Re: Grid To Hex Conversion

Re: Grid To Hex Conversion

[D. Cooper Stevenson]

On Tuesday 30 November 2004 18:34, you wrote:
> Hey Coop!
>
> Eric and I have been wondering if you knew what you
> were getting into when you offered to host XConq,
> we've had 2500+ hits in the last two weeks on the
> sf.net site, alone.

I'm thoroughly enjoying Xconq's continued success. I'm really happy to see the 
project going so well. 

Eric's cranking out new revisions like there's no tomorrow, you've brought in 
an artist and are actively pursuing good game design, and Matt's come out in 
force with the terrain imaging!

For my part, it turns out that Xconq uses an "oblique cartesian coordinate 
system" for displaying tiles. The Xconq manual does an excellent job of 
documenting this:

  "However, if you take a rectangular array of data and just wrap an area 
(terrain ...)) form around it, then everything will appear to be tilting to 
the left...

  The coordinate system is Cartesian oblique, with the y axis tilted to form a 
60-degree angle with the x axis, so it can be difficult to relate typed-in 
characters to the final appearance."

This means exactly what it says: I you try to overlay a rectangular land use 
grid over Xconq it's going to lean to the left. A screen image of what this 
looks like upon request. The solution is to "...have your program map the 
cell at x, y in the rectangular array to x - y / 2, y before writing." 

I continue to work on this. Here's some links to give you a better idea.

Here's a link to the input file. Each number represents a hex on the Xconq 
map. "42," for example, represents a hex of forest type I believe:

  http://wiki.xconqgis.org/map_files/landcover_ascii.txt
  
Here's the script that converts from the input file to the Xconq file:

  http://wiki.xconqgis.org/map_files/ascii_to_xconq.sh

Here's the output:

  http://wiki.xconqgis.org/map_files/earth-gis.g
  
On the plus side, I have found a comprehensive GIS Landcover, DEM, and road 
dataset for the entire globe!

Here it is:
http://www.mapability.com/index1.html?http&&&www.mapability.com/info/vmap0_index.html

And here for more information:

  http://earth-info.nga.mil/publications/vmap0.html

By the way, the 3D work based on the global VMAP0 data was already written for 
us a' la FlightGear:

  http://www.terragear.org/docs/vmap0/

Note that global VMAP1 data (1:250,000) scale is declasiffied but is not yet 
completely available for download:

http://www.mapability.com/index1.html?http&&&www.mapability.com/info/vmap1_index.html


-Coop

 
> Anyway, I had a new idea for a 
> poll:

[snip]

Okay! I'll post tomorrow!


-Coop

Re: Grid To Hex Conversion

[Eric McDonald]

Hi Coop,

On Wed, 1 Dec 2004, D. Cooper Stevenson wrote:

> an artist and are actively pursuing good game design, and Matt's come out in 
> force with the terrain imaging!

Matt's just getting started, I think. The best is yet to come.
(Matt, I saw your revised patch in the patch submission system, I 
will apply and test it tonight after I get home.)

> I continue to work on this. Here's some links to give you a better idea.

Glad to see you're back in action. I hope your system recovery 
went well.

> Here's a link to the input file. Each number represents a hex on the Xconq 
> map. "42," for example, represents a hex of forest type I believe:

It is also the answer to life, the universe, and everything. But, 
I digress....

> On the plus side, I have found a comprehensive GIS Landcover, DEM, and road 
> dataset for the entire globe!
> Here it is:
> http://www.mapability.com/index1.html?http&&&www.mapability.com/info/vmap0_index.html

Excellent.

Eric

Re: Grid To Hex Conversion

[D. Cooper Stevenson]

On Thursday 11 November 2004 23:59, you wrote:

> I'm not certain which images you mean, but if you mean the JPEGs in
> ansuz.sooke.bc.ca/temporary where I posted the Antarctic files, those are
> from the Helsinki leg of my Summer 2003 vacation in Finland and Sweden.
> The building in np020.jpg (which I'm guessing is the one you mean) is the
> Tuomiokirkko - a Lutheran cathedral in Helsinki.  

Yes, those are they. They are beautiful!

> I have a bunch more 
> photos (from the Swedish part of my trip) that I still have to scan,
> organize, and write captions for, even though it's been more than a year
> now.

They're interesting!


> What do you have in mind?  What I was thinking of was rather than a
> script, to modify the actual image-handling code in XConq more or less as
> discussed on the list several weeks ago.  

Yes. Modifying Xconq to parse whole images seems a cleaner solution than an 
outside script.

> That's going to involve some 
> careful data-structures work, though, and I won't be able to do it until I
> find a nice block of time.  XConq's image handling wasn't really designed
> to handle images bigger than a hex, nor thousands of images, so some
> reorganization will be needed to make it load large images and then cut
> thousands of little chunks out of them, as Eric and I discussed
> doing.  Nothing insurmountable, but it won't be a trivial hack either.

It's worth the wait. That is just incredible.

>
> I was also hoping that we'd have an official CVS again by the time I get a
> chance to work on image mapping, since the current pattern of getting
> updates as source tarballs makes it difficult for me to make modifications
> if my modifications will require more development time than the time
> between tarballs (hard to merge in new updates - that's part of the
> problem CVS is supposed to solve).

I hope we see this happen on SourceForge or Savannah. As you know, others have 
asked for this.

>
> If you're talking external script, then what you have in mind is probably
> something different from what I was envisioning.  Maybe it'd be easier to
> do; so tell me, what do you imagine an "image mapping script" as actually
> doing?  What sort of input would it take and what sort of output would it
> generate?  And if we're going to be discussing technical details, should
> we take this back onto the mailing list?  Feel free to reply to and
> quote this message on the list instead of directly to me, if you wish.

Yeah, I agree that carefully crafting Xcong to handle images natively is the 
best way to go.


-Coop