[dajones]

It turns out osgEarth uses an EllipsoidModel class from OSG to convert back and forth between latitude, longitude, height and ECEF. And a comment in that class contains this link http://www.colorado..../gif/llhxyz.gif for the math. The WGS84 equator radius is 6378137.0 and the polar radius is 6356752.3142.

Here is a link to osgEarth if anyone is interested.

Doug

[BillC]

 Genma Saotome, on 11 April 2015 - 08:53 AM, said:

@Bill -- Cool. I was reading some other part of the web page that spoke to inputs and it only mentioned the Space Shuttle data. V. good to see 1/9th arc .img data canbe processed.

But processed into what? I dunno Tranz... no idea what their tile files are like. And then there is the payware issue.

Dave,

For myself using TransDem is no brainer as I already have copies of Trainz 2006 and 2012, and played around with it doing toy routes.TrainZ it has good support and tutorials. As I indicated in a previous post Walt has done an analysis on TranzDem see this post #83. For approximately $30 USD this program is a bargain. Here is a post by Dr. Ziegler (geophil) on his work with 3D UTM tiles. Unfortunately support is currently Trainz related. If OR could define a new baseboard structure based on a UTM structure, perhaps Dr. Ziegler could be convinced to support OR on a sales basis.

The point I am trying to make here, that OR needs to select a baseboard not the harrow of Goode-Homosoline. A viable solution would be UTM, whether commercial support, or roll your own open source. Much of source input NED data, topographic maps, etc are based on UTM. Trains do not, except in rare instances travel long distances, as aircraft. So no need for large circle navigation that international jets and ocean liners use. Heck, as a private pilot straight line route planning was good enough. Even flying the iron compass, which was the railroad ;-) . IMHO a need to select a common supported base system, to have a RE in the foreseeable future. Preaching to the choir on this.

[Genma Saotome]

 BillC, on 11 April 2015 - 04:11 PM, said:

The point I am trying to make here, that OR needs to select a baseboard not the harrow of Goode-Homosoline.

Agreed. The reason I posted the basenote is to try to show why OR should not use Goode-Homosoline.

Going forward tho... I understand your point about using TransDEM; my reservations come largely from the experience of DEMEX, which was a GREAT start, middling middle age, and an untimely death. And we're still trying to do stuff with its bones 10 years later. That's why I'd prefer to cast my vote for an Open Source tool.

Naturally, the downside of an Open Source GIS tool is the OR team would most likely have to adapt it for this project... not sure there is anyone both capable and willing to do that. Assuming there is someone who can do the initial work then the long term viability of the tool is in the project teams hands and not some solo actor on the outside and I do that that is important.

I have found and installed a couple Open Source/free GIS tools -- qdig, QGIS wien, Grass, and SAGA. About all I can say is GIS is certainly a complex topic; these programs do a whole heck of a lot of stuff I don't understand. Of these three I've gotten a bit further with QGIS wien, SAGA, and Grass... very basic stuff like importing .img files.

Grass has been fairly easy to figure out but ti my eye seems to be cluttered with more advanced GIS features than the others. I don't know what exactly that means for OR... left as is would be a massive degree of confusion for an average route builder.

Which leads to a question: Let's say there develops a preference for one of the Open Source GIS tools. Is it really feasible to strip out all of the extraneous features and still have something that can track along with the original (I'm thinking about some future format change by one or more data sources, as had happened in the past)?

[WaltN]

 Genma Saotome, on 11 April 2015 - 05:10 PM, said:

Is it really feasible to strip out all of the extraneous features and still have something that can track along with the original (I'm thinking about some future format change by one or more data sources, as had happened in the past)?

Yes. It's called DotSpatial. Others (third parties) are doing just that, and charging for their work product.

The big question is buried in your parenthetical remark: What is going on with the support base for MapWindow/DotSpatial at the University of Idaho? The project seems to be leaderless based on observational evidence.

But, regardless, it's going to take a committed individual in Open Rails to take on the cartography job. But, we've got time. Version 4 is a ways off.

[WaltN]

 WaltN, on 13 April 2015 - 03:03 AM, said:

Yes. It's called DotSpatial. Others (third parties) are doing just that, and charging for their work product.

The big question is buried in your parenthetical remark: What is going on with the support base for MapWindow/DotSpatial at the University of Idaho? The project seems to be leaderless based on observational evidence.

