Re: [Talk-us] Unpaved streets

2009-09-07 Thread Paul Johnson
On Mon, 2009-09-07 at 11:40 -0700, Apollinaris Schoell wrote:
>  I think people most  
> tag according to importance anyway.
> this brings us to the topic of tiger import again. there are too many  
> roads tagged as residential.

Agreed.  You'd think after people complaining after the first one, the
availability of the generic highway=road and the fact TIGER calls
everything it can't identify as a residential street regardless of real
status would have been some clues not to make the same mistake twice.


signature.asc
Description: This is a digitally signed message part
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


[Talk-us] Another shot at a whole-US Garmin map

2009-09-07 Thread Dave Hansen
I decided to try another approach.  I used the tile picker from here:

http://ulrichkuester.de/OSM/CoordinateToOSMTile.html

and got the tiles from here:

http://osm.ammit.de/osm/latest/img/

with a little wget and a bash script.  I then used the mkgmap commands
what Lambertus talks about here:

http://forum.openstreetmap.org/viewtopic.php?id=2625&p=10

java -Xmx2048M -jar mkgmap.jar 
--series-name='OSM World Routable' 
--latin1 
--code-page-1252 
--description='OSM World Routable' 
--product-id=53 
--tdbfile 
--gmapsupp *.img

The result is here:


http://dev.openstreetmap.org/~daveh/garmin/USA-fromtiles.09-06-2009.gmapsupp.img.bz2

It looks quite a bit better than the previous one that I posted.  I
think those tiles have a bit more detail and don't have the error
annotations that the Cloudmade ones do.  It still doesn't want to do
real routing on my Nuvi, but at least the map looks OK.

-- Dave


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


Re: [Talk-us] Is there a solution to Yahoo WMS poor image quality?

2009-09-07 Thread Michel Marti
Alan Mintz wrote:
> As is well known apparently, at some point in the last couple of months, 
> the Yahoo imagery available through WMS became pixelated and difficult to 
> use at zoom levels high enough to trace individual streets.
[snip]
> 2. Is there a better solution (in places where USGS urban imagery is not 
> available)? Is anyone working on one?

Yes, use Potlatch... See this JOSM TRAC Ticket:

http://josm.openstreetmap.de/ticket/1868

Cheers,

- Michel


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


Re: [Talk-us] question: relations & divided highways

2009-09-07 Thread Apollinaris Schoell
+1
given the length of some interstates this is highly recommended.


On 7 Sep 2009, at 12:12 , Richard Welty wrote:

> given that there is apparent concensus that Interstate relations be  
> done
> on a state-by-state
> basis, perhaps the language on the Interstate_Highways_Relations page
> should be updated:
>
> Avoid relation proliferation, if possible. If a relation already  
> exists
> for the route you are
> tagging, you can reuse the existing relation in your area. In  
> Potlatch,
> do a relation search
> on the existing relation's number.
>
> perhaps should become:
>
> Avoid relation proliferation, if possible. Interstate Highway  
> relations
> should be on a state-by-state
> basis, e.g. I-90 in MA, I-90 in NY, I-90 in PA, and so forth.
> If a suitable relation already exists for the route you are tagging,  
> you
> can reuse the existing relation
> in your area. In Potlatch, do a relation search on the existing
> relation's number.
>
>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us


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


Re: [Talk-us] question: relations & divided highways

2009-09-07 Thread Richard Welty
given that there is apparent concensus that Interstate relations be done 
on a state-by-state
basis, perhaps the language on the Interstate_Highways_Relations page 
should be updated:

Avoid relation proliferation, if possible. If a relation already exists 
for the route you are
tagging, you can reuse the existing relation in your area. In Potlatch, 
do a relation search
on the existing relation's number.

perhaps should become:

Avoid relation proliferation, if possible. Interstate Highway relations 
should be on a state-by-state
basis, e.g. I-90 in MA, I-90 in NY, I-90 in PA, and so forth.
If a suitable relation already exists for the route you are tagging, you 
can reuse the existing relation
in your area. In Potlatch, do a relation search on the existing 
relation's number.



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


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Apollinaris Schoell
full support for all your arguments.

there are a few places where the wiki is different and I think we  
should change these definitions.

