Re: [OSM-talk] Map Co-ordinates for towns, etc in UK

2012-03-02 Thread mick 
On Fri, 02 Mar 2012 21:14:47 -0600
"John F. Eldredge"  wrote:

> mick  wrote:
> 
> > On Fri, 02 Mar 2012 20:16:18 +
> > Philip Barnes  wrote:
> > 
> > > I have found some interesting stuff whilst playing with routing on
> > > http://open.mapquest.org (which uses OSM). Have found that it cannot
> > > route to Shrewsbury. 
> > > 
> > > Have found that the town waypoint has been put in the middle of a
> > retail
> > > area, with pedestrianised streets around. Am guessing it is because
> > it
> > > is too far from a road. It works if I ask for High Street,
> > Shrewsbury.
> > > 
> > > Have moved it so that it is close to High Street and once mapquest
> > has
> > > updated the map I will see if it works again. But maybe this is
> > > something else we should consider.
> > > 
> > > A good example of TomTom/google getting it wrong is Ironbridge,
> > where it
> > > leads you into a back street, rather than the centre i.e. the
> > bridge.
> > > 
> > 
> > My Navman (MY50 I think) also has problems generating a route to just
> > a town or suburb without a street name.
> > 
> > At the times these towns developed very few people had a car and even
> > fewer had sat-nav units so the 'rule of thumb' didn't need to take
> > vehicular access into account. Now social engineers have had their
> > evil way, the 'rule' joins dinosaurs, woolly mammoths and this old
> > System/370 operator on the dusty shelves of the museum and its up to
> > the new generations to clean up our mess.
> > 
> > mick
> > 
> 
> Another issue you are likely to encounter is a town that has grown in an 
> asymmetric manner, so that the current geometric center is offset, perhaps by 
> a large amount, from the historic center point.  This is particularly true 
> where a natural barrier, such as a lake, adjoins the town.
> 
> Here in the USA, some small towns that have experienced most of their growth 
> during the automobile age are essentially one-dimensional, extending for 
> several miles along a main road, but extending only a block or two at right 
> angles to that main road.
> 
My original interest was if there was a specific point that said 'this is 
Sometown', where distances to adjacent towns were measured from, similar to the 
Australian convention where the "Zero Point" was set at the roadside, at the 
Post Office which was usually next door to or across the road from a 'coaching 
inn'.

This point rarely had anything to do with the geographic centre of town but 
served only as a survey benchmark.

As Phillip, yourself and a few other people have pointed out these points have 
little remaining relevance in current times, especially for routing.

The only place where I've found this concept still in use is Queensland Rail's 
Brisbane suburban network, where the track at stations is marked with the 
distance to Central Station and the markings are maintained.

mick

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

Re: [OSM-talk] Map Co-ordinates for towns, etc in UK

2012-03-02 Thread John F. Eldredge 
mick  wrote:

> On Fri, 02 Mar 2012 20:16:18 +
> Philip Barnes  wrote:
> 
> > I have found some interesting stuff whilst playing with routing on
> > http://open.mapquest.org (which uses OSM). Have found that it cannot
> > route to Shrewsbury. 
> > 
> > Have found that the town waypoint has been put in the middle of a
> retail
> > area, with pedestrianised streets around. Am guessing it is because
> it
> > is too far from a road. It works if I ask for High Street,
> Shrewsbury.
> > 
> > Have moved it so that it is close to High Street and once mapquest
> has
> > updated the map I will see if it works again. But maybe this is
> > something else we should consider.
> > 
> > A good example of TomTom/google getting it wrong is Ironbridge,
> where it
> > leads you into a back street, rather than the centre i.e. the
> bridge.
> > 
> 
> My Navman (MY50 I think) also has problems generating a route to just
> a town or suburb without a street name.
> 
> At the times these towns developed very few people had a car and even
> fewer had sat-nav units so the 'rule of thumb' didn't need to take
> vehicular access into account. Now social engineers have had their
> evil way, the 'rule' joins dinosaurs, woolly mammoths and this old
> System/370 operator on the dusty shelves of the museum and its up to
> the new generations to clean up our mess.
> 
> mick
> 

Another issue you are likely to encounter is a town that has grown in an 
asymmetric manner, so that the current geometric center is offset, perhaps by a 
large amount, from the historic center point.  This is particularly true where 
a natural barrier, such as a lake, adjoins the town.