But, regardless, it's going to take a committed individual in Open Rails to take on the cartography job. But, we've got time. Version 4 is a ways off.

I decided to push things a bit, and I went for the father of MapWindow. Here is his reply:

Quote

Hi Walt,

MapWindow 4.x continues to grow and new releases are being planned by the Dutch team lead by Paul Meems. The discussion forum you applied to has actually been recently decommissioned in favor of the forums at mapwingis.codeplex.com and mapwindow4.codeplex.com.

You are tight that DotSpatial is also becoming more active as a .NET/C# mapping library.

- Dan

Dan is now at BYU.

[Genma Saotome]

Walt, have you taken a look at QGIS wien, Grass, or SAGA? All freeware and some variant of an Open Source license. IMO you have a better eye in assessing these products than I do.

[WaltN]

I took a brief look at Grass a few days ago, and then Doug's post about ECEF absorbed all of my attention.

The problem with these GIS systems is that, unless you spend a lot of time with cartography, investigating a candidate thoroughly and objectively takes a lot of time and energy. If I was sufficiently motivated, I would take a test set of DEMs and shape files and do a test scenario. I think that's the only way to do a fair assessment.

That said, I personally would start with MapWindow/DotSpatial. Several years ago, when I stumbled into it, it struck me as more professionally executed than the few freeware GIS systems I had seen. (I had spent considerable time with Microdem, and had looked at TransDEM and talked to Ziegler.) Subsequently, an old IBM colleague, Alan Jones, contacted me. He was doing some work in seismic analysis for CBS News, and he wondered if I knew of any program that would read a GIS shape file and turn it into a format that was easy to do subsequent processing on. I suggested MapWindow. He took me at my word and, in one day, downloaded and installed the program, indoctrinated himself, read the shape file, and did his processing on it. Job done! That impressed me.

The other thing that impressed me was that Dan Ames started MapWindow at Idaho State as an effort to produce the GIS functionality of ArcGIS as open source. It was something more than a guy in his basement. It was an ISU-chartered project. It is open source, it is free, and it is written in C#. Much more than that, I don't know -- until I take that next step.

[WaltN]

This today from Dan Ames:

Quote

Walt, thanks for recommending these libraries. There is always more work to be done on both toolkits, but they are quite complete and usable in their current form. The main thing lacking for both DotSpatial and MapWinGIS is up to date documentation. But that is typical for open source. We're working on addressing it...

James, if you're listening, take a look at this page. If you were to consider doing an Open Rails GIS utility kind of thing, which would you try first: DotSpatial or MapWinGIS? What I'm asking for is your sense of which is the better toolkit from a systems point of view, not a GIS point of view. I'm assuming they're equal from a GIS point of view.

[James Ross]

 WaltN, on 14 April 2015 - 09:34 AM, said:

James, if you're listening, take a look at this page. If you were to consider doing an Open Rails GIS utility kind of thing, which would you try first: DotSpatial or MapWinGIS? What I'm asking for is your sense of which is the better toolkit from a systems point of view, not a GIS point of view. I'm assuming they're equal from a GIS point of view.

Okay, let's see... I've only looked over them a little, using the information from http://dotspatial.codeplex.com/ and http://mapwingis.codeplex.com/ and linked documentation/intros.

They both appear to provide the same basic, core function: a library you can use for your GIS/mapping needs, including the visualisation aspect. Although MapWinGIS is in C++, by compiling to an ActiveX control it will have good interoperation with C#/.NET; but if we were at any point interested in utilising the source code we'd have to go with DotSpatial (for our needs) due to it being written in C#/.NET and not C++.

So all things being equal on the abilities and relevant functions, I'd lean towards DotSpatial - if only so we can look at the code and explore it more easily with our existing programmers.

[WaltN]

 James Ross, on 14 April 2015 - 12:54 PM, said:

They both appear to provide the same basic, core function: a library you can use for your GIS/mapping needs, including the visualisation aspect. Although MapWinGIS is in C++, by compiling to an ActiveX control it will have good interoperation with C#/.NET; but if we were at any point interested in utilising the source code we'd have to go with DotSpatial (for our needs) due to it being written in C#/.NET and not C++.

That was my sense too, but I thought I'd employ a second opinion to boost my confidence. Thanks.

If I decide to test the water, I'll get familiar with MapWindow first -- to become acquainted with the jargon and work flow. Then on to DotSpatial. That ought to keep me out of trouble for a while.