Work at SourceForge, help us to make it a better place! We have an immediate need for a Support Technician in our San Francisco or Denver office.

gpsbabel-misc

 [Gpsbabel-misc] Circles to make a route... From: Jason Slagle - 2005-07-18 20:41:46 ```I'm trying to calculate the proper locations of circles along a route to maintain at least X distance between them where they intersect. Anyone here know the generic formula? Going to use it hopefully to calculate the required placement of pocket queries to do a "x miles from route" query. Would such a function be useful as a filter? Jason -- Jason Slagle /"\ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . \ / ASCII Ribbon Campaign . X - NO HTML/RTF in e-mail . / \ - NO Word docs in e-mail . ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Ron Parker - 2005-07-18 20:50:06 ```At 03:51 PM 7/18/2005, Jason Slagle wrote: >I'm trying to calculate the proper locations of circles along a route to >maintain at least X distance between them where they intersect. Anyone >here know the generic formula? Do you mean you want to compute circles such that the distance between centers is maximized while keeping the minimum distance from route to intersection (zero, one, or two such intersections) above a given number? If there is a generic formula, it's going to be rather ugly. >Would such a function be useful as a filter? I think I'd find a use for it if it were sufficiently configurable. ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Jason Slagle - 2005-07-18 20:56:33 ```On Mon, 18 Jul 2005, Ron Parker wrote: > Do you mean you want to compute circles such that the distance between > centers is maximized while keeping the minimum distance from route to > intersection (zero, one, or two such intersections) above a given number? That would be it. I figure it's going to be ugly, but you have some hints I think. > If there is a generic formula, it's going to be rather ugly. Yeah. Moving along a single line (Just along x) isn't bad, but when you throw y in there too, you get all goofy. > I think I'd find a use for it if it were sufficiently configurable. It would have to be. My idea is a way to take google route, generate locations for suitable pocket queries, create them, and process the results. I can do everything but calculate the suitable pocket query locations as it stands now. Jason -- Jason Slagle /"\ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . \ / ASCII Ribbon Campaign . X - NO HTML/RTF in e-mail . / \ - NO Word docs in e-mail . ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Ron Parker - 2005-07-18 21:06:21 ```At 04:06 PM 7/18/2005, Jason Slagle wrote: >On Mon, 18 Jul 2005, Ron Parker wrote: > >>Do you mean you want to compute circles such that the distance between >>centers is maximized while keeping the minimum distance from route to >>intersection (zero, one, or two such intersections) above a given number? > >That would be it. I figure it's going to be ugly, but you have some hints >I think. Possibly. I don't have anything off the top of my head, but I'll see if I can dredge up my previous work on the same problem. >Yeah. Moving along a single line (Just along x) isn't bad, but when you >throw y in there too, you get all goofy. With appropriate coordinate transformations, a single line is not a problem in any case. It's the case where the center of the new circle isn't moving along the line that joins it to the center of the first circle, which will happen quite frequently in the case under discussion. In addition, you're constrained by the endpoints of the current line segment. And so on. Computing the minimum distance from the route to the intersection points would also be tricky for such a route. And I can't shake a nagging suspicion that there exists a pathological route that would violate the implicit assumption that the the given conditions are equivalent to the condition you really want, which is that no point within a given distance of the route be outside of the union of all your circles. ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Robert Lipe - 2005-07-18 23:31:21 ```[ This will be kind of a geocaching-centric conversation, so some of you may tune out while we figure out if this can be solved in a better way here. ] > Computing the minimum distance from the route to the intersection points > would also be tricky for such a route. And I can't shake a nagging > suspicion that there exists a pathological route that would violate the It's more complicated than that becuase there's another variable involved. It's not merely "how to I place the circles to cover the distance?" it's "how do I place the circles to cover the distance and not exceed the max number of points per circle that the server can hand me?" Remember, there's a max of 500 in a PQ. So you have to factor caches per square mile into your calcs. Since this information can only be gotten from the server - the very server you're computing these circles to query - you have problem that is, well, circular. 500 points will get you a very tiny circle in Santa Clara but a much larger circle in New Mexico. http://forums.groundspeak.com/GC/index.php?showtopic=68615 I'll admit I've always built them by hand and have find it kind of frustrating to have to futz with it but it's not insurmountable. I'll use the 'preview' function to get an idea how much distance I can cover with one. Then I'll check that last page for the most distant in the directions of interest and then futz with centerpoints of the next one unt If I have time up front, I'll just stock up on PQs of the area and pile them into S&T. I'll then use the "find nearest by radius" and circle drawing with the various PQ's set for different colored pushpins. I can sometimes squeeze things down another 20 or 30% by eliminating overlap and overspray just by moving the centers (sometimes in non-obvious ways) to maximize the coverage along the route I cared about. What non-obvious ways? Use the state selection to be sure you get only what you want. For example, when I drive from Nashville to St. Louis I'll go through KY and IL but never get near IN. So when picking states, be sure that MO and IL are ticked by TN, AK, and IN are not so you don't pick those up in the PQ. Think of the state filter as a mask over the radius filter. I've long meant to write this technique up ala http://www.gpsbabel.org/formats/s_and_t/TripPlanning.html but just never gotten around to it. RJL ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Jason Slagle - 2005-07-19 00:23:21 ```On Mon, 18 Jul 2005, Robert Lipe wrote: > [ This will be kind of a geocaching-centric conversation, so some of you > may tune out while we figure out if this can be solved in a better way > here. ] We can always move the discussion to the other place. The three of us are there are we not? > It's more complicated than that becuase there's another variable > involved. It's not merely "how to I place the circles to cover the > distance?" it's "how do I place the circles to cover the distance and > not exceed the max number of points per circle that the server can hand > me?" Remember, there's a max of 500 in a PQ. Yes, thats true also, but I have in the past usually used uniform sized circles, something like 50 miles, and then in areas where that is an issue filled in with smaller ones. If doing processing for a trip, I could always spit out a warning if you hit 500 and are nowhere near the radius of the circle. The only place I could see that breaking is where you have a whole TON of caches in a small area with nothing on the outlying area. It's not an insurmountable problem, and in a generic fashion not something you necessarily need to solve. > I'll admit I've always built them by hand and have find it kind of > frustrating to have to futz with it but it's not insurmountable. I'll > use the 'preview' function to get an idea how much distance I can cover > with one. Then I'll check that last page for the most distant in the > directions of interest and then futz with centerpoints of the next one > unt Yeah, I have a way I use to create the PQ's and then I just set them to not run. Then I go in and manually look at them to see they do what I like and set them to run. Then I have scripts I wrote to produce the output. > If I have time up front, I'll just stock up on PQs of the area and pile > them into S&T. I'll then use the "find nearest by radius" and circle > drawing with the various PQ's set for different colored pushpins. I can > sometimes squeeze things down another 20 or 30% by eliminating overlap > and overspray just by moving the centers (sometimes in non-obvious ways) > to maximize the coverage along the route I cared about. > > What non-obvious ways? Use the state selection to be sure you get only > what you want. For example, when I drive from Nashville to St. Louis > I'll go through KY and IL but never get near IN. So when picking > states, be sure that MO and IL are ticked by TN, AK, and IN are not so > you don't pick those up in the PQ. Think of the state filter as a mask > over the radius filter. Yeah. This would be good if you absolutely care about number of PQ's. The trip from vegas to toledo only took 22 50 mile pq's, and IIRC, we only had a size issue once, and it was near denver. Poke me offlist if you wish to see the results. I still maintain an algorithm could be written to do this, and will continue playing. I'll attach my slashnet client if you wanna chat about it. Jason -- Jason Slagle /"\ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . \ / ASCII Ribbon Campaign . X - NO HTML/RTF in e-mail . / \ - NO Word docs in e-mail . ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Ron Parker - 2005-07-19 01:25:26 ```Jason Slagle wrote: > On Mon, 18 Jul 2005, Robert Lipe wrote: > >> [ This will be kind of a geocaching-centric conversation, so some of you >> may tune out while we figure out if this can be solved in a better way >> here. ] > > > We can always move the discussion to the other place. The three of us > are there are we not? If "the other place" is anywhere but -code, the answer is no. ```
 Re: [Gpsbabel-misc] Circles to make a route... From: Jason Slagle - 2005-07-19 15:38:56 ```On Mon, 18 Jul 2005, Ron Parker wrote: > At 04:06 PM 7/18/2005, Jason Slagle wrote: >> On Mon, 18 Jul 2005, Ron Parker wrote: >> >>> Do you mean you want to compute circles such that the distance between >>> centers is maximized while keeping the minimum distance from route to >>> intersection (zero, one, or two such intersections) above a given number? >> >> That would be it. I figure it's going to be ugly, but you have some hints >> I think. > > Possibly. I don't have anything off the top of my head, but I'll see if I > can dredge up my previous work on the same problem. Thanks to ClayJar for the general formula - a couple of us here at work simplified it. Let W = the desired width (2x the "distance from path) Let D = the distance between 2 circles. Let r1 = the radius of circle 1 Let r2 = the radius of circle 2. To maximize D for a given W: D=sqrt(r1^2-(W^2/4))+sqrt(r2^2-(w^2/4)) Just set the center points that far apart. > With appropriate coordinate transformations, a single line is not a problem > in any case. It's the case where the center of the new circle isn't moving > along the line that joins it to the center of the first circle, which will > happen quite frequently in the case under discussion. In addition, you're > constrained by the endpoints of the current line segment. And so on. Endpoints are easy. If you set them at r-(1/2*w) from the endpoints, you're guaranteed at least 1/2 w at any point for all r > 1/2w. > Computing the minimum distance from the route to the intersection points > would also be tricky for such a route. And I can't shake a nagging suspicion > that there exists a pathological route that would violate the implicit > assumption that the the given conditions are equivalent to the condition you > really want, which is that no point within a given distance of the route be > outside of the union of all your circles. Clayjar and myself discussed how a spiral is probably the worst case scenerio. It would work, just be a lot of extra circles. I'm working out how to properly handle intersections now. If you blindly follow the line, you may end up broken, but I'll see what my trig can drag up now that I've had a refresher from ClayJar! Jason -- Jason Slagle /"\ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . \ / ASCII Ribbon Campaign . X - NO HTML/RTF in e-mail . / \ - NO Word docs in e-mail . ```
 [Gpsbabel-misc] Delorme Topo routes From: M. Bramel - 2005-07-19 20:37:51 ```Goal: get routes from Delorme Topo5 to explorist. This was a direct = serial communication on the Meridian, but T5 lacks USB explorist support at present. I'm trying to take the anr route files saved by Delorme Topo 5 and get = them into some sort of Magellan form. Using the GPSBabelWrapper, I ran = something like this command line: -w -i saroute -f "blah.anr" -o mapsend -F "blah.wpt" That appears to work ok, but the Magellan Conversion Manager gives me an error message when I try to send the wpt file to my explorist. In a fit = of what passes for brilliance for me, I tried to open the wpt file into = MapSend DirectRoute, and it refused it as either the wrong format or corrupted. Any insight, speculation, or general reflections on life? (I'm not too = bent out of shape if I just have to wait for DeLorme to catch up on this). Max ```
 Re: [Gpsbabel-misc] Delorme Topo routes From: Robert Lipe - 2005-07-19 21:12:53 ```On 7/19/05, M. Bramel wrote: > Goal: get routes from Delorme Topo5 to explorist. =20 The Topo formats are reverse engineered and change from version to version, so it's not shocking to me if we can't read any given Topo file. > I'm trying to take the anr route files saved by Delorme Topo 5 and get th= em > -w -i saroute -f "blah.anr" -o mapsend -F "blah.wpt" -w means "waypoint" yet you say you're taking a route. I just ran gpsbabel -r -i saroute -f track/i65.anr -o mapsend -F /tmp/blah.rte and that loaded into DR with only the warning that the number of routepoints had been exceeded. If I get froggy and do: gpsbabel -r -i saroute,turns_important -f track/i65.anr -x simplify,count=3D45 -o mapsend -F /tmp/blah.rte I get a perfectly lovely route in DR that hugs the road and contains 45 po= ints. (And THIS is why route filters are hard to do with a general purpose GUI - you really need special knowledge of both the input and output sides....) ```
