Re: [OSM-talk] Editing road geometry Australia

[Jem]

Spot on. Although the routing engine data could impose a turn restriction
here based upon geometry as part of their data pipeline.

I wonder if it is legal to turn there and, if not, does that form part of
the ground truth IRT OSM, regardless of whether there is a sign present.


On Fri, 11 Jan 2019 at 22:53, Marc Gemis  wrote:

> If you miss the on-ramp and are waiting for the traffic signals, a
> router can recalculate the route in the meantime and still try to let
> you turn left at the traffic signals.
>
> m.
>
> On Fri, Jan 11, 2019 at 1:47 PM Maarten Deen  wrote:
> >
> > I agree that Markus' solution is more elegant (and I was more looking to
> > the offramp itself). I would normally also map it like that but I also
> > don't go out of my way to correct situations like that.
> > The way it is mapped now is more organic, more as you would actually
> > drive. As such I don't see it as wrong.
> >
> > I would not add a turn restriction. For routers it is useless because
> > you never get that route anyway.
> >
> > Regards,
> > Maarten
> >
> > On 2019-01-11 13:23, Jem wrote:
> > >> I'd map that place like that:
> > >
> https://wiki.openstreetmap.org/wiki/File:ID_Screen_Shot_from_-32.0914374,_116.0129206.png
> > >
> > > I agree. And a supplementary question... would you also add a
> > > no-left-turn restriction from https://osm.org/way/581948344 at
> > > https://osm.org/node/5680879176? I would, and have done in the past.
> > > But to be honest, I'm not sure if a turn like that (having already
> > > passed the slip lane designated for the turn) is legal or not.
> > >
> > > On Fri, 11 Jan 2019 at 20:47, Markus 
> > > wrote:
> > >
> > >> On Fri, 11 Jan 2019 at 07:40, Maarten Deen  wrote:
> > >>>
> > >>> On 2019-01-11 07:16, Petra Rajka - (p) wrote:
> > >>>
> > >>>>
> > >>>> See below two cases where we would simplify the geometry:
> > >>>>
> > >>>>   * -32.0914374, 116.0129206
> > >>>
> > >>> Is seen no big problem in how the roads are layed out there.
> > >> Coming from
> > >>> the motorway there is a clear divider where the offramp connects
> > >> to the
> > >>> Albany Highway.
> > >>
> > >> <https://www.openstreetmap.org/way/596272469> and
> > >> <https://www.openstreetmap.org/way/596272466> form a
> > >> double-rectangle,
> > >> but there isn't such a divider. I'd map that place like that:
> > >>
> > >>
> > >
> https://wiki.openstreetmap.org/wiki/File:ID_Screen_Shot_from_-32.0914374,_116.0129206.png
> > >>
> > >>> I have more problems with the tags of the on- and offramp. They
> > >> are
> > >>> mapped as motorway when they should be mapped as motorway_link.
> > >> The two
> > >>> bridges in the on- and offramp are mapped as motorway_link.
> > >>
> > >> +1. I'd also delete the descriptions like Tonkin Highway Southbound
> > >> Ramp off to Albany Highway in the name tag unless the ramps are
> > >> signed
> > >> like that on site.
> > >>
> > >>>>   * -35.3409195, 149.1616891
> > >>>
> > >>> Ways 77001149 and 77000891 should IMHO not be mapped like that but
> > >>> mapped with turn:lanes.
> > >>
> > >> +1
> > >>
> > >> Regards
> > >>
> > >> Markus
> > >>
> > >> ___
> > >> talk mailing list
> > >> talk@openstreetmap.org
> > >> https://lists.openstreetmap.org/listinfo/talk
> > > ___
> > > talk mailing list
> > > talk@openstreetmap.org
> > > https://lists.openstreetmap.org/listinfo/talk
> >
> > ___
> > talk mailing list
> > talk@openstreetmap.org
> > https://lists.openstreetmap.org/listinfo/talk
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Editing road geometry Australia

[Jem]

> I'd map that place like that:
https://wiki.openstreetmap.org/wiki/File:ID_Screen_Shot_from_-32.0914374,_116.0129206.png

I agree. And a supplementary question... would you also add a
no-left-turn restriction from https://osm.org/way/581948344 at
https://osm.org/node/5680879176? I would, and have done in the past. But to
be honest, I'm not sure if a turn like that (having already passed the slip
lane designated for the turn) is legal or not.

