Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 David Megginson wrote: On 31/10/2007, Christian Buchner [EMAIL PROTECTED] wrote: I would like to get a few pointers where to look for information about the terrain engine that is currently used by Flight Gear. In particular about the irregular terrain mesh - how is it ... Our current terrain engine was great in the late 1990s (when MSFS still had a flat world), but it's pretty stale now, mainly from a lack of interest and the difficulty of understanding the code. It's actually deteriorated in some ways -- note that all the Great Lakes have been forced to sea level now, so that Lakes Huron, Michigan, and Superior are in deep gorges, and all the cities beside them are perched on cliff faces (that wasn't happening a couple of years ago). I'm not sure that we should really be maintaining our own terrain engine any more, though -- we should concentrate on aircraft models and flight dynamics, and try to interface with an existing engine. That depends on where your interests lie, I suppose. A flight simulator has specialized needs from a terrain engine: good close-up detail for near or on- earth operations, terrain stretching to a far horizon, no-pause loading of terrain from everywhere on the round earth, accurate material properties for ground polygons. I don't know of any free project that offers this. One promising prospect is Demeter, which works with OpenSceneGraph (already supported by FGFS): http://www.tbgsoftware.com/index.html Unfortunately, Demeter itself seems to have gone a bit stale -- some of the links from the site don't work any more, and it won't compile with the latest version of OSG. It does, however, support LOD and lots of other nice stuff (check out the screen shots!). Maybe someone should take it over if the maintainer has abandoned it, or fork it if the maintainer hasn't but is stalling on new releases. It doesn't do round-earth at all... The best thing we could do in FGFS, in the meantime, would be to isolate scenery-engine dependencies so that it's easy to switch to a new engine when it's ready. That's not a bad idea; isolating the scenery engine would make it viable as an independent project as well. I think we have the expertise, both on the production / GIS side and the programming side, to create a first-class world terrain and terrain engine that would be taken up by other projects too. I personally want to work on this, but I need to help get our OSG house in order first. I'll start a Wiki page with some of my ideas (which aren't very original). Tim - -- Red Hat France SARL, 171 Avenue Georges Clemenceau 92024 Nanterre Cedex, France. Siret n° 421 199 464 00056 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.7 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFHKYymeDhWHdXrDRURAqb6AJ9z96Jop7f3p7W4Ry64VMvJkE90LQCdF/Sf lWYlhMue+BYRO3Q7e1Dt9ew= =2vto -END PGP SIGNATURE- - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Problem compiling Atlas
On 10/31/07, Durk Talsma [EMAIL PROTECTED] wrote: As I was typing this reply, I got your other mail. Yes, you're pretty much right that Atlas no longer compiles against Simgear CVS. But, most fixes are pretty straightforward. Provided one already has that optind=0; stuff, the attached patch should take care of the linking flags and some simgear type changes I encountered. (Minus some local hackery I removed; it does compile and seemingly run as of last week's cvs HEAD of simgear , qualified by that evil for me) HTH K. Hoercher diff -Nur Atlas.cleanorig/src/Makefile.am Atlas.cleanhack/src/Makefile.am --- Atlas.cleanorig/src/Makefile.am 2007-10-26 13:19:27.0 +0200 +++ Atlas.cleanhack/src/Makefile.am 2007-10-26 13:19:02.0 +0200 @@ -28,6 +28,7 @@ Preferences.cxx Preferences.hxx Atlas_LDADD = \ + -lsgstructure -lsgprops \ -lsgmagvar -lsgtiming -lsgmisc -lsgio -lsgserial -lsgdebug -lsgbucket\ -lsgxml -lplibsg -lplibpu -lplibul -lplibfnt -lplibnet \ $(opengl_LIBS) \ @@ -45,8 +46,9 @@ fg_mkdir.cxx fg_mkdir.hxx Map_LDADD = \ + -lsgstructure -lsgprops -lsgserial -lsgxml \ -lsgmagvar -lsgtiming -lsgmisc -lsgio -lsgdebug \ - -lplibsg -lplibpu -lplibul -lplibfnt -lsgbucket -lsgscreen \ + -lplibsg -lplibpu -lplibul -lplibfnt -lplibnet -lsgbucket -lsgscreen \ $(opengl_LIBS) \ -lpng -lz -ljpeg @@ -61,8 +63,9 @@ Scenery.cxx Scenery.hxx MapPS_LDADD = \ + -lsgstructure -lsgprops -lsgserial -lsgxml \ -lsgmagvar -lsgtiming -lsgmisc -lsgio -lsgdebug \ - -lplibsg -lsgbucket \ + -lplibsg -lplibnet -lsgbucket \ -lpng -lz GetMap_SOURCES = \ diff -Nur Atlas.cleanorig/src/MapMaker.cxx Atlas.cleanhack/src/MapMaker.cxx --- Atlas.cleanorig/src/MapMaker.cxx 2007-10-26 13:19:27.0 +0200 +++ Atlas.cleanhack/src/MapMaker.cxx 2007-10-26 13:19:02.0 +0200 @@ -653,8 +653,8 @@ /* convert point_list of wgs84 nodes to a list of points transformed into the maps local coordinate system */ - const point_list wgs84_nodes = tile.get_wgs84_nodes(); - for ( point_list::const_iterator node = wgs84_nodes . begin(); + const std::vectorSGVec3d wgs84_nodes = tile.get_wgs84_nodes(); + for ( std::vectorSGVec3d::const_iterator node = wgs84_nodes . begin(); node != wgs84_nodes . end(); node++ ) { @@ -674,8 +674,8 @@ } // same as above for normals - const point_list m_norms = tile.get_normals(); - for ( point_list::const_iterator normal = m_norms.begin(); + const std::vectorSGVec3f m_norms= tile.get_normals(); + for ( std::vectorSGVec3f::const_iterator normal = m_norms.begin(); normal != m_norms.end(); normal++ ) { // Make a new normal - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] Colombian Air Force virtual T-37
Hi, I received the following request from a university group: VOLTA is a research group of Militar Nueva Granada University, in this moment we are development a project for Colombian Air Force. In reference with a e-mail send by JHON JAIRO PALOMINO LOSADA (it was send in 25-oct-2007), We need the T-37 for FSX with all the instruments at the control panel and painted with the Colombian Air Force logo and colors. If you can do it, send us the quotation (how much is it, time and payment way) This requirement is very important and urgent for us, we thank for its aid and quick answer Now since I don't have time to do it, nor the expertise to create a FSX model I wonder is someone here might be interested in this proposal. I wouldn't mind if someone would make some bucks out of this using my T-37 model (both FDM and 3d model) provided that the changes to the existing T-37 model will be backported to the FlightGear model as well. Anyone who is interested could contact me privately for the contact information. Erik - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Compiling TerraGear !
Well I have downloaded the latest Simgear CVS and ran the autogen.sh but is missing ./autogen.sh Host info: Linux i686 automake: 1.10 (110) Running aclocal aclocal: `configure.ac' or `configure.in' is required Running autoheader autoheader: `configure.ac' or `configure.in' is required ERROR: autoheader didn't create simgear/simgear_config.h.in! Why are these missing? Kind Regards, Will [EMAIL PROTECTED] 10/31/07 16:53 PM On 10/31/07, Jon Stockill wrote: will Pink wrote: I have the lastest CVS version of SimGear but the package doesn't include configure.ac or configure.in! Any reason why? Because it's the cvs version, and those are built by autotools There is definitely a configure.ac in the cvs checkout. You need to run ./autogen.sh followed by ./configure , followed by make ... You also need to have OSG available since that is a dependency for simgear. Curt. -- Curtis Olson: http://baron.flightgear.org/~curt/ Unique text: 2f585eeea02e2c79d7b1d8c4963bae2d This message has been checked for all known viruses by the MessageLabs Virus Control Centre on behalf of the Marshall Group of Companies. - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
On 01/11/2007, Tim Moore [EMAIL PROTECTED] wrote: ... we should concentrate on aircraft models and flight dynamics, and try to interface with an existing engine. That depends on where your interests lie, I suppose. By we, I meant the Flightgear project, not the individuals in it. I would also be very interested in contributing to a terrain engine, but I think it probably would need its own maintainer and team of core developers who were focusing primarily on it, rather than building it as a sideshow to FlightGear, especially since it would have to come with a suite of editing tools to help people manipulate GIS data and manually tweak scenery. All the best, David - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
Hello, Few thoughts as FlightGear moves to osg there is a nice virtual terrain project which is integrates with osg. The site for the project is http://www.vterrain.org/ and some papers are here http://www.vterrain.org/LOD/Papers/. Personally I like Thatcher Ulrich adaptive quadtrees approach ( which allows to load chunks in runtime and thus it is possible to ) the approach is implemented in Soul Ride game which source could be obtained here http://sourceforge.net/projects/soulride/ ( this one specific implementation does not feature loading new chunks but I seen the java code with the functionality. anyhow there are quite a bit of approaches which are presented at http://www.vterrain.org/LOD/Papers/ and some are more hardware friendly than adaptive quadtrees. Regards Sergey Kurdakov http://www.sim-ai.org/simplestgame.htm http://www.sim-ai.org/FlightGearlesson.htm On 11/1/07, Tim Moore [EMAIL PROTECTED] wrote: