Colours and heights

So I’m getting back to iCartographer.
I still haven’t been able to get advertisements on the site, and I think I’ve stopped trying. Maybe after I put a few more updates on, then Google will accept me into their evil empire. But at present there is no incentive for me to tell people about the site; I can’t monetize my friends. Yet.
So, I’m backing away from updating the website. Although I have started using php to support it. That should allow bots to troll it better.
I recently got a desktop atlas. The maps I’m making did not compare well with it. The land I generate is all greyscale, and let’s face it, kind of depressing. So I decided to do some experimenting. I took the exact colours (taking a digital colour picker to their website) from the book and started to use them in my program. This revealed two new problems.
The first problem is that the book’s colours were picked to allow named places to show up clearly; they are muted pastels. I never changed my ocean colours which are vibrantly blue. This makes it so that the land is actually hard to see as it is washed out. I could reduce the vividness of the oceans, but I like having colourful maps. I think I’ll have to acknowledge that the land colours need to be selected differently.
The second problem is that the colour is giving me a better idea of the actual elevations that my program is generating. In the atlas, huge tracts of land are below 500 meters above sea level. In my program, barely any is. If you look at an elevation histogram of earth you can see that land elevation is like a hockey stick (backwards in the linked picture). So I’ve started doing some new experiments with the program to better approximate that. I’ve found that putting the elevation to a cubed power and then scaled back down works fairly well. But this has revealed a third problem.
This new problem is a fundamental flaw with the program. Mainly that I do not understand how fractals really work. Most of my knowledge has come from a few books that just give me enough information to be dangerous. But at a fundamental level, I don’t know what I’m doing. This comes into glaring realization when I try and increase the granularity of an already generated map. If any modifications have been made to the map, there is a good chance the new granularity will not be at an appropriate scale.
This is something I’ve got to work on. I need to better understand my project.