-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
Curtis Olson wrote: | On Fri, Mar 21, 2008 at 11:00 AM, AnMaster wrote: | |> -----BEGIN PGP SIGNED MESSAGE----- |> Hash: SHA512 |> |> It does indeed look like 0.9.8 had best coast line. Why is it so much |> worse in more recent |> scenery? Wouldn't it be possible to get the same good coastline as in |> 0.9.8? | | | As with anything, it's not quite that simple. The GHSSH coastline database | (used in 0.9.8) is indeed more accurate for ocean coastlines. However, for | any inland fresh water, it is much worse/less detailed than vmap0. So in | 0.9.8 we used GHSSH for coastlines and VMAP0 for lakes & rivers. | | However, we found that there were major discrepancies between the two | databases. For instance, lake washington near seattle is classified as lake | by GHSSH and ocean by vmap0, so it simply did not appear in 0.9.8. | Similarly there were things that were classified as ocean by GHSSH and lake | by vmap0 so they were both drawn on top of each other. | | The decision was made to go with vmap0 entirely. We gave up accuracy around | the coast lines, but we gained a much more consistent picture of the world | ... with no major missing bits and no overlapping sections. | | Originally we only used GHSSH defined waterways, but this let to major | problems, like the great lakes only being half defined (i.e. the canadian | side of the lakes was not mapped so there was only land there.) And | strangely, some Canadian's complained about that! :-) So we went to the | mixed GHSSH-Ocean/VMAP0-fresh-water scheme, but people complained about the | missing water that fell in the gaps between both dataset classifications. | So then we went to vmap0, but now people are complaining about less detailed | coast lines! | | What combination should we try next? :-) 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? Regards, Arvid Norlander -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.7 (GNU/Linux) iD8DBQFH4+bmWmK6ng/aMNkRCrPAAJ9cnNn819/rhieMAOLM/aRP/VXc4QCgkHPA dW6e+qXIMwMmKnKKZoShDJE= =wyoA -----END PGP SIGNATURE----- ------------------------------------------------------------------------- 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