A flight simulator has specialized needs from a terrain engine: good close-up detail for near or on- earth operations, terrain stretching to a far horizon, no-pause loading of terrain from everywhere on the round earth, accurate material properties for ground polygons. I don't know of any free project that offers this. I think we have the expertise, both on the production / GIS side and the programming side, to create a first-class world terrain and terrain engine that would be taken up by other projects too. I personally want to work on this, but I need to help get our OSG house in order first. I'll start a Wiki page with some of my ideas (which aren't very original). - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
On 01/11/2007, Sergey [EMAIL PROTECTED] wrote: as FlightGear moves to osg there is a nice virtual terrain project which is integrates with osg. The site for the project is http://www.vterrain.org/ and some papers are here http://www.vterrain.org/LOD/Papers/. Excellent. It looks like a good start, but there is a strange culture around the site -- for example, you have to request a download instead of simply downloading the source, even though it seems to have a BSD-style license. I wonder if most of the developers are coming from the Windows world, and don't quite get the norms of open source yet. All the best, David - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
Hello David. you have to request a download instead of simply downloading the source, even though it seems to have a BSD-style license. I do not know exact rationale why this system to get a code is in place, still the pass goes without problems and the author is from linux/unix world himself ( to my know ). Regards sergey On 11/1/07, David Megginson wrote: On 01/11/2007, Sergey wrote: as FlightGear moves to osg there is a nice virtual terrain project which is integrates with osg. The site for the project is http://www.vterrain.org/ and some papers are here http://www.vterrain.org/LOD/Papers/. Excellent. It looks like a good start, but there is a strange culture around the site -- for example, you have to request a download instead of simply downloading the source, even though it seems to have a BSD-style license. I wonder if most of the developers are coming from the Windows world, and don't quite get the norms of open source yet. - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Colombian Air Force virtual T-37 for FSX
Dear Sir: We are very interested in develop the T-37 for FSX in the way that you said. Please contact us whit the persons who could help us (if is different to you), this to establish the specifications, develop time, cost and mode of payment. One of the requirements for the virtual T-37 for FSX is the time of development which can`t be greater to fifteen day. we thank for its aid VOLTA Research group William Gómez Rivera M.Sc. Mobile Phone: 57-3112324076 Dear Sir/Madam, I've asked around in the FlightGear team to see if others are interested in adjusting this model to FSX and there seams to be some interest in doing so. If you are still interested in this effort please let me know (or you could post a message the the flightgear mailinglist: flightgear-devel@lists.sourceforge.net). Regards, Erik Hofman Grupo de Investigación VOLTA wrote: Hello Mr. Erik: VOLTA is a research group of Militar Nueva Granada University, in this moment we are development a project for Colombian Air Force. In reference with a e-mail send by JHON JAIRO PALOMINO LOSADA (it was send in 25-oct-2007), We need the T-37 for FSX with all the instruments at the control panel and painted with the Colombian Air Force logo and colors. If you can do it, send us the quotation (how much is it, time and payment way) This requirement is very important and urgent for us, we thank for its aid and quick answer VOLTA Research group William Gómez Rivera M.Sc. Mobile Phone: 57-3112324076 VOLTA Grupo de Investigación en Energias Renovables y Aplicaciones Mecánicas Lider: M.Sc.Ing. William Gómez Rivera - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FlightGear terrain engine