1) National Forest Service, Bureau of Indian Affairs, Power  
Administration, and Bureau of Land Management routes
.
highway (required) track or tertiary depending on unpaved or paved  
respectively.
this should be changed to  track,tertiary,service,unclassified  
depending on the class and importance of the road

the main wiki for highway was updated recently too and  
United_States_roads_tagging page should follow. I think people most  
tag according to importance anyway.
this brings us to the topic of tiger import again. there are too many  
roads tagged as residential. We should try to find a way to identify  
all residential roads without any address and downgrade.
not sure if tiger data contains some attributes to make this possible.
on a smaller scale roads in national parks, forests, other park areas  
are never residential with very few exceptions. changing them can be  
done easily by any mapper. highest tag for unnamed  roads should be  
service and in a lot of cases only track.
If others agree we should put it on the tiger cleanup pages.
not everyone is subscribed to the talk list.



On 7 Sep 2009, at 6:41 , Greg Troxel wrote:

> Paul Johnson  writes:
>
>> OK, but can you drive them in a PT Cruiser Touring Edition, a Porche
>> 911, or some other low-slung, stiff-suspension vehicle on it without
>> problems?  How about a road bicycle (which absolutely depends on
>> pavement)?  Probably not going to be happening.
>
> Yes, you'd be ok in those on some roads, and of course not on some.
>
> But 'residential' isn't about road quality, it's about whether a  
> road is
> a "public way" or "private way" vs. something that happens to exist  
> in a
> farm, forest, or some other sub-legal-road setting.
>
> In Massachusetts, roads (and private ways, but not driveways or farm
> tracks) are separate lots on assessor's plots and the at the  
> registry of
> deeds.
>
> At
>
>  http://wiki.openstreetmap.org/wiki/Key:highway
>
> track says:
>
>  Roads for agricultural use, gravel roads in the forest etc.; usually
>  unpaved/unsealed but may occasionally apply to paved tracks as  
> well, see
>  tracktype=* for more guidance.
>
>
> Most proprietary maps I've seen show only what I'd call
> highway=residential and up.  Driveways and tracks in forests are
> omitted.  I think that's because they are car navigation centric, and
> also because of the legal classification of roads.
>
>
>
> I think the real problem here is that different people want to attach
> different semantics to tags, but we have nonorthogonal tags.  For  
> roads,
> there are two mostly orthogonal concerns:
>
>  legal status
>
>  physical condition
>
> I am arguing that highway=residential speaks to legal status but  
> doesn't
> say much about physical condition, and that physical condition tags  
> are
> needed.  If I understand you correctly, you would label a track an
> unpaved road that is a distinct parcel owned by the town, has a name,
> houses with numbers and addresses on that road (and appears as a  
> road on
> most other maps).  I see there is some support for this notion at
>
>  http://wiki.openstreetmap.org/wiki/United_States_roads_tagging
>
> and wonder if it is new - I don't remember seeing that on my previous
> reading, and I don't think it's the right thing to do.  My wiki-fu is
> too weak to do 'svn blame' on the source...
>
>
> Separately, we need an equivalent tag to residential for roads that  
> are
> less than unclassified but don't meet the residential notion.
>
>
> ___
> Talk-us mailing list
> Talk-us@openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-us


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


Re: [Talk-us] TopOSM Colorado

2009-09-07 Thread Lars Ahlzen
Lars Ahlzen wrote:
> Either way, I guess the thing for me to do would be to clean things up 
> just enough to be readable and then put the code and instructions up on 
> the wiki. Then we can all play with it. :)

If you want to try it out, I put the scripts and other files up at 
http://toposm.com/src

There are some very brief instructions at 
http://wiki.openstreetmap.org/wiki/TopOSM_howto

It's still pretty unpolished, but should be enough for the adventurous. :)

Let me know if I forgot anything.

- Lars

-- 
Lars Ahlzen
l...@ahlzen.com

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


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Bill Ricker
On Mon, Sep 7, 2009 at 3:20 AM, Paul Johnson wrote:
>> There are streets in my town (Stow, MA) that are definitely not paved
>> that I have driven on with a Saturn SC2 and not had a problem.
>
> OK, but can you drive them in a PT Cruiser Touring Edition, a Porche
> 911, or some other low-slung, stiff-suspension vehicle on it without
> problems?


Safer there than primary streets in Boston. A friend lost money on
every test drive when rebuilding Lotus Europa cars Boston as he had to
replace at least one strut. Between utility dig crews' bad patches and
Jack Frost's po-tholes and frost heaves, our paved streets are often
in worse condition than unpaved -- which can be much more cheaply
resurfaced in the spring with a grader.

However the Condition=Intolerable on Greg's Stowe streets sounds like
they've not been regraded and while there may be clearance for a
Porsche or Lotus (not so rutted at to be two wheel tracks), it might
be as hazardous for the sporty stiff suspension and decorative hubcaps
as my city streets.

Surface=gravel would be a data improvement to much of the tiger
imports in my original hometown upcountry in Maine, but I've heard
some of the winterized cottage neighborhoods have been paved in the
last few decades, either being converted from private to public
ownership or by subscription. Alas the cost to pave the one that would
improve the value of the backside of Dad's woods far exceeds the value
he'd gain.
-- 
Bill in Boston
n1...@arrl.net bill.n1...@gmail.com

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


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Greg Troxel
Paul Johnson  writes:

> OK, but can you drive them in a PT Cruiser Touring Edition, a Porche
> 911, or some other low-slung, stiff-suspension vehicle on it without
> problems?  How about a road bicycle (which absolutely depends on
> pavement)?  Probably not going to be happening.

Yes, you'd be ok in those on some roads, and of course not on some.

But 'residential' isn't about road quality, it's about whether a road is
a "public way" or "private way" vs. something that happens to exist in a
farm, forest, or some other sub-legal-road setting.

In Massachusetts, roads (and private ways, but not driveways or farm
tracks) are separate lots on assessor's plots and the at the registry of
deeds.

At

  http://wiki.openstreetmap.org/wiki/Key:highway

track says:

  Roads for agricultural use, gravel roads in the forest etc.; usually
  unpaved/unsealed but may occasionally apply to paved tracks as well, see
  tracktype=* for more guidance.


Most proprietary maps I've seen show only what I'd call
highway=residential and up.  Driveways and tracks in forests are
omitted.  I think that's because they are car navigation centric, and
also because of the legal classification of roads.



I think the real problem here is that different people want to attach
different semantics to tags, but we have nonorthogonal tags.  For roads,
there are two mostly orthogonal concerns:

  legal status

  physical condition

I am arguing that highway=residential speaks to legal status but doesn't
say much about physical condition, and that physical condition tags are
needed.  If I understand you correctly, you would label a track an
unpaved road that is a distinct parcel owned by the town, has a name,
houses with numbers and addresses on that road (and appears as a road on
most other maps).  I see there is some support for this notion at

  http://wiki.openstreetmap.org/wiki/United_States_roads_tagging

and wonder if it is new - I don't remember seeing that on my previous
reading, and I don't think it's the right thing to do.  My wiki-fu is
too weak to do 'svn blame' on the source...


Separately, we need an equivalent tag to residential for roads that are
less than unclassified but don't meet the residential notion.


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


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Richard Weait
Hi Paul and Greg and talk-us,

Your points regarding tagging of gravel roads, highway vs. path vs.
trail, and vehicle choice are points that come up in the OSM community
from time to time.  For further discussion and to see how previous
discussions have resolved, consider looking up the discussion on the
smoothness tag, the thread on track vs. path on the talk-list, and the
very excellent presentation from SotM by Peter Miller called Community
Smoothness.

In many cases these issues boil down to a couple of points.  First, a
map is an abstraction and can never reflect all of reality in pristine
perfection. Second, we as mappers can't be expert in all of the ways
to map all of the wide world.  Third, we can not predict all of the
ways our data will be used by the present and future users of OSM.

Having One True Way to map each thing makes no sense because it fails
to consider context.  And context changes, sometimes drastically,
depending on location.  So while you are discussing specific roads in
Massachusetts and Oregon, and finding the things are a bit different
for each of you, imagine how your argument applies to roads at McMurdo
Station on Antarctica or in a rain forest, or on a desert?  Things are
Just Too Different.

So we as a community have to trust ourselves.  We have to trust our
local experts.  We have to grow more local experts by having these
conversations and improving how we tag things in the local context and
in the wider context.  And in some ways we have to Just Do Our Best.
It won't always be perfect and that is okay.

There is a history in OSM of saying that if you tag it to reflect the
reality on the ground, then that is okay.  Even if it is unusual.

So if highway=residential; surface=gravel/unpaved/unsurfaced makes
sense?  Use it.

If highway=track; track_type=2 makes sense?  Use it!

And don't let anybody tell you that you are wrong, just because they
wouldn't do it that way.

So, by all means, continue the discussion.  You may both be right.

Best regards,
Richard

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


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Paul Johnson
On Sun, 2009-09-06 at 21:49 -0400, Greg Troxel wrote:
> Paul Johnson  writes:
> 
> > On Fri, 2009-09-04 at 07:30 -0400, Greg Troxel wrote:
> >
> >> It depends on what the road is like.  If it's a decent dirt road that
> >> normal cars routinely drive on, has a street name, is considered a
> >> public or private way by the town, then it's highway=residential
> >> surface=unpaved.
> >> 
> >> If when driving on it in a car you wince and wish you had a 4WD truck,
> >> it doesn't a name, and it isn't recognized as a 'real road', it might be
> >> highway=track.
> >
> > How are you coming to that conclusion, anyway?  Also, show me an unpaved
> > residential street in the western US that you didn't wish you had a 4WD
> > for, and I'll show you a street that really is paved, but hasn't been
> > swept or had rain wash the dirt off all summer.
> 
> There are streets in my town (Stow, MA) that are definitely not paved
> that I have driven on with a Saturn SC2 and not had a problem.

OK, but can you drive them in a PT Cruiser Touring Edition, a Porche
911, or some other low-slung, stiff-suspension vehicle on it without
problems?  How about a road bicycle (which absolutely depends on
pavement)?  Probably not going to be happening.



signature.asc
Description: This is a digitally signed message part
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us


Re: [Talk-us] Unpaved streets

2009-09-07 Thread Paul Johnson
On Sun, 2009-09-06 at 21:12 -0700, Dion Dock wrote:
> >
> >> Hello,
> >>
> >> In my neighborhood there are some unpaved residential streets. How
> >> should those be tagged?
> >>
> >> highway=track
> >>
> >> highway=residential
> >> surface=unpaved
> >
> > I would tend to go with highway=track unless the street in question  
> > is a
> > gravelled over macadam or some other semi-paved surface mostly because
> > most routing engines and all renderers at this point are more likely  
> > to
> > use the highway tag to determine and render such objects correctly.
> 
> Mapnik renders highway=unsurfaced. That seems like a pretty good sign  
> it's a common tag, even if the OSM wiki doesn't include it.
> 
> I think this makes a lot of sense, as I've never seen an arterial that  
> was gravel.

I have.  County roads are very rarely paved outside urban counties in
Oregon.  There was also an approximately 8 month span when the
Beaverton-Tigard Freeway was a four-lane, center-divided gravel road
thanks to a failed experiment at resurfacing the freeway with recycled
shoes instead of asphault (thank Nike for that idea...).  It's safe to
assume this is a common problem in countries that can't afford to
maintain infrastructure (such as most of Africa, much of southern Asia,
and demonstrably the US what with the above mentioned gravel freeway
incident and collapsing freeway bridges every couple years as long as
I've been alive).

>   I view the surface= tag as being more descriptive of the  
> highway= value.  For example, if the road is unsurfaced (which to me  
> means gravel), I might want to know how big the gravel is or whether  
> it is more dirt than gravel.
> 
> I view highway=track to mean dirt roads with no gravel (e.g. 4wd roads).

That would be a track grade 1, 2 or 3.  Smooth gravel surface that
you're capable of taking any car including something low to the ground
like a PT Cruiser at trunk speeds on would be a 5.  Track by itself just
means "it's not paved."



signature.asc
Description: This is a digitally signed message part
___
Talk-us mailing list
Talk-us@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk-us