Free maps for Garmin

[andygates]

Added to my to-do. 

[andygates]

(There will be a slight delay to updates while I wonder where Doncaster has gone.  Talk among yourselves...)

[David Martin]

(There will be a slight delay to updates while I wonder where Doncaster has gone.  Talk among yourselves...)

Does anyone care?

(ducks and runs)

[piedwagtail91]

(There will be a slight delay to updates while I wonder where Doncaster has gone.  Talk among yourselves...)

i haven't seen anything on the news about that, unless of course you mean it's just gone from the map!
thanks for all the good work.

[andygates]

Ah, it was me having a brainfart, that's all.  Normal service is resumed.

[piedwagtail91]

yes, just downloaded and installed! thanks

[amaferanga]

I used the maps provided by andygates today for the first time.  I've been using Garmin City Navigator, but thought I'd give these a try.  Overall I'm impressed - I like the look of the maps and there weren't any missing roads on my route today.

But I do have an issue - there were a few stretches of road, usually where there was tree cover, where the road on the map was up to about 20 metres away from where my GPS reckoned I was.  I was following a track plotted on Google maps (bikehike) and it coincided exactly with the road I was on.  So I guess whoever mapped this road had pretty poor GPS accuracy at the time.

This got me thinking - do people actually need to cycle or drive down a road in order to map it in OSM, or can you just plot a route in (e.g.) bikhike, save it as a .gpx track and then use that to fill in OSM?  This approach would seem to me to avoid the above issues with people adding roads when their GPS accuracy was poor.  So do people actually do this?

[David Martin]

This got me thinking - do people actually need to cycle or drive down a road in order to map it in OSM, or can you just plot a route in (e.g.) bikhike, save it as a .gpx track and then use that to fill in OSM?  This approach would seem to me to avoid the above issues with people adding roads when their GPS accuracy was poor.  So do people actually do this?

No, because that is copying copyright data. You have to map from scratch.

With thenew OpenDataintiative from OS, you should be able to map by tracing the open map data.

[amaferanga]

This got me thinking - do people actually need to cycle or drive down a road in order to map it in OSM, or can you just plot a route in (e.g.) bikhike, save it as a .gpx track and then use that to fill in OSM?  This approach would seem to me to avoid the above issues with people adding roads when their GPS accuracy was poor.  So do people actually do this?

No, because that is copying copyright data. You have to map from scratch.

With thenew OpenDataintiative from OS, you should be able to map by tracing the open map data.

Of course.  Well maybe people should check the accuracy of their recorded tracks using Google maps before adding to the OSM maps?  No point in having an inaccurate map....

[Polar Bear]

You can always record and update OSM.   GPS isn't touted as pinpoint accurate.

I recall being in a car being driven down the M5 with my Garmin GPS60CSx.  It was consistently showing that we were travelling to the west of the M5 on the mapping provided with the unit.

[fuaran]

Of course.  Well maybe people should check the accuracy of their recorded tracks using Google maps before adding to the OSM maps?  No point in having an inaccurate map....

That's still like copying from Google Maps. Also Google Maps can be rather inaccurate in places.
The best way to get an accurate map is (usually) to get multiple GPS traces for each road, preferably several in each direction, and at different times / days etc. Then draw the way as an average of these, which should average out any errors.

Also it depends on how accurate it has to be anyway. You say there was a 20m error - you won't notice that unless you are using a GPS device, and zoomed right in. If its on something like a 1:25,000 printed map its less than 1mm difference, which is about the same as the width or the road.

[Tewdric]

I built my map set from Worldwide routable Garmin maps from OpenStreetMap and one of the downloads that it provides is a MapSource installer. I just ran the executable provided and then selected that map from the MapSource menu as described above. I'm not using Win7 but I didn't have to type any techy commands. Hope that helps, sorry if I missed the point.

I use the lovely munkymap on the GPS but for mapsource purposes the above source is close enough - I don't think it has the little postboxes but it has the little foaming beer glasses for pubs 

Thanks - worked for me!

[amaferanga]

Of course.  Well maybe people should check the accuracy of their recorded tracks using Google maps before adding to the OSM maps?  No point in having an inaccurate map....

That's still like copying from Google Maps. Also Google Maps can be rather inaccurate in places.
The best way to get an accurate map is (usually) to get multiple GPS traces for each road, preferably several in each direction, and at different times / days etc. Then draw the way as an average of these, which should average out any errors.

Also it depends on how accurate it has to be anyway. You say there was a 20m error - you won't notice that unless you are using a GPS device, and zoomed right in. If its on something like a 1:25,000 printed map its less than 1mm difference, which is about the same as the width or the road.

On my GPS (at 200ft scale) there was about a 1cm difference between where I was (on the road) and where the OSM map had the road and that lasted for about 1km.  It was enough of an error that If I didn't know the area I'd have been wondering if I was actually on the road I had planned to be on.  In contrast, with City Navigator and routes plotted with Google maps I've never noticed any significant discrepancies.

I'm not trying to knock OSM as I do really like the look of andgates' maps, but little things like this might turn me back to Garmin's maps.

[piedwagtail91]

i had the same probelm on todays ride. i followed my ride planned on tracklogs as i usually do and was spot on the tracklogs  track but way off the osm in a few places.
when i loaded the rides trace into osm ( not the traklogs one!) i just moved the osm road to my gps trace. should be ok next time i go round.

[nuttycyclist]

...
I'm not trying to knock OSM as I do really like the look of andgates' maps, but little things like this might turn me back to Garmin's maps.

I use Garmin maps.  If I turn off the "follow road" feature then I'm not always on the mapped road (more than one device).

Out mapping the other day with my Edge 605 I found that the bridleway I was coming back along was 120feet away from where I'd mapped it an hour earlier.

Mapping with that Edge sometimes had me on top of the Etrex GPS tracks I'd uploaded to OSM the previous week.

However when mapping the area I used one road many times, so all those GPS traces were uploaded.  All varied slightly but the average (of others as well as mine) gave a good indication of where the road went.


I get cross that people rely on GPS believing it gives pin point accuracy.

[amaferanga]

...
I'm not trying to knock OSM as I do really like the look of andgates' maps, but little things like this might turn me back to Garmin's maps.

I use Garmin maps.  If I turn off the "follow road" feature then I'm not always on the mapped road (more than one device).

Out mapping the other day with my Edge 605 I found that the bridleway I was coming back along was 120feet away from where I'd mapped it an hour earlier.

Mapping with that Edge sometimes had me on top of the Etrex GPS tracks I'd uploaded to OSM the previous week.

However when mapping the area I used one road many times, so all those GPS traces were uploaded.  All varied slightly but the average (of others as well as mine) gave a good indication of where the road went.


I get cross that people rely on GPS believing it gives pin point accuracy.

I must have missed the post where someone said they believe GPS always gives pinpoint accuracy, so who are you getting cross with? 

I thought we were discussing the accuracy of maps and my point was that I was on the road, my GPS indicated I was on the road according to Google maps (based on the fact that I was bang on the track that I'd drawn in bikehike along that road), but according to OSM I was a long way from the road.  Perhaps you need to get cross with the person who added that stretch of road as they clearly didn't map it numerous times and then take an average.  Tut tut....

[sbseven]

Back to your discussion, I've only ever seen one OSM mapped road seriously out of wack with my road position in the 18 months I've had it on my Garmin. So I corrected it. Virtually everything I've ridden on using the mapping appears to be within GPS tolerances and I imagine, virtually all of it was created via one pass rather than the multiple passes mentioned above.

I often look at the map while I'm riding past features (bridges, post boxes, farm tracks etc) and they always seem to be where they should be.

