From: Bryce L N. <bno...@fs...> - 2007-07-31 20:02:52
|
> Our recent experience is that using Java Collections for these kind of > ideas is a bad move. It has lulled our user community into some really > bad mistakes; hense the motivation for this proposal. Really? It seems so innocent. What happened? (Wait a minnit: I was talking about writing new implementations of Map<Geometry, ?>, etc. to accelerate storing/retrieving spatial data. Were you talking about using the current implementations? That could be bad.) > Okay if I understand you; if we ever care about finding a standard > interface for these ideas we can go look up coverage stuff. Got it. > Until then we are on our own and much happier for it. Yes and no. Discrete Coverages would be a natural first client of an efficient spatial collections suite. But there's no explicit encapsulation of spatial collections even in coverages. With respect to spatial collections, you're on your own even after you decide to go look for a standard. You should be very happy! > > Bryce "Repetitiveness is my job" Nordgren :) > > > It is more that you have to wait (sometimes years) for us to catch up. That is less irritating than waiting years for my own brain to catch on. ;) Bryce |