Request Realtime terrain generator

[deltawing777]

Ok that sounds good ill go with that.Both good ideas :thumbup:

---------- Post added at 04:22 PM ---------- Previous post was at 04:12 PM ----------

I have very detailed plans in this direction, and it is on my TO DO list. Just a few notes:

Because of licensing reasons, Google will not be the default data source. However, I will enable users to configure arbitrary map sources by means of editing / overwriting a configuration file. I've had a look at how Google maps tiles are encoded, and it should be no problem at all, as soon as you are legally allowed, to set Google as the data source.

I will probably allow users to configure how much tile cache should be stored on the local hard disk. This setting may also be relevant for your agreement with your map provider. In the case of Google, there seem to be two sides: on one hand, their license does not allow 'mass downloads', probably due to the fear that someone might download an entire copy of the map. A system with a large cache could be interpreted as a mass download system. On the other hand, people have reported being blocked from the Google maps servers after experimenting with high-traffic map applications. This is understandable, as bandwidth has a certain cost for Google, but the best way to reduce bandwidth usage is to have a large cache. So I'm not sure what Google prefers.

I'm not going to ask Google anything, because I have nothing to do with Google. I'm going to make sure that my software can be used with at least one free map provider, and I'll leave it to others, if they wish to use an alternative map source (such as Google), to get permission for that.

Have one computer getting the data(server) from the source(google)then we(clients) connect to the server.man in the middle configuration i guess im trying to say

 

[cjp]

Have one computer getting the data(server) from the source(google)then we(clients) connect to the server.man in the middle configuration i guess im trying to say

The legal problem is bigger than the technological one. Have one computer distribute Google maps data to multiple users is a sort of copying, and it can be a legal problem. I've asked around to some experts, and it seems that even when data is made available for free to all on the Internet, that doesn't mean it has become public domain.

On the other hand, such a server doesn't really need to be more sophisticated than an ordinary HTTP proxy. And deploying HTTP proxies is such a common practice that I can't see a problem there.

Remaining issue with the server idea: having all the bandwidth available for multiple orbinauts flying at the same moment isn't going to be cheap. As long as you don't use some P2P system, this could be a show stopper for a zero-fund project.

 

[deltawing777]

The legal problem is bigger than the technological one. Have one computer distribute Google maps data to multiple users is a sort of copying, and it can be a legal problem. I've asked around to some experts, and it seems that even when data is made available for free to all on the Internet, that doesn't mean it has become public domain.

On the other hand, such a server doesn't really need to be more sophisticated than an ordinary HTTP proxy. And deploying HTTP proxies is such a common practice that I can't see a problem there.

Remaining issue with the server idea: having all the bandwidth available for multiple orbinauts flying at the same moment isn't going to be cheap. As long as you don't use some P2P system, this could be a show stopper for a zero-fund project.

Yeah I could see where the upload bandwidth would be a problem and to get equal upload and download bandwidth isnt cheap you're right about that for sure! Do you have a idea for the time acceleration hurdle? That could be a show stopper also,im guessing thats where the large cache would come into play?

 

[cjp]

Do you have a idea for the time acceleration hurdle? That could be a show stopper also,im guessing thats where the large cache would come into play?

For the terrain generator, flying with time acceleration is not really different from flying very fast. The idea in libProcTer is that, if the terrain generating can't keep up, it will automatically switch back to a less detailed mode. You can already test the current version to see what that looks like.

The cache will help increase the speed for nearly all people: that's why browsers also have caches. And if it doesn't increase the speed (e.g. you have a very fast Internet connection), then at least it won't make much difference. Don't expect much faster speeds than the current version though: the current version of OrbiterProcTer loads it data from disk (for Earth, that is), which is about the same as it would do when a cache is used. The only way to make it faster is to optimize the speed of already existing functionality in libProcTer.

I think the current loading speed in OrbiterProcTer is CPU-limited, and not disk-limited. To reduce CPU usage as much as possible, OrbiterProcTer will save the cache in a format that has already done most of the image processing, such as re-projecting and color re-calibration. I may also keep a small cache of non-processed images, to optimize the situation where re-projection places one source tile onto multiple destination tiles, but I can probably keep that cache entirely in RAM.

 

[deltawing777]

For the terrain generator, flying with time acceleration is not really different from flying very fast. The idea in libProcTer is that, if the terrain generating can't keep up, it will automatically switch back to a less detailed mode. You can already test the current version to see what that looks like.

I have tested it as of your post and I like the direction you are going. I would like to make a few comments if I may. First of all I am no coder and I wish I knew 1/16 of what you know about coding. First of all there needs to be some kind of collision detection.The scenery above the moutains looks great but flying through them not so great.I also had some video issues that instead of me trying to explain them I thought I would take some time and make you a video so you could see them first hand from my perspective. My computer is core 2 duo 3 ghz cpu with Nvidia geforce GTS 250 video card with 1 Gig ram. You will see in the video some of the abnormal things that happen. But yes your addon is really looking good and thank you for taking the time to code it so we can all enjoy it. Cant wait to see the finished product :thumbup:
The link for the video is here I hope this helps you In some way:

---------- Post added at 09:53 PM ---------- Previous post was at 09:50 PM ----------

This is on a clean install of orbiter 2010 windows 7

---------- Post added at 10:34 PM ---------- Previous post was at 09:53 PM ----------

Anyway to combine your addon with artlav's meshland 2 addon for collision detection? Or maybe any chance both you and artlav could collaborate work together on the same goal of collision detection and terrain mapping?

 

[cjp]

First of all there needs to be some kind of collision detection.
[snip]
I also had some video issues that instead of me trying to explain them I thought I would take some time and make you a video so you could see them first hand from my perspective.
[snip]
http://www.youtube.com/watch?v=RB_BdWQhHWkThis is on a clean install of orbiter 2010 windows 7

Anyway to combine your addon with artlav's meshland 2 addon for collision detection? Or maybe any chance both you and artlav could collaborate work together on the same goal of collision detection and terrain mapping?

The video issues you showed are all known to me. The large distance issue will be solved in the future. But the good news for me is that OrbiterProcTer works with Orbiter 2010:jj:. I still have the 2006P1 version, but I will soon upgrade!

For collision detection, I do wish to collaborate with Artlav, or at least use his software. I know by experience how hard it is to get collision detection right, and it's not my ambition to make the best collision detection engine ever. I'll leave that to Artlav:tiphat:.

Man, there's still so much work to do!

 

[4throck]

Take a look at http://onearth.jpl.nasa.gov/ for some data that I think might be usable. Really not that different to the WorldStudio approach because the servers don't seem to be fast enough for real time access.

Google Earth imagery is a dead end, it's only licensed to their programs and on-line usage thought their API.

 

[deltawing777]

The video issues you showed are all known to me. The large distance issue will be solved in the future. But the good news for me is that OrbiterProcTer works with Orbiter 2010:jj:. I still have the 2006P1 version, but I will soon upgrade!

For collision detection, I do wish to collaborate with Artlav, or at least use his software. I know by experience how hard it is to get collision detection right, and it's not my ambition to make the best collision detection engine ever. I'll leave that to Artlav:tiphat:.

Man, there's still so much work to do!

Sounds good Ill be on the look out for any releases :tiphat:

 

[fort]

I started to reuse wind world last year for Landsat 7 and i cubed data (and incidentally virtual earth), an can be twice a month, since, and if the server is unreachable at times remained ten days, it has nevertheless always restarted. It works by example well tonight here. But this behavior is not reassuring. And I have, paradoxically, never seen so many queries from users on the world wind forum, when the server was down. How many use it ?

For this and other reasons, I use another program, sas Planet ( http://sasgis.com/sasplanet/) to create maps and those of Yahoo ( landsat7 icubed i think ) are virtually and mostly identical to those of WW. But that is another question.

This only to say that WW is, today, still alive. :yes:

 

[deltawing777]

I again started to reuse wind world last year for Landsat 7 and i cubed data (and incidentally virtual earth), an can be twice a month, since, and if the server is unreachable at times remained ten days, it has nevertheless always restarted. It works by example well tonight here. But this behavior is not reassuring. And I have, paradoxically, never seen many queries from users on the world wind forum, when the server was down. How many use it ?

For this and other reasons, I use another program, sas Planet ( http://sasgis.com/sasplanet/) to create maps and those of Yahoo are virtually identical to those of WW, the colors around. That is another question.

But this to say that WW is, today, still alive. :yes:

hummm very intresting