Clearly the OSM mapping is open to human error because of its collaborative nature. It's possible since you found several roads not where they should be, they were all added by the same person with either poor technique, a poor GPS reading or even maliciously. You can check if the same person was responsible as there's a change history available for each way (track/road).

You may notice discrepancies if you zoom right in on the map, but at a reasonable 100m/150m zoom, any error due to GPS inaccuracy (the map creator and yours combined) should be bearly detectable.

Shaun

[piedwagtail91]

i agree with sbseven. i only move roads that are outside the circle of error on the gps. there aren't that many outside that range. When i check the accuracy of the gps it's about 12 feet, which if thats anywhere near right is not a problem.

[andygates]

OSM mapping is declaredly "good enough" not "pinpoint accurate".  Pinpont accurate requires surveyors and expensive kit and is incredibly slow.  The wiki nature means people will sketch things in sometimes, before nailing 'em down.

Ceci n'est pas le cadastre.

[piedwagtail91]

thats a good point andy, i know i've put a couple of roads in cos i know they exist ,then  a week or so later ridden down them to correct them.

[nuttycyclist]

...  Pinpont accurate requires surveyors and expensive kit and is incredibly slow. ...

... and is only pinpoint accurate if the end user has the correct quality equipment to use the map.  The google maps mentioned upthread are actually only a data gather by google from other mapping sources. 

I've had problems with GPS not actually tracking the road to pin point accuracy (with "follow road" turned off) on google maps and Garmin maps.  I've heard quite often of plans designed on paper that didn't work out on the ground (re my thread on the Southend Cycle scheme they found that when they painted the central bays they didn't actually fit on the road because the surveyed plans were a different scale to real life.)

[andygates]

Sure. As is said upthread, GPS isn't pefectly accurate.  Remember, the little circle around your position isn't really accuracy, it's confidence.  "I'm pretty sure you're in this circle, probably around the middle."

Sometimes a whole GPS session can be off - it looks like it's slipped North 10m, say.  Most of the time, after a couple of hours of operation, it's very repeatable and that's as good as we can hope to get.  The device is very confident of its silly position. 

When there's only one trace, unless you've got other cues to tell you that it's off (like known junctions all being shifted the same way), you've got to treat the GPS trace as authoritative: It's the only data you have.  Apply a brain-based smoothing algorithm ("that bit was straight really") and cartographic elegance and you're winning.

When I'm mapping a burb, I like to walk / ride the main roads a couple of times.  It usually happens by default just through "getting" everything, and it makes it easier to see where the real centre of a junction is, or the real shape of a curve.

[nuttycyclist]

...  Pinpont accurate requires surveyors and expensive kit and is incredibly slow. ...

... and is only pinpoint accurate if the end user has the correct quality equipment to use the map.  The google maps mentioned upthread are actually only a data gather by google from other mapping sources. 

...

Which reminds me.  Google streetview has the village names completely wrong in my area (half the borough has been lost!)

If it was OSM I'd go and correct it.  As it's google all I've been able to do is to report a problem and hope that somebody will pick up that report and go back to the source data providers and ask them to correct their data.

[David Martin]

Which reminds me.  Google streetview has the village names completely wrong in my area (half the borough has been lost!)

If it was OSM I'd go and correct it.  As it's google all I've been able to do is to report a problem and hope that somebody will pick up that report and go back to the source data providers and ask them to correct their data.

Good luck. I reported that my street was named incorrectly over a year ago and it still hasn't been fixed..

..d

[frankly frankie]

The google maps mentioned upthread are actually only a data gather by google from other mapping sources.  

AIUI, Google Maps now originate their own data, and are using that wherever possible.  What you currently see is a mix of original data and data gathered from the other major sources.  They are effectively setting up as a 3rd major data source alongside Navtech and Tele Atlas (and undermining their business model).  
The endgame could render the OSM project rather futile, unless OSM actually succeeds in being better (rather than just good and free).