Talk:Philippines/Coastline corrections

From OpenStreetMap Wiki
Jump to: navigation, search

Next 31-60 biggest islands

Will be posted after the first 11-30 is complete

  1. Samal - done
  2. Camiguin (Province) - done
  3. Panaon - done?
  4. Lubang - done?
  5. Olutanga
  6. Calayan
  7. Alabat
  8. Camiguin (Island in Cagayan)
  9. Bucas Grande
  10. Bugsuk
  11. Bantayan - done?
  12. Homonhon - 95%
  13. Linapacan - done
  14. Sibutu - done?
  15. Poro
  16. Pacjan
  17. Fuga
  18. Pangutaran
  19. Darman
  20. Panglao - done
  21. Batan
  22. Patnanongan
  23. Itbayat
  24. Romblon - done
  25. Caburrayan
  26. Siasi
  27. Tubattaha - (is this the reef?)
  28. Ilin
  29. Coron - done
  30. Babuyan
Hmmm, based on the progress, it doesn't seem to be useful to post these islands. Maintaining the status of these next 30 islands is much more work than actually fixing those islands. :-) --seav 06:26, 24 July 2010 (UTC)

Origin of the number 7,107

Speaking of islands; where is that original list that determined that the Philippines has exactly 7,107 islands? I assume that such a list was constructed by the Americans when they created maps of the colonial Philippine Islands. --seav 10:16, 3 June 2010 (UTC)

Based on my own database, we have approximately 500++ islands with > 100 hectares. Then 1800++ between 1-100 hectares. The rest are probably just rocks. The GNS data listed 3,800 island points (a lot are duplicates). I really don't know where the 7,107 number came from. Maning 10:36, 3 June 2010 (UTC)
I guess I really have to go to NAMRIA since they are the best people to answer the provenance of the 7,107 number. They probably inherited all the military maps from the American period. I want to create a complementary project to re-catalogue all the Philippines' islands again. For instance, Cebu City's Kawit Island got swallowed up by the reclamation project so we technically have 7,106 islands (or so), but Island Cove in Cavite reclaimed land and created a new island so are we back to 7,107? :-P --seav 02:24, 4 June 2010 (UTC)

Automatically determine where the sawtooth coasts are

I've thought about this way before but I've never had the time to code it. A better way to determine which coasts still need de-sawtooth-ing is to display a map (a la Coastline Checker) pointing out where there are still sawtooth coasts. It's quite easy to determine where they are: just check for three or more consecutive nodes in a natural=coastline way where each adjacent pair of nodes have the same latitude or longitude. Showing a map that highlights these coasts is better than zooming in and hunting for those jaggies. --seav 03:28, 9 June 2010 (UTC)

Cool suggestion! If you have a working code maybe we can make a simple map. I remember bart committed some server space. Maning 10:48, 9 June 2010 (UTC)
Let me see what I can come up with. I think a command-line script that takes an .osm file (containing natural=coastline ways as input and outputs the coordinates of the starting node of each chain of sawtoothed nodes is a good way, no? The output can then be fed into a webmap by plotting those coordinates. There's no need to highlight the whole sawtooth. Just point to the starting location (same as how keep right points to problematic ways) of the sawtooth. Based on the algorithm I see, it will take linear time and constant memory space with respect to the amount of input data. I'll try to whip something up this long weekend. --seav 05:03, 10 June 2010 (UTC)
Script done and sent to maning. --seav 19:41, 3 July 2010 (UTC)
Webmap showing the output of the script is up and running. See the main page for the link. --seav 02:09, 14 July 2010 (UTC)