Here in the USA, some small towns that have experienced most of their growth 
during the automobile age are essentially one-dimensional, extending for 
several miles along a main road, but extending only a block or two at right 
angles to that main road.

-- 
John F. Eldredge --  j...@jfeldredge.com
"Reserve your right to think, for even to think wrongly is better than not to 
think at all." -- Hypatia of Alexandria

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

Re: [OSM-talk] Map Co-ordinates for towns, etc in UK

2012-03-02 Thread mick 
On Fri, 02 Mar 2012 20:16:18 +
Philip Barnes  wrote:

> I have found some interesting stuff whilst playing with routing on
> http://open.mapquest.org (which uses OSM). Have found that it cannot
> route to Shrewsbury. 
> 
> Have found that the town waypoint has been put in the middle of a retail
> area, with pedestrianised streets around. Am guessing it is because it
> is too far from a road. It works if I ask for High Street, Shrewsbury.
> 
> Have moved it so that it is close to High Street and once mapquest has
> updated the map I will see if it works again. But maybe this is
> something else we should consider.
> 
> A good example of TomTom/google getting it wrong is Ironbridge, where it
> leads you into a back street, rather than the centre i.e. the bridge.
> 

My Navman (MY50 I think) also has problems generating a route to just a town or 
suburb without a street name.

At the times these towns developed very few people had a car and even fewer had 
sat-nav units so the 'rule of thumb' didn't need to take vehicular access into 
account. Now social engineers have had their evil way, the 'rule' joins 
dinosaurs, woolly mammoths and this old System/370 operator on the dusty 
shelves of the museum and its up to the new generations to clean up our mess.

mick

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

Re: [OSM-talk] problem when trying to use osmconvert / osmupdate / osmfilter to keep a filtered planet extract up to date

2012-03-02 Thread Martijn van Exel 
On Fri, Mar 2, 2012 at 6:52 PM, Martijn van Exel  wrote:
> Hi,
>
> I thought I had found a nice way to keep a filtered planet extract (by
> geographic extent *and* specific tags, in my case bridges in the US)
> up-to-date by using osmconvert[1], osmupdate[2] and osmfilter[3]. This
> was going to be my workflow:
>

I asked a similar question, but geared towards osmosis, some time ago on help:
http://help.openstreetmap.org/questions/4065/getting-specific-poi-data-and-keeping-them-up-to-date

I don't currently have read access to the part of my brain that holds
my memory of implementing Frederik's suggestions, but I will revisit
that as well.
Anyway, I thought it an interesting exercise to try and use this
toolset on a similar problem and gain some experience using the tools
mentioned here. They are very, very fast. Updating the US planet
extract with about a day's worth of changes takes ~3 minutes, the
filtering ~4 minutes (on an AMD Phenom 965 machine with 16GB RAM).

-- 
martijn van exel
geospatial omnivore
1109 1st ave #2
salt lake city, ut 84103
801-550-5815
http://oegeo.wordpress.com

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

[OSM-talk] problem when trying to use osmconvert / osmupdate / osmfilter to keep a filtered planet extract up to date

2012-03-02 Thread Martijn van Exel 
Hi,

I thought I had found a nice way to keep a filtered planet extract (by
geographic extent *and* specific tags, in my case bridges in the US)
up-to-date by using osmconvert[1], osmupdate[2] and osmfilter[3]. This
was going to be my workflow:

Setup steps:
A1. Grab a planet file
A2. Derive the extract using osmconvert
bzcat /osm/planet/planet-latest.osm.bz2 | ./osmconvert -
-B=/osm/poly/us.poly  -o=/osm/planet/us.o5m
A3. Filter the desired features into a new file
osmfilter /osm/planet/us.o5m --keep= --keep-ways="bridge=" --out-o5m >
/osm/planet/bridges.o5m
A4. Use osmosis to import the initial filtered planet extract into PostGIS.

