Re: [OSM-talk] [Talk-us] Find missing roads

[Paul Johnson]

On Thu, Oct 1, 2015 at 10:58 AM, Martijn van Exel  wrote:

> Paul,
>
> The GPS data is from Scout (US as well as Skobbler / Global) users. It is
> collected during navigation sessions as well as in Scout Global ‘free
> drive’ mode.
>
> The fact that you see incomplete sets is to do with pretty high thresholds
> we set for the validity of a tile as a whole. We want to prevent that we
> present too many false positives that would make the experience frustrating.
>
> It turns out that Scout is in fact used a fair bit by commercial drivers,
> so you’ll see a bunch of tiles depicting access=private service roads on
> factory lots and so on. There’s not really a great way for us to filter
> those out up front, other than convincing drivers to shut down their nav
> session before they enter a private road ;)
>

I actually consider this a good thing since private roads are something
that are underserved right now.  It would be nice to be able to enable a
more complete data layer in order to see what was filtered out to avoid a
false positive, however, since it can be hard to determine if a driveway is
the one commonly used or if a driveway with a gate on it is manned or
simply there for the convenience of a fire crew to crash open in an
emergency (is my region the only one where fire trucks are equipped with
ramming bumpers and fire crews prepared to run over fences and locked
gates?)


> If you look closely you’ll also see traces from trains, airplanes etc.
> People use Scout in some places we had not really imagined! We try to
> filter these out based on speed and we’re improving the algorithm to filter
> out traces that match to railways in the future as well.
>

Granted, plane might not be very useful, but for the other modes, having
some kind of color coding to share what vehicle type is being used would be
awesome.

As a Scout user, I'd really find it handy if it had some of the warning
alarms and lane guidance that Osmand has...I'd really love to be able to
hold an Android app up as a real Garmin killer in terms of functionality.
Maybe include a truck mode where you can enter vehicle width, height,
weight and length so those restrictions can be taken into consideration as
well.  Osmand has a truck mode, given the people I've met on the road
driving a truck myself, I think Osmand in general is probably a lot too
fiddly for the segment of "commercial drivers who aren't GIS nerds", which
is only slightly larger than the segment of "commercial drivers who are not
me."
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Find missing roads

[Martijn van Exel]

Marc — you should be able to click on the ‘mark as fixed’ (lock) and ‘mark as 
invalid’ (exclamation mark) buttons in the missing roads panel after you make 
the missing roads layer active and select one or more tiles. Let me know if it 
doesn’t work for you.

Martijn

> On Oct 1, 2015, at 7:42 AM, Marc Zoutendijk  wrote:
> 
> Op 30 sep. 2015, om 17:11 heeft Martijn van Exel  > het volgende geschreven:
>> 
>> Take it for a spin and let me know what you think, what we could improve, or 
>> just if you like it!
> 
> I did and it works, but how can I indicate that I “solved” the tiles? In JOSM 
> the tabs with the lock cannot be unlocked.
> 
> Marc.
> 

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] [Talk-us] Find missing roads

[Martijn van Exel]

:D

We may need to add this to the disclaimer: ‘may contain traces of nuts’!

Martijn

> On Sep 30, 2015, at 11:45 PM, Jack Burke  wrote:
> 
> Clearly, your sample contained nuts. 
> 
> 
> On September 30, 2015 2:40:37 PM EDT, Paul Johnson  
> wrote:
> On Wed, Sep 30, 2015 at 10:11 AM, Martijn van Exel  > wrote:
> Our OSM team cooked up something new. A missing roads plugin for JOSM. I 
> think it's pretty nice but I would really like to hear what you think. 
>  
> Neat, has potential.
> 
> You can read some more about it on my diary (http://bit.ly/missingroads 
> ) but it's basically what it says on the tin. The 
> plugin will show where we think roads are missing from OSM based on GPS data 
> so you can add them :)
> 
>  Based on whose GPS data from where?  I trust the source in this question is 
> completely halal, kosher, gluten-free, vegan, organic, free-range, 
> all-natural data, but some of the output I'm getting from it is rather odd.  
> Take, for example, this screenshot in JOSM near node 148319848 
>  in my 
> neighborhood.
> 
> 
> ​
> Other than highlighting my own inadvertently selective blindness (what with 
> having not mapped the large and aging chain link fence factory on the 
> northeast corner of the intersection; Brinks behind my favorite QuikTrip 
>  only 
> recently appeared on Bing), I do find two things remarkable about this 
> plugin's output:
> It seems to have picked out an incomplete set based on the paths relative to 
> imagery.
> I have no way of being able to survey the exact location of the GPS output 
> from the plugin from the ground (it's inside a fence factory, of course it's 
> fenced off!), so I can only assume the GPS was located on one of those big 
> diesel-powered forklifts.
> 
> 
> 
> Talk-us mailing list
> talk...@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-us 
> 
> 
> -- 
> Typos courtesy of fancy auto-spell technology.
> ___
> Talk-us mailing list
> talk...@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-us

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] [Talk-us] Find missing roads

[Martijn van Exel]

Paul,
 
The GPS data is from Scout (US as well as Skobbler / Global) users. It is 
collected during navigation sessions as well as in Scout Global ‘free drive’ 
mode.

The fact that you see incomplete sets is to do with pretty high thresholds we 
set for the validity of a tile as a whole. We want to prevent that we present 
too many false positives that would make the experience frustrating.

It turns out that Scout is in fact used a fair bit by commercial drivers, so 
you’ll see a bunch of tiles depicting access=private service roads on factory 
lots and so on. There’s not really a great way for us to filter those out up 
front, other than convincing drivers to shut down their nav session before they 
enter a private road ;)

If you look closely you’ll also see traces from trains, airplanes etc. People 
use Scout in some places we had not really imagined! We try to filter these out 
based on speed and we’re improving the algorithm to filter out traces that 
match to railways in the future as well.

Thanks for sharing your observations!
Martijn

> On Sep 30, 2015, at 12:40 PM, Paul Johnson  wrote:
> 
> On Wed, Sep 30, 2015 at 10:11 AM, Martijn van Exel  > wrote:
> Our OSM team cooked up something new. A missing roads plugin for JOSM. I 
> think it's pretty nice but I would really like to hear what you think. 
>  
> Neat, has potential.
> 
> You can read some more about it on my diary (http://bit.ly/missingroads 
> ) but it's basically what it says on the tin. The 
> plugin will show where we think roads are missing from OSM based on GPS data 
> so you can add them :)
> 
>  Based on whose GPS data from where?  I trust the source in this question is 
> completely halal, kosher, gluten-free, vegan, organic, free-range, 
> all-natural data, but some of the output I'm getting from it is rather odd.  
> Take, for example, this screenshot in JOSM near node 148319848 
>  in my 
> neighborhood.
> 
> 
> ​
> Other than highlighting my own inadvertently selective blindness (what with 
> having not mapped the large and aging chain link fence factory on the 
> northeast corner of the intersection; Brinks behind my favorite QuikTrip 
>  only 
> recently appeared on Bing), I do find two things remarkable about this 
> plugin's output:
> It seems to have picked out an incomplete set based on the paths relative to 
> imagery.
> I have no way of being able to survey the exact location of the GPS output 
> from the plugin from the ground (it's inside a fence factory, of course it's 
> fenced off!), so I can only assume the GPS was located on one of those big 
> diesel-powered forklifts.
> 
> ___
> Talk-us mailing list
> talk...@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-us

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

[OSM-talk] ODBL, Trivial_Transformations and Specialised format databases

[THEVENON Julien]

Hi all,

I ask myself a question about specialised proprietary format database used in 
some applications like by example smartphone routing applications.
Quite often these applications propose you to dowwnload some map files if you 
want to use them Offline.
The issue I see with this is that user is dependant of the application provider 
for map update or scope cover by map extract.
I looked at this page : 
http://wiki.openstreetmap.org/wiki/Open_Data_License/Trivial_Transformations_-_Guidelineand
 especially part called Specialised (e.g: mobile) format databases

It is said that this is a trivial transformation and that format could be kept 
closed if data is also provided in an open format.
Ok but how to be sure that data provided in Open Format is exactly the same 
than the one contained in closed format ? by example how to check that non-OSM 
data have not been included in closed format ?

To make a parallell ( perhaps wrong ), GPLv3 requires that someone providing a 
binary using GPL code must provide the source code ( including modifiactions if 
any ) eveything needed to rebuild the binary.
"Both versions of the GPL require you to provide all the source necessary to 
build the software, including supporting libraries, compilation scripts, and so 
on. They also draw the line at System Libraries: you're not required to provide 
the source for certain core components of the operating system, such as the C 
library."
source: http://www.gnu.org/licenses/quick-guide-gplv3.en.html
and "GPLv3 stops tivoization by requiring the distributor to provide you with 
whatever information or data is necessary to install modified software on the 
device."



Do the ODBL requires the same for a closed format including OSM data ? or more 
clearly does it require that application provider also provide the tool used to 
perform the conversion between OSM data and closed format ?
If yes, by consequence this would allow user to be independant of application 
provider and to regenerate by itself the maps directly from OSM data.

If this topic has already been covered somewhere sorry and I'll please you to 
provide me some pointers to the discussion

Thanks by advance for you answer
Cheers
Julien

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Find missing roads

[Marc Zoutendijk]

Op 30 sep. 2015, om 17:11 heeft Martijn van Exel  het volgende 
geschreven:
> 
> Take it for a spin and let me know what you think, what we could improve, or 
> just if you like it!

I did and it works, but how can I indicate that I “solved” the tiles? In JOSM 
the tabs with the lock cannot be unlocked.

Marc.

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Traffic Signs

[Florian Lohoff]

On Wed, Sep 30, 2015 at 04:04:32PM +0200, Martin Koppenhoefer wrote:
> sent from a phone
> 
> > Am 30.09.2015 um 14:57 schrieb Florian Lohoff :
> > 
> > Given that mapillary stays i can
> > go back to those photos and put details in the map i am not even
> > thinking about today
> 
> besides that your photos will get old like any imagery and not reflect the 
> future state for many of these details that you then want to map ;-)

How often do street lights get moved? I guess the ones around my 
house are there for like 30-50years ... 

You are right that there are certain aspects which changes much more
often like Pubs etc. But there are aspects which change only each
century.

Flo
-- 
Florian Lohoff f...@zz.de
 We need to self-defense - GnuPG/PGP enable your email today!


signature.asc
Description: Digital signature
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk