On Friday 21 March 2008 17:13, Curtis Olson wrote: > On Fri, Mar 21, 2008 at 12:07 PM, Ralf Gerlich wrote: > > AnMaster wrote: > > > Good question, I guess combining them and manually fixing the > > > problems > > > > would be too much > > > > > work. I got no really good solution. But the current > > > coastlines are very > > > > bad in many cases. > > > > > What about only using GHSSH for those coastlines around > > > continents? With > > > > that I mean coast > > > > > line around, say, North and south America, > > > Europe/Africa/Asia (that, > > > > apart from the Suez > > > > > channel, are connected), Australia and any islands, and > > > simply discard > > > > any coastlines > > > > > inside these "blocks" and use vmap0 there. That is: don't > > > trust how > > > > vmap0/GHSSH classify > > > > > the data. Would that be feasible? > > > > Feasible, as GSHHS explicitly makes the outer coastlines > > available and differentiates them from inner shorelines, but it > > wouldn't solve the problems with inconsistent waterways at the > > coastlines of continents. > > > > Even though that is a lot of work, manually adapting our > > VMAP0-based data to the GSHHS-data is the only solution I > > currently see. > > Right, this is basically what we did for 0.9.8 and ended up with > a bazillion inconsistancies ... > > Areas marked as lake/river in GSHHS but ocean in vmap0 will be > entirely skipped. > Areas marked as ocean in GSHHS and lake/river in vmap0 will be > doubled up and overlapped. > VMAP0 rivers may run short of the GSHHS coastline. > etc. > > There's no combination of these two datasets you can do perfectly > with an automated system. You would need a tremendous amount of > effort to visually inspect the entire data set and resolve any > problems manually. > > Regards, > > Curt.
So it looks like we either live with the problem until someone else creates a new database with all the problems fixed, or bite the bullet and fix it manually ourselves. Would it be possible to cobble together a small utility that would allow small parcels of the scenery database e.g. 1x1 deg tiles, to be checked and corrected manually without setting up the full scenery build system? That way, many people could work on it whenever they feel like it. I've had to do this sort of manual data correlation/verification a couple of times over the years, on different projects I've worked on, and while it sounds like an onerous and tedious task it's not too bad if you can just do bits of it, now and then, as a break from your primary tasks. Obviously, it would be a slow and long-term undertaking but at least the problem would be fixed eventually, whereas the alternative seems to be that it never gets fixed and we just have to live with it. LeeE ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