Hello! Georg Vollnhals schrieb: 1. Getting better scenery details for wide areas without buying any data - the Landsat7 project http://www.custom-scenery.org/Building-Scener.331.0.html Here a semi-automatic scanning of prepared Landsat 7 pictures can result into a much more detailed scenery. There are two areas available for FlightGear, the Berlin Linuxtag scenery and (as far as I remember) the Bodensee scenery. Although the process of scanning these pictures is not easy due to the low resolution and the software and AI is under development the result was fascinating in my eyes. The aim of this project is to enhance the local scenery worldwide automatically by scanning the LS7 stuff. This is far! from working but a very interesting branch of development. Well, it _is_ working, as you can see from the Custom Scenery examples on http://mapserver.flightgear.org/. The limiting factors currently are the simplification required so the TerraGear and FlightGear don't choke on the load of details we get from the low resolution imagery ;-) , and the second is making the process workable for average Joe User (which is mainly a problem of simplification, as well ;-) ) I have documented the process of classification at http://www.custom-scenery.org/Building-Scener.331.0.html Martin is currently working on trying the process himself and maybe will be able to establish a webservice where users can retrieve parts of the Landsat imagery for their favourite area, produce a training vector map, e.g. using QGIS, submit that to the service again and retrieve automatically classified vector data. He has established close ties with the OSGeo foundation and we are allowed to use some of their topnotch hardware for this task. So, no, this is not far from working. In contrast, we are finally approaching a point where this can be applied generally, and it might be that Seb's work will allow us to use all the detail we can retrieve from Landsat. Cheers, Ralf - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] FlightGear terrain engine
Ralf Gerlich schrieb: Hello! Georg Vollnhals schrieb: 1. Getting better scenery details for wide areas without buying any data - the Landsat7 project http://www.custom-scenery.org/Building-Scener.331.0.html Here a semi-automatic scanning of prepared Landsat 7 pictures can result into a much more detailed scenery. There are two areas available for FlightGear, the Berlin Linuxtag scenery and (as far as I remember) the Bodensee scenery. Although the process of scanning these pictures is not easy due to the low resolution and the software and AI is under development the result was fascinating in my eyes. The aim of this project is to enhance the local scenery worldwide automatically by scanning the LS7 stuff. This is far! from working but a very interesting branch of development. Well, it _is_ working, as you can see from the Custom Scenery examples on http://mapserver.flightgear.org/. The limiting factors currently are the simplification required so the TerraGear and FlightGear don't choke on the load of details we get from the low resolution imagery ;-) , and the second is making the process workable for average Joe User (which is mainly a problem of simplification, as well ;-) ) I have documented the process of classification at http://www.custom-scenery.org/Building-Scener.331.0.html Martin is currently working on trying the process himself and maybe will be able to establish a webservice where users can retrieve parts of the Landsat imagery for their favourite area, produce a training vector map, e.g. using QGIS, submit that to the service again and retrieve automatically classified vector data. He has established close ties with the OSGeo foundation and we are allowed to use some of their topnotch hardware for this task. So, no, this is not far from working. In contrast, we are finally approaching a point where this can be applied generally, and it might be that Seb's work will allow us to use all the detail we can retrieve from Landsat. Cheers, Ralf Hi Ralf, you know yourself how much I appreciate your work and like the output - just remember my words about the Berlin terrain in the GFF. The aim of this project is to enhance the local scenery worldwide automatically by scanning the LS7 stuff. This worldwide automatically = press the button I'll do the rest is really far from working as you describe yourself. Actually there is a lot of knowledge necessary to do the work and even if Martin will finish his part you have to go through the training process for every LS7 picture as far as I understood all. But your work will result in a big step forward to a better scenery, together with tasks from other people, and even going through several procedures before you'll get the result is nothing compared to the manually Landsat7 picture scanning like I did for OSG areas and obviously many OSG people do - even if OSG has a Tools with JOSM and the LS7 background-area function a tool we are only dreaming of. submit that to the service again and retrieve automatically classified vector data I doubt that this will help most of the scenery-developers/users as long as you cannot process these data due to that cannot compile, cannot use TerraGear problem. Regards Georg - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Colombian Air Force virtual T-37 for FSX
On Thursday 01 November 2007 18:26, Erik Hofman wrote: Grupo de Investigación VOLTA wrote: Dear Sir: We are very interested in develop the T-37 for FSX in the way that you said. Please contact us whit the persons who could help us (if is different to you), this to establish the specifications, develop time, cost and mode of payment. One of the requirements for the virtual T-37 for FSX is the time of development which can`t be greater to fifteen day. So far I've had one response of someone who is skilled enough to do the 3d modeling but doesn't know much about FSX and other parts of the simulator. If anyone else is interested to join in please let it know. (I was thinking maybe Bill Galbraith might be interested). If this doesn't work out I think I want to let them know as soon as possible. Erik Alexis did a great job of making the A-10 usable, so in that respect it would be in good hands. Has MS FSX actually been released yet? LeeE - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Material animation problem
SydSandy a écrit : On Wed, 31 Oct 2007 20:51:40 + Jon Stockill [EMAIL PROTECTED] wrote: I'm having problems with the following seemingly simple bit of animation (running the cvs osg version): animation object-nameWarnLoVolt/object-name object-nameWarnGen/object-name object-nameWarnLoOil/object-name object-nameWarnFuel/object-name object-nameWarnStarter/object-name condition not propertyinstrumentation/warning-panel/test/property /not /condition typematerial/type emission red0.0/red green0.0/green blue0.0/blue factor-propinstrumentation/warning-panel/brightness/factor-prop /emission /animation animation object-nameWarnLoVolt/object-name object-nameWarnGen/object-name object-nameWarnLoOil/object-name object-nameWarnFuel/object-name object-nameWarnStarter/object-name condition propertyinstrumentation/warning-panel/test/property /condition typematerial/type emission red1.0/red green0.0/green blue0.0/blue factor-propinstrumentation/warning-panel/brightness/factor-prop /emission /animation It's supposed to light all the lamps on a warning panel when the test button is pressed. If used as shown, the lamps never light. Switching the order of the 2 animation blocks results in the lamps lighting when the test switch is pressed - but they remain on when it is released (the controlling property value in the browser does change). With the lamps lit if the brightness is changed and the test button pressed again the brightness of the lamps changes - so the first animation can still be triggered - it appears that the second is always ignored though. Have I done something silly, or have I discovered a bug? Jon Hi Jon , I THINK type has to come right after animation ... Cheers Another tip: you better use a group in the .ac (3D) model and use the group name when you want an animation on all objects together, doing so create a different branch in the tree. If you specify all the objects namely in the same block, you may have difficulties to animate them separately later. You can also use another method which is a bit more verbose: a select animation with the condition, and later in the xml the material animation without condition. You will have in your .ac (3D) file of the panel the lamp off and the lamp on. then in the xml you can select the 'lamp on' object upon a property. Again use separate block for single objects called by their name or a group of them if needed. Alexis - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Demeter (was Re: FlightGear terrain engine)
On Thursday 01 November 2007 09:32, Curtis Olson wrote: Certainly we could and should take a look at cleaning up the scenery render engine API within FlightGear. I think the problem here is about sceney _data_ generation as much as about the scenery engine itself. The sceney generation process is overly complex, many people who tried didn't get far for one reason or another. Then, we have experts who know how everything ties together but simply don't have the time to maintain the code. Finally, when scenery actually gets generated, it either lost a lot of details or worst, has a lot of errors. There are a few opensource libraries that are dedicated to the features that you have mentioned. GTS, for example, can do continious LODs as well as CSG. The great thing about using existing libraries is that these libraries are purposely designed, dedicated to what they do, and are maintained by experts, instead of being hacks that are written once and hardly get looked at afterwards. So, not only will we able to cut holes into the terrain for runways, we will also be able to do so better and much more efficiently. Ampere - This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now http://get.splunk.com/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel