This is the mail archive of the xconq7@sources.redhat.com mailing list for the Xconq project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Antarctic Theatre


Mark,

I loaded and played your Antarctic game.  It looks and plays great. have
to tell you that it looks and plays great. It works fine given your
indications in the documentation. So far I've not had any trouble. 

Well, that's not exactly true: It seems one cannot the enemy base with
bombers "in one go." Do you think you could fix that?  :)

I wrote up a quick installation document:

http://wiki.xconqgis.org/index.php?UserInstallationInstructions

I'll have more concrete news as I study what you've achieved. It's like,
I think I know what you did but if I'm right then you've achieved
nothing more than the incredible. 

As I write this I go back and look at your documentation, especially in
the area of deriving cell elevation from the ASCII DEM data. I seriously
think to myself, "oh, man."

If I understand this correctly, your system basically uses the DEM data
to

  1) Converted the 14 million data points from the RADARSAT ASCII file
into the Lambert projection. This was necessary given the special nature
of mapping land mass around the Poles (Mercator gets it "wrong.")

  2)  Converted all the x and y coordinates into hex cells

  3) Computed the mean elevations for each cell

  4) Created the terrain file based upon a fairly dynamic "bumpiness"
algorythm. For example, if the cell's standard deviation is less than 20
and  it's mean elevation is less than 40 then it's pretty "smooth and
low" so you classify it as "ice-edge."

 5) Then you one step further and say, okay, height matters. Just
because the terrain's bumpy doesn't necessarily mean that it's a
highland. Your script distinguish basically between, "bumpy high" and
"bumpy low."

  6) Then you added the features 

Oh, man. If one were to tweak the rules a bit to reflect the region's
(anyone know where to get their hands on a "Middle Earth" DEM?) 
characteristics, they would basically have themselves a "turnkey" map
attribute generator, wouldn't one? Of course, I'm making it simpler than
it is (projection factor, for example) but basically, right?

It occurs to me that your terrain types were really special types given
the uniqueness of Antarctica. If one were to stick with "standard"
terrain types,  "antarctica" and "antar-stdterr" wouldn't be necessary,
correct? 

Please correct me if I'm mistaken, but it appears as if you've already
done the "hard" work of parsing the ASCII data into a two dimensional
array (terrain-type/elevation). It seems to me that if I can get Xconq
to render an individual image image per cell then having an Xconq game
based on real GIS maps is pretty much covered, correct or not correct?

Hey, you know what? It just occured to me that I have landcover data. I
don't believe I have to extrapolate terrain type; that's what the
landcover GIS layer is for! All I have to do is take the mean value per
cell to extrapolate terrain!

And if the above is true, this means automation:

  1) Pull down GIS elevation and landcover data for NW Lat/Long, SE
Lat/Long 

  2) Import landcover and elevation data  ( r.bin.in landcover,
elevation)

  3) Export landcover and elevation to ascii

  4) Run these through Matt's "cell division/mean elevation" algorithms,
output results to two-dimensioned terrain file 
  
  6) Rasterize shaded relief image (r.slope.aspect elevation=elevation
aspect=aspect)

  7) Output to a GIF or BMP for Xconq

  8) "Cellularize map" Using algorithm that I will write, hopefully with
your kind support

  9) Play game

Instant maps. Optimistic, yes, but I think it can be done.


-Coop




Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]