 Re: [Gpsbabel-misc] Delorme Topo routes From: Ron Parker - 2005-07-19 21:25:55 ```At 04:11 PM 7/19/2005, Robert Lipe wrote: >On 7/19/05, M. Bramel wrote: > > Goal: get routes from Delorme Topo5 to explorist. > >The Topo formats are reverse engineered and change from version to >version, so it's not shocking to me if we can't read any given Topo >file. Topo 5 should be supported. Just today I discovered that I can WRITE an ANR file that can be understood by Topo 5 (though only if I start with an ANR file from some flavor of Street Atlas; there's too much data in an ANR file to synthesize one in the general case.) ```
 RE: [Gpsbabel-misc] Delorme Topo routes From: M. Bramel - 2005-07-19 23:38:38 ```It doesn't take much to amaze me, but consider me so rendered. I edited your command line for my system, and was able to get the goodies into DirectRoute as you and Ron foretold. There's still a little funny something happening...if I take the babel-generated wpt file and try to run it through the Conversion = Manager to the explorist, I still get an error. However, if I open said file into DirectRoute and then save it--without doing anything else to it--that = goes into the explorist just hunky-dory. I've found a path, I'm happy enough. Thankee, gentlemen. Max > -----Original Message----- > From: gpsbabel-misc-admin@...=20 > [mailto:gpsbabel-misc-admin@...] On Behalf=20 > Of Robert Lipe > Sent: Tuesday, July 19, 2005 5:12 PM > To: mbramel@... > Cc: gpsbabel-misc@... > Subject: Re: [Gpsbabel-misc] Delorme Topo routes >=20 >=20 > I just ran > gpsbabel -r -i saroute -f track/i65.anr -o mapsend -F /tmp/blah.rte >=20 > and that loaded into DR with only the warning that the number=20 > of routepoints had been exceeded. If I get froggy and do: >=20 > gpsbabel -r -i saroute,turns_important -f track/i65.anr -x=20 > simplify,count=3D45 -o mapsend -F /tmp/blah.rte >=20 > I get a perfectly lovely route in DR that hugs the road and=20 > contains 45 points. >=20 > (And THIS is why route filters are hard to do with a general=20 > purpose GUI - you really need special knowledge of both the=20 > input and output > sides....) >=20 ```
 Re: [Gpsbabel-misc] Delorme Topo routes From: Robert Lipe - 2005-07-20 05:34:58 ```M. Bramel wrote: > It doesn't take much to amaze me, but consider me so rendered. That's why we get the big bucks here... > There's still a little funny something happening...if I take the > babel-generated wpt file and try to run it through the Conversion > Manager to the explorist, I still get an error. However, if I open > said file into DirectRoute and then save it--without doing anything > else to it--that goes into the explorist just hunky-dory. Huh. So CM is more picky (or broken) about _something_ than DR is? Out of masochism, send me your wpt file and I'll put it on the microscope and see if I can figure out why it's angering CM. But didn't we establish that Conversion manager doesn't actually convert anything (well, at least for maps, tracks, waypoints, and routes) it just a big bloated wrapper for "copy"? RJL ```
 RE: [Gpsbabel-misc] Delorme Topo routes From: M. Bramel - 2005-07-20 13:57:36 ```> -----Original Message----- > From: Robert Lipe [mailto:robertlipe@...]=20 > Sent: Tuesday, July 19, 2005 11:41 PM > Huh. So CM is more picky (or broken) about _something_ than DR is? >=20 > Out of masochism, send me your wpt file and I'll put it on=20 > the microscope and see if I can figure out why it's angering CM. We aim to keep our masochists happy (file to follow). >=20 > But didn't we establish that Conversion manager doesn't=20 > actually convert anything (well, at least for maps, tracks,=20 > waypoints, and routes) it just a big bloated wrapper for "copy"? I don't see that this necessarily contradicts anything known/believed to date...although I haven't mucked with route files on the explorist until now. I was using CM to conveniently get the right file name endings and = get the files placed in the right folders. Is there any difference in a upt file and a corresponding rte file? Max ```
 Re: [Gpsbabel-misc] Delorme Topo routes From: Robert Lipe - 2005-07-20 15:39:49 ```> > Out of masochism, send me your wpt file and I'll put it on > We aim to keep our masochists happy (file to follow). Masochists are only happy when they're unhappy. (Think about it.) I've analyzed your file and have basically concluded that Conversion Manager is just being hyperactive. I've sent the analysis leading to that conclusion to engineering at Magellan. > Is there any difference in a upt file and a corresponding rte file? Not that I know of, but I haven't checked. Compare the output of CM with: gpsbabel -r -i mapsend -f MartShip.wpt -o magellan -F blah.rte -- Support GPSBabel by helping to improve it or fund those that that have done so. Visit: http://sourceforge.net/donate/index.php?group_id=58972 ```