Replication steps:
B1. Update the planet extract:
osmupdate /osm/planet/us.o5m /osm/planet/us-new.o5m -B=/osm/poly/us.poly
B2. Filter the desired features into a second file, analogous to step A3
B3. Derive an osc file from the two filter files using osmconvert.
osmconvert bridges.o5m bridges-new.o5m --diff -o=diffbridges.osc
B4. Using osmosis --rxc --wpc, update the PostGIS database
B5. Clean up

I started testing this and quite soon I hit an error in step B4:
SEVERE: Thread for task 1-rxc failed
java.lang.NullPointerException

Trying to load the osc file[4] into JOSM I got the same error.

I can't be sure but I think this has to do with referential integrity
issues caused by missing nodes outside the bounding polygon. The
--complete-ways option in osmconvert is supposed to solve that, but is
not available either when the input is STDIN (step A2) or when
osmconvert is called from osmupdate (step B1). The first case could be
solved by just getting the initial file via some other means. I don't
know how to solve the second case within the context of this set of
tools. Is that possible? Suggestions for an alternative toolchain to
accomplish this? Back to osmosis for the entire process, sacrificing
speed? Would I run into different problems?

[1] https://wiki.openstreetmap.org/wiki/Osmconvert
[2] https://wiki.openstreetmap.org/wiki/Osmupdate
[3] https://wiki.openstreetmap.org/wiki/Osmfilter
[4] This is the file that was created when testing the process and
gave the errors: https://gist.github.com/8001978303cb1077d7c2


-- 
martijn van exel
geospatial omnivore
1109 1st ave #2
salt lake city, ut 84103
801-550-5815
http://oegeo.wordpress.com

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

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Frans Thamura 
hehe, i see my self was brainwashed by API jargon :0

F

On Sat, Mar 3, 2012 at 5:51 AM, andrzej zaborowski wrote:

> On 2 March 2012 23:00, Frans Thamura  wrote:
> > does this mean that OSM API is not usefull for integration?
>
> What Toby says is they don't need to use the API.  They source the
> tiles from mapbox, who in turn use the OSM planet files or diffs as
> their interface to OSM.  And they use leaflet as their javascript
> thing.
>
> The OSM toolchain, like Unix, is naturally made up of small utilities
> each of which solves one problem at a time.
>
> Cheers
>
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread andrzej zaborowski 
On 2 March 2012 23:00, Frans Thamura  wrote:
> does this mean that OSM API is not usefull for integration?

What Toby says is they don't need to use the API.  They source the
tiles from mapbox, who in turn use the OSM planet files or diffs as
their interface to OSM.  And they use leaflet as their javascript
thing.

The OSM toolchain, like Unix, is naturally made up of small utilities
each of which solves one problem at a time.

Cheers

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

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Chris Hill 

On 02/03/12 22:07, Frans Thamura wrote:

i wish OSM can provide like what GMAP API can :0

rather using this model.

API wrapper to TILE will be interesting

I think you are confusing what Google call their API with what OSM call 
our API. They are not the same thing.


What Google call their API is JavaScript to display a map on a web page 
as a series of tiles. OSM use open tools like OpenLayers or Leaflet to 
do the same job.


OSM API retrieves and updates the data that is later used to create the 
tiles. Of course you cannot access the data Google has.


These are very different things. If you want to display OSM map tiles on 
a web page take a look at OpenLayers or Leaflet. You can see more about 
this in http://switch2osm.org/


--
Cheers, Chris
user: chillly


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

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Frans Thamura 
i wish OSM can provide like what GMAP API can :0

rather using this model.

API wrapper to TILE will be interesting

F


On Sat, Mar 3, 2012 at 5:03 AM, Tom MacWright  wrote:

> The OSM API(s) are certainly useful for integration, but a different kind
> - if they were pulling small chunks of data, etc., then they'd be using an
> API, but at this point they're mainly using tiles. More to come, but at
> this point the process looks like OSM Planet + update chunks -> TileMill
> rendering -> MapBox Hosting -> Foursquare.
>
> Tom
>
> On Fri, Mar 2, 2012 at 5:00 PM, Frans Thamura  wrote:
>
>> does this mean that OSM API is not usefull for integration?
>>
>> F
>>
>>
>>
>> On Sat, Mar 3, 2012 at 2:32 AM, Toby Murray wrote:
>>
>>> Frousquare is not using any OSM API. They are just using map tiles
>>> provided by MapBox with the leaflet library to display it in the
>>> browser.
>>>
>>> Toby
>>>
>>>
>>> On Fri, Mar 2, 2012 at 1:14 PM, Frans Thamura 
>>> wrote:
>>> > hi all
>>> >
>>> > we have great news that foursquare using OSM now
>>> >
>>> > anyone working with Fq? which API that osm using ?
>>> >
>>> > the ruby one?
>>>
>>
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/talk
>>
>>
>
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Tom MacWright 
The OSM API(s) are certainly useful for integration, but a different kind -
if they were pulling small chunks of data, etc., then they'd be using an
API, but at this point they're mainly using tiles. More to come, but at
this point the process looks like OSM Planet + update chunks -> TileMill
rendering -> MapBox Hosting -> Foursquare.

Tom

On Fri, Mar 2, 2012 at 5:00 PM, Frans Thamura  wrote:

> does this mean that OSM API is not usefull for integration?
>
> F
>
>
>
> On Sat, Mar 3, 2012 at 2:32 AM, Toby Murray  wrote:
>
>> Frousquare is not using any OSM API. They are just using map tiles
>> provided by MapBox with the leaflet library to display it in the
>> browser.
>>
>> Toby
>>
>>
>> On Fri, Mar 2, 2012 at 1:14 PM, Frans Thamura  wrote:
>> > hi all
>> >
>> > we have great news that foursquare using OSM now
>> >
>> > anyone working with Fq? which API that osm using ?
>> >
>> > the ruby one?
>>
>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk
>
>
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Frans Thamura 
does this mean that OSM API is not usefull for integration?

F


On Sat, Mar 3, 2012 at 2:32 AM, Toby Murray  wrote:

> Frousquare is not using any OSM API. They are just using map tiles
> provided by MapBox with the leaflet library to display it in the
> browser.
>
> Toby
>
>
> On Fri, Mar 2, 2012 at 1:14 PM, Frans Thamura  wrote:
> > hi all
> >
> > we have great news that foursquare using OSM now
> >
> > anyone working with Fq? which API that osm using ?
> >
> > the ruby one?
>
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Map Co-ordinates for towns, etc in UK

2012-03-02 Thread Philip Barnes 
I have found some interesting stuff whilst playing with routing on
http://open.mapquest.org (which uses OSM). Have found that it cannot
route to Shrewsbury. 

Have found that the town waypoint has been put in the middle of a retail
area, with pedestrianised streets around. Am guessing it is because it
is too far from a road. It works if I ask for High Street, Shrewsbury.

Have moved it so that it is close to High Street and once mapquest has
updated the map I will see if it works again. But maybe this is
something else we should consider.

A good example of TomTom/google getting it wrong is Ironbridge, where it
leads you into a back street, rather than the centre i.e. the bridge.

Phil


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

Re: [OSM-talk] FourSquare and OSM

2012-03-02 Thread Toby Murray 
Frousquare is not using any OSM API. They are just using map tiles
provided by MapBox with the leaflet library to display it in the
browser.

Toby


On Fri, Mar 2, 2012 at 1:14 PM, Frans Thamura  wrote:
> hi all
>
> we have great news that foursquare using OSM now
>
> anyone working with Fq? which API that osm using ?
>
> the ruby one?

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

[OSM-talk] FourSquare and OSM

2012-03-02 Thread Frans Thamura 
hi all

we have great news that foursquare using OSM now

anyone working with Fq? which API that osm using ?

the ruby one?

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

Re: [OSM-talk] 'Foursquare Says Farewell to Google Maps, Joins OpenStreetMap Movement'

2012-03-02 Thread Oleg 
Great choice! OSM is way better that oddy googlish scratches on the tiles
:) Rock'n'Roll!

2012/3/2 Maarten Deen 

> From 
> 
> >:
>
>  Foursquare is parting ways with Google Maps in favor of
>> crowdsourced maps created by the OpenStreetMap project.
>>
>
> I think this is a huge pat on the back for the OSM community.
>
> Maarten
>
>
> __**_
> talk mailing list
> talk@openstreetmap.org
> http://lists.openstreetmap.**org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk