we keep reinventing the wheel? I think that's actually a very interesting question, and I think an answer would have to address the ideal level of complexity in a public library. This is what I can figure: implementing a 2-D cartographic map feature, in any programming language, from the native features of that language, is easier than understanding the tangle loosely related, poorly-branded everything-to-everyone geospatial libraries available. I have a choice at the outset. On the one hand, I can think in terms of the set of abstractions that you or your community has chosen to represent the problem space. I can learn to _think like you_. That seems easy to you, but it's actually hard for me. Or, I can think in terms of Cartesian mathematics. Pythagorean theorem and whatnot. That stuff gets pounded into you in the fifth grade. It's _cannon_. For many, thinking in terms of the cannon is more attractive than thinking like you. So, for many, reimplementing from scratch will be faster and more fun. How could we overcome this? A new user's first introduction to a system of abstraction needs to be excruciatingly gentle and non-abstract. Right: http://www.feedparser.org/. Wrong: http://www.gdal.org/gdal_tutorial.html. This may seem off topic, but it fits well with Anselm's point that insularity is causing a lot of duplicated work on both sides of the insider/outsider boundary. > > Projects like tonchidot, work in ambient computing and augmented reality > coming out of university research labs, or google, and random hacker teams - > and seem to in part be re-inventing the wheel. They have their own formats, > they seem to emerge full cloth with no history, they act as if nothing else > exists. Why aren't more crazy new ideas coming from (or supported in part > by) established players such as say MetaCarta or ESRI or um, even (although > perhaps not strictly fair because we've seen a fair degree of innovation > here) from Poly9 or Urban Mapping? My bad. Sorry. See you in Portland, B _______________________________________________ Geowanking mailing list [email protected] http://lists.burri.to/mailman/listinfo/geowanking