On Fri, 11 Jan 2019 at 20:47, Markus  wrote:

> On Fri, 11 Jan 2019 at 07:40, Maarten Deen  wrote:
> >
> > On 2019-01-11 07:16, Petra Rajka - (p) wrote:
> >
> > >
> > > See below two cases where we would simplify the geometry:
> > >
> > >   * -32.0914374, 116.0129206
> >
> > Is seen no big problem in how the roads are layed out there. Coming from
> > the motorway there is a clear divider where the offramp connects to the
> > Albany Highway.
>
>  and
>  form a double-rectangle,
> but there isn't such a divider. I'd map that place like that:
>
>
> https://wiki.openstreetmap.org/wiki/File:ID_Screen_Shot_from_-32.0914374,_116.0129206.png
>
> > I have more problems with the tags of the on- and offramp. They are
> > mapped as motorway when they should be mapped as motorway_link. The two
> > bridges in the on- and offramp are mapped as motorway_link.
>
> +1. I'd also delete the descriptions like Tonkin Highway Southbound
> Ramp off to Albany Highway in the name tag unless the ramps are signed
> like that on site.
>
> > >   * -35.3409195, 149.1616891
> >
> > Ways 77001149 and 77000891 should IMHO not be mapped like that but
> > mapped with turn:lanes.
>
> +1
>
> Regards
>
> Markus
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Multiple errors in the same location

[Jem]

> It is a CanVec import from 4 years ago

Is there subtext to this? I saw the weird natural=wood CanVec features
yesterday (polys cut up into quadtrees) and wondered about its validity. Is
the CanVec import notable for being problematic?

On Wed, 21 Nov 2018 at 19:16, sandor  wrote:

> Mateusz, this was really a quick and simple answer, probably made on
> reading the title only.
>
> The issue is much more complicated than you can imagine. You could really
> help me (and the original mapper) if you describe your suggestion how to
> resolve the lake and the hole mismatch in the case from the link. More
> precisely, assume the other mentioned problems are resolved but the (newer)
> lake – hole in forest fitting problem. So, how to move/transform these two
> objects to fit together? Further, how to do the same for really large
> number of cases where a manual procedure by me or “wait until it is done by
> someone else” is, for many reasons, unrealistic. Thanks.
>
>
>
> Sent from Mail  for
> Windows 10
>
>
>
> *From: *Mateusz Konieczny 
> *Sent: *søndag 18. november 2018 21:12
> *Cc: *talk@openstreetmap.org
> *Subject: *Re: [OSM-talk] Multiple errors in the same location
>
>
>
> 16. Nov 2018 17:06 by sandor...@gmail.com:
>
> When multiple errors appear in the same location the question is what to
> do?
>
>
>
> The same as with a single error - fix the problem (how it should be done
> depends on situation) or
>
>
>
> wait until it is done by someone else.
>
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Fwd: Short ways added to substitute barriers

[Jem]

Your assumptions are spot-on. Thanks for the advice.

On Mon, 29 Oct 2018 at 20:12, Mateusz Konieczny 
wrote:

> 29. Oct 2018 04:08 by jem.maw...@gmail.com:
>
>
> Re: https://www.openstreetmap.org/way/634085262 and several more like it
> in the area.
>
> It seems that new, short ways have been introduced to replicate the
> purpose of the existing barrier nodes. i.e. to prevent routing for
> vehicular traffic. I believe it is incorrect and just adds complexity.
>
> I plan to contact the user to discuss, but want to make sure I'm right.
> Can any experienced members please advise?
>
>
> I am assuming that there is a gate here and there is no short segment
> where
>
> motor vehicles are forbidden - though weirder thing happened and maybe
> there is a sign
>
> meters before gate from each side "motor vehicles forbidden".
>
>
> I would consider it as a tagging for renderer, and it would be preferable
> to avoid it (tagging
>
> access on gate should be sufficient). On the other hand it is one of the
> least harmful ones
>
> so I would it phrase it "it is not necessary to do that" rather "it is
> harmful, stop immediately,
>
> I reverted your edits".
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

[OSM-talk] Fwd: Short ways added to substitute barriers

[Jem]

Re: https://www.openstreetmap.org/way/634085262 and several more like it in
the area.

It seems that new, short ways have been introduced to replicate the purpose
of the existing barrier nodes. i.e. to prevent routing for vehicular
traffic. I believe it is incorrect and just adds complexity.

I plan to contact the user to discuss, but want to make sure I'm right. Can
any experienced members please advise?
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Is it time to redevelop JOSM?

[Jem]

Putting aside the string of 0-day exploits on applet plugins (which
absolutely nobody uses) that we saw a few years ago, I'm finding it
difficult to get info that backs up the claim that Java is much less secure
than .NET. What I have found indicates they are on par.  e.g.
https://www.veracode.com/sites/default/files/Resources/Reports/state-of-software-security-focus-on-application-development.pdf

It feels a bit like the old "java is slow" argument, which hasn't been true
for over a decade.

On Mon, 8 Oct 2018 at 07:40, James  wrote:

> Android(google) has been sued over implementing their own JVM, so they do
> not depend on Oracle(oracle lost that fight too)
>
> On Sun., Oct. 7, 2018, 5:38 p.m. Richard,  wrote:
>
>> On Sun, Oct 07, 2018 at 05:11:20PM -0400, James wrote:
>> > people will switch to openjdk. problem solved.
>>
>> my hope too. Although Oracle's touch of death has left an impressive
>> trace
>> in the IT departments Java is much more than that single implementation
>> by
>> Sun/Oracle. Think Android, gcj, c-lang and a few dozens third party
>> compilers
>> and implementations, also many new languages depending on Java.
>>
>> > > iD runs using an Internet connection and a browser.  I think there is
>> a
>> > > need for something like JOSM which can work completely off line if
>> need be.
>>
>> https://wiki.openstreetmap.org/wiki/Merkaartor ?
>> Vespucci?
>>
>> Richard
>>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Proximity

[Jem]

> One method might be to plot a path from each building to the health
centre by walking, car and bicycle then see how many can reach it within x
minutes.  That is a lot of routing calculation to do but it can be done
overnight or even over a couple of days.

Consider using a reverse walk from the destination and aggregating stats on
the travel times to origins. That will speed things up nicely.

On Sat, 29 Sep 2018, 10:01 AM john whelan  wrote:

> https://www.mapnificent.net/ does a close enough job for public transit
> certainly locally and demonstrates it can be done.
>
> I thank Fredrick for his comments as well.  If a more refined solution is
> required then there is enough information given to make a start coding.
>
> Thanks John
>
> On Fri, 28 Sep 2018, 6:51 pm Imre Samu,  wrote:
>
>> > If you live within a 5 minute walk of a bus stop on a route that goes
>> past the hospital then its easy to get to but how do you find these
>> locations using OpenStreetMap data?
>> >Many cities have had their bus stops imported.  If you are in one of
>> these what else is needed to work it out?
>>
>> A lot of cities has a GTFS feed - with the latest public transport info
>> it can be combined with  OSM hospital info
>> see: https://www.mapnificent.net/  [ *"Shows you areas you can reach
>> with public transport in a given time"* ]
>>
>> Best,
>>  Imre
>>
>>
>>
>> john whelan  ezt írta (időpont: 2018. szept. 28.,
>> P, 23:24):
>>
>>> I drifted down to a government conference on open data and software
>>> today and whilst there a question came up concerning proximity to a
>>> hospital.
>>>
>>> Just a planner wondering where to put it to maximise the ease of access
>>> for as many people as possible.
>>>
>>> You can route plan for walking and driving a car but what can you do for
>>> public transport.
>>>
>>> Essentially it is how many buildings are within 1 km for pedestrians, 3
>>> km for cyclists, 7 km for a car.  I've chosen arbitrary numbers but public
>>> transit is quite different.  If you live within a 5 minute walk of a bus
>>> stop on a route that goes past the hospital then its easy to get to but how
>>> do you find these locations using OpenStreetMap data?
>>>
>>> Many cities have had their bus stops imported.  If you are in one of
>>> these what else is needed to work it out?
>>>
>>> The information is worth money.  The right location makes a business or
>>> amenity more valuable.
>>>
>>> What do we have?
>>>
>>> Is this the right forum to raise the problem?
>>>
>>> Thanks John
>>> ___
>>> talk mailing list
>>> talk@openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk
>>>
>> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Waterway rel with mix of line & poly

[Jem]

Thank you both. That's very helpful.

On Thu, 20 Sep 2018 at 22:25, Dave F  wrote:

> Hi
>
> Short answer: Yes
>
> There's a few problems here:
>
> Relations should not be used to collect thing together.
> There shouldn't be tags on the ways which conflict with those in the
> relations
> MP relations require a 'type' tags and 'inners' & 'outers' roles
> In this case the Southern section shouldn't be a polygon
> MP relations should be restricted to the areas which have inners:
> https://www.openstreetmap.org/relation/2571440#map=19/51.15275/-2.05045
> The Islet, in this case, isn't included in the relation.
> There are no defined 'outers'
> There should be a complimentary tag to natural=water, such as
> water=stream/river etc.
>
> Cheers
> DaveF
>
>
>
> On 19/09/2018 01:35, Jem wrote:
>
> Is there any problem with defining a water feature that is a mix of
> polygons & lines? e.g. https://www.openstreetmap.org/relation/6447531
>
> Should it be fixed, or is it ok?
>
>
> ___
> talk mailing 
> listtalk@openstreetmap.orghttps://lists.openstreetmap.org/listinfo/talk
>
>
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

[OSM-talk] Waterway rel with mix of line & poly

[Jem]

Is there any problem with defining a water feature that is a mix of
polygons & lines? e.g. https://www.openstreetmap.org/relation/6447531

Should it be fixed, or is it ok?
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk