Feature Proposal - RFC - More Consistency in Railway Tagging

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
33 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins

Hi all,

In an earlier thread I lamented that railways are tagged inconsistently
throughout the world, and in some cases even within a certain area. I
attribute this to a lack of definition in the railway tagging scheme.

With that in mind, I've written up a proposal that attempts to:

- Refine the basic railway=* tagging to have a more specific definition,
taking inspiration from the tagging conventions around highway=* .

- Adapt the "lanes" tagging scheme from highways to allow descriptions
of individual tracks of a railway where railway mappers find that detail
to be interesting.

- Tackle in particular the (previously very under-defined) mechanism by
which we can tag tram lines that run inside highways, which is close to
my heart as a mapper in a city that has lots of them.

Here's the proposal:
 
http://wiki.openstreetmap.org/wiki/Proposed_features/Railway_Schematic_Mapping

I apologize for the length; I was trying to be thorough and avoid
ambiguity. All constructive feedback is welcomed!

Thanks,
Martin



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Kytömaa Lauri
Martin Atkins wrote:

>Refine the basic railway=* tagging to have a more specific definition,
>taking inspiration from the tagging conventions around highway=* .

IMO this is flawed in two ways:
- on empty highways, one can drive in circles on the whole road surface (not that one may or should, but they can). For anything that moves on tracks, the switches are the only place one can change course. This makes the individual tracks and their connectivity "the network", that is the roads they happen to run on are not the network itself. Drawing a single way where there are multiple tracks is not wrong, but it's just an intermediary solution before somebody has the time to add the switches and separate the tracks.

and more importantly
- there are an abundance of places where the tracks switch from the simplest case "in the road" to short (or long) bits where a separate way is required ; that is, for example where the tracks diverge from the lanes and a stop platform is between the tracks and the lanes for other motorized traffic. Were the adjoining simple case ways just one tag on the road, you would introduce arbitrary(!) turns and kinks in the course of the tracks. Other variations exist, too. There's an example of such kink in your proposal's example photos, the one where the two tracks suddendly come together when they go into a tunnel.

It's easier to go back (in code, that is) from more detailed ways to a generalized display, than it is to reconstruct the actual possible routes and the placement of the tracks if they are just tags on ways that do not follow the course of the tracks meticulously.

Tracks shown "outside" the highway where they are not, is just a rendering issue in one map, at extreme zoom levels.

There's already tram:lanes=yes| etc. as an additional attribute tag on the highway way to say "the lanes of this highway contain tram tracks, which may well be drawn separately".

--
Alv

_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Rovastar
I too think that tracks should just be a single lane. From the reasons stated above put more by Kytömaa Lauri more eloquently than I will. If a road has a physical split it is 2 separate one way streets. Tracks are physically separate so they should be mapped as such.

Roads and the crossings of multiple tracks are a little tricky but maybe this can be done as a relation rather than a node. I am no expert on tagging railway crossings.

As San Fran doesn't look like it has many railways I suggest you look at locations around the world maybe UK that has a detailed rail infrastructure so you get a better understanding about how it is done there.

Most of your issues seem to be about the rendering of the map rather than the map data.
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
In reply to this post by Kytömaa Lauri
On 04/13/2013 04:21 AM, Kytömaa Lauri wrote:
> Martin Atkins wrote:
>
>> Refine the basic railway=* tagging to have a more specific definition,
>> taking inspiration from the tagging conventions around highway=* .
>
> IMO this is flawed in two ways:
> - on empty highways, one can drive in circles on the whole road surface (not that one may or should, but they can). For anything that moves on tracks, the switches are the only place one can change course. This makes the individual tracks and their connectivity "the network", that is the roads they happen to run on are not the network itself. Drawing a single way where there are multiple tracks is not wrong, but it's just an intermediary solution before somebody has the time to add the switches and separate the tracks.

While this is true from the perspective of the railway operator, I would
argue that passengers of the railway perceive it a single "road" where
they travel in terms of the route relations placed on top of it, not in
terms of the physical track layout. When I ride on a railway, I don't
care about where the switches are, I just care about whether the train
will stop at the two stations I'm interested in (in the correct order!)
and whether the route it's going to take is a sensible one in terms of
distance.

I agree that the next level of detail can be mapped too, but I don't
agree that level of detail is interesting for the common cases of
routing that schematic mapping is aimed at, for members of the public.
Therefore it should be handled via a separate set of tags, just as
highway tagging has a progression of tagging detail from the basic ways,
to the individual lanes, to the actual area occupied by the roadways,
with each "layer" complementing the previous rather than conflicting
with it.

> and more importantly
> - there are an abundance of places where the tracks switch from the simplest case "in the road" to short (or long) bits where a separate way is required ; that is, for example where the tracks diverge from the lanes and a stop platform is between the tracks and the lanes for other motorized traffic.  Were the adjoining simple case ways just one tag on the road, you
would introduce arbitrary(!) turns and kinks in the course of the tracks.

Someone (you?) noted this on the talk page too, and I updated that
example to talk about the platform, but I'm not sure this example is
specifically about railways... even if the railway weren't there and the
platform were for boarding a bus (a situation which also occurs in San
Francisco) it would still be necessary to split the highway if one
wished to show the platform location; if we were to just draw the
platform next to the separated tramway without splitting the highway, it
would be unclear that the platform is within the roadway at all, just as
it would be unclear whether the tramway is in the roadway.

 > Other variations exist, too. There's an example of such kink in your
proposal's example photos, the one where the two tracks suddenly come
together when they go into a tunnel.

Assuming you're talking about the rendering example I showed [1], I
considered that case an example of two mappers with different opinions
about how railways should be mapped, rather than a necessary kink in the
railway; I was using it to illustrate the idea that the lack of detailed
guidance on how to tag railways has caused differing approaches in
different places.

> Tracks shown "outside" the highway where they are not, is just a rendering issue in one map, at extreme zoom levels.

Rendering is just one symptom of the inconsistency between railways and
highways. Another symptom is that e.g. level crossings get tagged twice
-- once for each track -- even though they operate as one autonomous
unit that only risks stopping the traffic once. This can have an impact
on automobile routing.

> There's already tram:lanes=yes| etc. as an additional attribute tag on the highway way to say "the lanes of this highway contain tram tracks, which may well be drawn separately".
>

I was only able to find what you described on a proposal page on the
wiki[1], but in the following form that seems compatible with my proposal:

railway:lanes=|tram|tram
direction:lanes:forward=backward|backward|forward
direction:tram:lanes:forward=|both_ways|both_ways|


[1] http://wiki.openstreetmap.org/wiki/File:Doublewaytram.png
[2]
http://wiki.openstreetmap.org/wiki/Proposed_features/lanes_General_Extension/ProposalPreVoting



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
In reply to this post by Rovastar
On 04/13/2013 04:54 AM, Rovastar wrote:
>
> As San Fran doesn't look like it has many railways I suggest you look at
> locations around the world maybe UK that has a detailed rail infrastructure
> so you get a better understanding about how it is done there.
>

Yes, I have looked at examples from elsewhere too; several of the
examples in my proposal are actually from the UK.

The only consistent thing I see with railway tagging, regardless of
where I look, is inconsistency.

To pick some examples from places I've lived or visited and am familiar
with:

- The tramway along Damrak in Amsterdam is a two-way tramway in the
middle of the street, just like my San Francisco examples. The tramway
is a separate way from the highway here. However, elsewhere in
Amsterdam, Ferdinand Bolstraat is a single way with both
highway=unclassified and railway=tram .

- In Greater London, the London Underground Metropolitan Line crosses
the LU Central Line and the Chiltern Main Line. All three of these
railways are represented as one-way-per-track, which causes the data to
suggest that the metropolitan line crosses under eight separate bridges,
when in reality it's just two wide ones. Arguably this is a problem with
the tagging scheme for bridges rather than for railways, but still.

- Continuing down the same line (pun intended!), at Finchley Road
(heading towards central London) the Metropolitan Line and the Jubilee
Line part ways in two separate tunnels, each with two tracks. The
Jubilee Line is tagged as a single way representing the entire tunnel
and both tracks, while the Metropolitan Line is tagged as two ways. I'd
argue that it's sufficient to tag both as a single way, but if either
one were to be separated it seems like the Jubilee Line is a better
candidate, since it's a pair of narrow tunnels created with a tunneling
machine, while the Metropolitan Line is a cut-and-cover pair of tracks
running side-by-side in a wide tunnel.

- Moving on to England's main-line railways, the Great Eastern Main Line
is tagged as a pair of ways from London to Whitham, but then it
inexplicably becomes a single way at least as far as Manningtree (I
didn't look any further, because I've never traveled past this point and
so I have no idea what the railway looks like on the ground.)

So all that is to say that I think we're far from having a standard for
tagging railways in Europe, too.

First and foremost I believe that more standardization would be
beneficial to make the data more useful; currently the wiki says very
little about how railways are to be tagged outside of talk pages and
off-hand comments, and that leads to inconsistencies.

Secondarily I think OSM mapping should always *begin* with a simple
ground transportation network (of highways, railways and waterways) and
then build upon that with additional (optional) tagging schemes to add
the level of detail that desired for more detailed rendering at close
zoom levels and for less-common usecases like modeling the detailed
operational details of a railway, both without conflicting with the
basic use-case of a geo-spatial route graph.

My proposal achieves the simple network and *begins* the detail, leaving
the door open for an interested party to continue it to whatever degree
of detail the community finds useful, hopefully in a way that is
compatible with approaches to also map highways and waterways in detail.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Paul Johnson-3
In reply to this post by Martin Atkins

On Sat, Apr 13, 2013 at 1:11 AM, Martin Atkins <[hidden email]> wrote:
- Adapt the "lanes" tagging scheme from highways to allow descriptions of individual tracks of a railway where railway mappers find that detail to be interesting.

Dead on arrival, as far as I'm concerned, for reasons previously mentioned the last time this came up:  Railways and highways don't work the same way, you can't just jump between tracks.

_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
On 04/13/2013 09:40 AM, Paul Johnson wrote:

>
> On Sat, Apr 13, 2013 at 1:11 AM, Martin Atkins <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     - Adapt the "lanes" tagging scheme from highways to allow
>     descriptions of individual tracks of a railway where railway mappers
>     find that detail to be interesting.
>
>
> Dead on arrival, as far as I'm concerned, for reasons previously
> mentioned the last time this came up:  Railways and highways don't work
> the same way, you can't just jump between tracks.
>

I'm not sure why that matters. The way represents several parallel
tracks without making any statement about whether one can cross them.
Anyone using the data should assume implicitly that "changing lanes"
doesn't make sense on a railway, just as real-world automobile routing
algorithms don't tell cars to drive laterally across the roadway just
because that entire surface is available for vehicle use. Mapping the
transport network is about describing useful travel paths, not about
describing all possible ways a vehicle can physically move.

With that said, I'd love a link to the previous discussion if you have
it; all I was able to discern from the wiki was that we're really
inconsistent.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Rovastar
In reply to this post by Martin Atkins
Martin,

The example you gave for tunnels and bridges are the same for roads as well. If you have a bridge or tunnel with 2 roads (one for each one-way) and a train line(s) and footpath each will be a tagged with a separate bridge. So in that regard rail is actually are consistent with the road network.

Also you say you want it better for simple mapping and other can do more detailed mapping if they want to. I see no part of your proposal to add additional tracks like is now yet you imply in the posts here that it is. If you do think this then it doesn't not help the crossing example you gave as we will have the same problem again.

Do you propose that we change *all* the currently mapped multi track rails to conform to your new standard?
e.g. here there are hundreds of tracks/railways which IMHO accurately reflects what is on the ground.
http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M 

However what I do agree with you is that the rail guidelines should be more detailed but I would go the other way with saying that all tracks should be mapped not less for complete mapping. That is a common way of doing things and going forward especially as we get more detailed mapping (it's slowly coming to the US ;))

And I also agree that crossing multiple tracks are probably not the best representation but a proposal for having them as relations would be best where multiple lines are.
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Tordanik
In reply to this post by Martin Atkins
On 13.04.2013 08:11, Martin Atkins wrote:
> http://wiki.openstreetmap.org/wiki/Proposed_features/Railway_Schematic_Mapping

You are touching a tricky issue with a proposal that is honest about its
limitations, so I welcome the constructive input. But unfortunately, I
doubt it is the right way to go.

First, lets compare this to highway: There we map separate highway ways
when there is a physical separation, e.g. with dual-carriageway
motorways. This is not the same as lanes, where you may often change
lanes at any time e.g. for overtaking, and could often move to any lane
even if you are not allowed to. Between railway tracks, there is
*always* some physical separation - you cannot just move to a parallel
track. So it is not true that your proposal would be required to
maintain consistency with highways.

Applying a "lanes" tagging scheme to railway would just highlight its
limitations: Even with highways, we don't have any established solution
for transitions between way segments with different lane layouts yet. As
you admit in your proposal, this means that we would, for example, lose
our current ability to to easily and intuitively map railway switches.

There would many similar unsolved issues regarding accuracy, many of
which you actually point out yourself, too (such as the 90° turns at
intersections). But I think not all mappers would agree with your
judgement that this very generalized mapping is good enough. Too much
important detail is "left as an exercise for someone else who is
interested".

Therefore, it is not surprising that mapping of individual tracks tends
to dominate in well-mapped areas with decent aerial imagery. At least in
the areas I'm familiar with, drawing parallel railways as a single way
is mostly limited to two cases:
* early mapping (sometimes even GPS based), to be improved later
* trams, as the unsolved rendering issues are particularly obvious there

To sum this up: I'm aware that we have unsolved problems with our
railway mapping scheme, and that it is hard to serve the needs of many
different use cases at once. But your proposal feels like giving up and
focusing exclusively on a generalized, road-centric map, and we should
strive for more than that.

Tobias

_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
In reply to this post by Rovastar
On 04/13/2013 10:18 AM, Rovastar wrote:
> Martin,
>
> The example you gave for tunnels and bridges are the same for roads as well.
> If you have a bridge or tunnel with 2 roads (one for each one-way) and a
> train line(s) and footpath each will be a tagged with a separate bridge. So
> in that regard rail is actually are consistent with the road network.

Point taken: it is a general problem with bridge tagging, not one with
railway tagging. I found some discussions on the wiki about modelling
bridges as areas that would address this, but I don't really have any
interest in mapping bridges in particular, so example retracted for the
purpose of this discussion.

> Also you say you want it better for simple mapping and other can do more
> detailed mapping if they want to. I see no part of your proposal to add
> additional tracks like is now yet you imply in the posts here that it is. If
> you do think this then it doesn't not help the crossing example you gave as
> we will have the same problem again.

I have not described a way to describe the actual routes of tracks; I
lack the expertise (or interest, frankly) to describe that next level of
detail, I just propose that we separate that next level of detail from
this simpler level of detail, rather than using the same tags for both.

I've seen the area:highway proposal for mapping the detailed shape of
streets, sidewalks and footways. My assumption was that this proposal
could be extrapolated to include a similar model for railways, modeling
the precise shapes of the trackbed the rails run along, the positions of
the individual tracks within that trackbed, etc.

If I were trying to define such a thing my first thought would be to
define a new way tag to mean "the exact path of a track" and use
separate ways from the simple route network. e.g. railway:track=rail .
I've not spent nearly as much time pulling that idea apart as I have my
simple route-oriented proposal, so I'm sure someone who knows more about
railways than I do could find examples where that doesn't apply, but
it's a strawman to start with.

I could also compromise on making the schematic network be the thing
that gets new tags, but I think it's tough to say whether it's better to
suggest re-tagging detail work in dense areas where there are evidently
lots of avid mappers at work (and the re-tagging could thus happen
relatively quickly), or to suggest re-tagging the basic network in areas
where there is less detail and there are fewer (or no) active mappers.

> Do you propose that we change *all* the currently mapped multi track rails
> to conform to your new standard?
> e.g. here there are hundreds of tracks/railways which IMHO accurately
> reflects what is on the ground.
> http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M

Under my proposal it wouldn't do any harm to leave existing detailed
tagging in place where the railway doesn't cross the highway, since a
railway way represents "one or more tracks", and that holds for the
example you showed. Optionally one could add tracks=1 to the existing
ways to make it very clear.

The converse is not true: if you define that a railway way represents
exactly one track, then there's lots of work to do to turn miles of
existing one-way-per-railway tagging into one-way-per-track.

When it comes to tramways and level crossings, whichever approach we
take there are many counter-examples to be corrected, of course.

I would note that if we *did* adopt a separate tagging scheme for
detailed mapping of tracks then at least one could simply re-tag the
existing railway=rail as (e.g.) railway:track=rail and not destroy the
existing detailed geometry. Of course, someone would have to draw in the
basic route network too; I bet the data for that is somewhere buried in
the OSM historical record, since a schematic view of the UK railway
network was imported into OSM as a starting point many years ago and is
still the basis of simple mapping in many rural areas, but admittedly I
have no idea how or whether it could easily be recovered for situations
like your Battersea example.

> However what I do agree with you is that the rail guidelines should be more
> detailed but I would go the other way with saying that all tracks should be
> mapped not less for complete mapping. That is a common way of doing things
> and going forward especially as we get more detailed mapping (it's slowly
> coming to the US ;))
>

I'd be fine with that as long as the result includes details about how
to connect the road and rail networks in a clear, unambiguous way at all
levels of detail. It is the lack of definition around these interactions
that causes the most difficulty, I think.

London isn't a great example of the problem since it has many, many
railways but very few situations where railways connect with highways at
grade.

Croydon Tramlink is one counter-example, and I'd concede that someone
has done great work in accurately mapping the path of it that I wouldn't
want to destroy, but it is sadly completely disconnected from the
highway network, and that is what I'd like to address.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
In reply to this post by Tordanik

On 04/13/2013 10:36 AM, Tobias Knerr wrote:
[snip the details]
>
> To sum this up: I'm aware that we have unsolved problems with our
> railway mapping scheme, and that it is hard to serve the needs of many
> different use cases at once. But your proposal feels like giving up and
> focusing exclusively on a generalized, road-centric map, and we should
> strive for more than that.
>

Thank you for the detailed and thoughtful feedback. One of your points
was that I deferred detail mapping to others; would you be more open to
this proposal if I were to expand it to include a mechanism for
separately mapping the details?

I'd hoped that could be addressed as a separate proposal, but it seems
like the loss of detail is the main hang-up people have with it as it
stands, so I'd be willing to spend a little time devising a strawman for
also representing the details of track and switch/point locations and
the perimeter of trackbeds if it would move the discussion forward.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - [UPDATED] More Consistency in Railway Tagging

Martin Atkins
In reply to this post by Martin Atkins

Hello again,

Based on the great feedback I got about the proposal I've updated in an
attempt to address the following concerns:

- Representing multiple tracks with a single way doesn't make sense
because trains can't "change lanes": I scrapped that part of the
proposal in favor of a new separate-way-per-track tagging scheme
described in the next point.

- Deferring the detail mapping to a separate proposal isn't good enough
because there is already detailed mapping in the wild that needs to be
addressed: I proposed a compromise that allows existing detail tagging
to be adapted via a single tag change per way and no geometry at all,
separating the concepts of "railway" (rail right of way) from
"railway:track" (an individual track).

- Also filled out my initial detail tagging proposal with a scheme for
describing the two-dimensional area occupied by the railway, based on
the area:highway proposal. As well as providing a nice 2D outline of the
railway right-of-way, it also provides a boundary to software that can
be used to answer the question "should I use the railway or the
railway:tracks?"... if an area:railway contains railway:track ways, they
should be used in favor of any railway way by software that supports both.

I believe this proposal provides a way for the two levels of detail to
co-exist in the OSM database and each be used in situations where they
are most appropriate.

The updated proposal is in the same place as the old one:
http://wiki.openstreetmap.org/wiki/Proposed_features/Railway_Schematic_Mapping

I'd appreciate it if those who have participated in the discussion so
far would take the time to review the updated proposal and let me know
what you think.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

David Fisher
In reply to this post by Martin Atkins
> Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want
> to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.

As a Croydon mapper myself who has made changes to the tramlink network... firstly thanks for noticing that it looks OK!  It's by no means all my own work, but I like to think I've done my bit :)   There's actually only one stretch which is relevant to this debate: the Addiscombe Road section between East Croydon and Sandilands (I presume it is this section you refer to when you say "completely disconnected from the highway network"?).  All the other on-highway parts are only single-track, so are easily represented by a single way with nodes merged with the highway.

Anyway.  My two cents, for what it's worth:  I am strongly in favour of mapping highways and railways differently (one way per separated piece of tarmac for roads; one way per rail for railways).  One form of compromise, however, could be to treat specifically on-highway rail systems with the "highway" protocol.  Or, maybe for multi-rail on-highway sections, map them as separate ways (cf Addiscombe Road tramlink) and use a relation just to cover these sections?  I realise this is not ideal for cities with a large proportion of such sections, but long-term it may be a way to maintain detail whilst limiting complexity (since relations would not be needed for *every* section, just shared sections).

Cheers,

David Fisher.




On Sat, Apr 13, 2013 at 7:11 PM, Martin Atkins <[hidden email]> wrote:
On 04/13/2013 10:18 AM, Rovastar wrote:
Martin,

The example you gave for tunnels and bridges are the same for roads as well.
If you have a bridge or tunnel with 2 roads (one for each one-way) and a
train line(s) and footpath each will be a tagged with a separate bridge. So
in that regard rail is actually are consistent with the road network.

Point taken: it is a general problem with bridge tagging, not one with railway tagging. I found some discussions on the wiki about modelling bridges as areas that would address this, but I don't really have any interest in mapping bridges in particular, so example retracted for the purpose of this discussion.


Also you say you want it better for simple mapping and other can do more
detailed mapping if they want to. I see no part of your proposal to add
additional tracks like is now yet you imply in the posts here that it is. If
you do think this then it doesn't not help the crossing example you gave as
we will have the same problem again.

I have not described a way to describe the actual routes of tracks; I lack the expertise (or interest, frankly) to describe that next level of detail, I just propose that we separate that next level of detail from this simpler level of detail, rather than using the same tags for both.

I've seen the area:highway proposal for mapping the detailed shape of streets, sidewalks and footways. My assumption was that this proposal could be extrapolated to include a similar model for railways, modeling the precise shapes of the trackbed the rails run along, the positions of the individual tracks within that trackbed, etc.

If I were trying to define such a thing my first thought would be to define a new way tag to mean "the exact path of a track" and use separate ways from the simple route network. e.g. railway:track=rail . I've not spent nearly as much time pulling that idea apart as I have my simple route-oriented proposal, so I'm sure someone who knows more about railways than I do could find examples where that doesn't apply, but it's a strawman to start with.

I could also compromise on making the schematic network be the thing that gets new tags, but I think it's tough to say whether it's better to suggest re-tagging detail work in dense areas where there are evidently lots of avid mappers at work (and the re-tagging could thus happen relatively quickly), or to suggest re-tagging the basic network in areas where there is less detail and there are fewer (or no) active mappers.


Do you propose that we change *all* the currently mapped multi track rails
to conform to your new standard?
e.g. here there are hundreds of tracks/railways which IMHO accurately
reflects what is on the ground.
http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M

Under my proposal it wouldn't do any harm to leave existing detailed tagging in place where the railway doesn't cross the highway, since a railway way represents "one or more tracks", and that holds for the example you showed. Optionally one could add tracks=1 to the existing ways to make it very clear.

The converse is not true: if you define that a railway way represents exactly one track, then there's lots of work to do to turn miles of existing one-way-per-railway tagging into one-way-per-track.

When it comes to tramways and level crossings, whichever approach we take there are many counter-examples to be corrected, of course.

I would note that if we *did* adopt a separate tagging scheme for detailed mapping of tracks then at least one could simply re-tag the existing railway=rail as (e.g.) railway:track=rail and not destroy the existing detailed geometry. Of course, someone would have to draw in the basic route network too; I bet the data for that is somewhere buried in the OSM historical record, since a schematic view of the UK railway network was imported into OSM as a starting point many years ago and is still the basis of simple mapping in many rural areas, but admittedly I have no idea how or whether it could easily be recovered for situations like your Battersea example.


However what I do agree with you is that the rail guidelines should be more
detailed but I would go the other way with saying that all tracks should be
mapped not less for complete mapping. That is a common way of doing things
and going forward especially as we get more detailed mapping (it's slowly
coming to the US ;))


I'd be fine with that as long as the result includes details about how to connect the road and rail networks in a clear, unambiguous way at all levels of detail. It is the lack of definition around these interactions that causes the most difficulty, I think.

London isn't a great example of the problem since it has many, many railways but very few situations where railways connect with highways at grade.

Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.




_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging


_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

John F. Eldredge
David Fisher <[hidden email]> wrote:
> Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want
> to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.

As a Croydon mapper myself who has made changes to the tramlink network... firstly thanks for noticing that it looks OK!  It's by no means all my own work, but I like to think I've done my bit :)   There's actually only one stretch which is relevant to this debate: the Addiscombe Road section between East Croydon and Sandilands (I presume it is this section you refer to when you say "completely disconnected from the highway network"?).  All the other on-highway parts are only single-track, so are easily represented by a single way with nodes merged with the highway.

Anyway.  My two cents, for what it's worth:  I am strongly in favour of mapping highways and railways differently (one way per separated piece of tarmac for roads; one way per rail for railways).  One form of compromise, however, could be to treat specifically on-highway rail systems with the "highway" protocol.  Or, maybe for multi-rail on-highway sections, map them as separate ways (cf Addiscombe Road tramlink) and use a relation just to cover these sections?  I realise this is not ideal for cities with a large proportion of such sections, but long-term it may be a way to maintain detail whilst limiting complexity (since relations would not be needed for *every* section, just shared sections).

Cheers,

David Fisher.




On Sat, Apr 13, 2013 at 7:11 PM, Martin Atkins <[hidden email]> wrote:
On 04/13/2013 10:18 AM, Rovastar wrote:
Martin,

The example you gave for tunnels and bridges are the same for roads as well.
If you have a bridge or tunnel with 2 roads (one for each one-way) and a
train line(s) and footpath each will be a tagged with a separate bridge. So
in that regard rail is actually are consistent with the road network.

Point taken: it is a general problem with bridge tagging, not one with railway tagging. I found some discussions on the wiki about modelling bridges as areas that would address this, but I don't really have any interest in mapping bridges in particular, so example retracted for the purpose of this discussion.


Also you say you want it better for simple mapping and other can do more
detailed mapping if they want to. I see no part of your proposal to add
additional tracks like is now yet you imply in the posts here that it is. If
you do think this then it doesn't not help the crossing example you gave as
we will have the same problem again.

I have not described a way to describe the actual routes of tracks; I lack the expertise (or interest, frankly) to describe that next level of detail, I just propose that we separate that next level of detail from this simpler level of detail, rather than using the same tags for both.

I've seen the area:highway proposal for mapping the detailed shape of streets, sidewalks and footways. My assumption was that this proposal could be extrapolated to include a similar model for railways, modeling the precise shapes of the trackbed the rails run along, the positions of the individual tracks within that trackbed, etc.

If I were trying to define such a thing my first thought would be to define a new way tag to mean "the exact path of a track" and use separate ways from the simple route network. e.g. railway:track=rail . I've not spent nearly as much time pulling that idea apart as I have my simple route-oriented proposal, so I'm sure someone who knows more about railways than I do could find examples where that doesn't apply, but it's a strawman to start with.

I could also compromise on making the schematic network be the thing that gets new tags, but I think it's tough to say whether it's better to suggest re-tagging detail work in dense areas where there are evidently lots of avid mappers at work (and the re-tagging could thus happen relatively quickly), or to suggest re-tagging the basic network in areas where there is less detail and there are fewer (or no) active mappers.


Do you propose that we change *all* the currently mapped multi track rails
to conform to your new standard?
e.g. here there are hundreds of tracks/railways which IMHO accurately
reflects what is on the ground.
http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M

Under my proposal it wouldn't do any harm to leave existing detailed tagging in place where the railway doesn't cross the highway, since a railway way represents "one or more tracks", and that holds for the example you showed. Optionally one could add tracks=1 to the existing ways to make it very clear.

The converse is not true: if you define that a railway way represents exactly one track, then there's lots of work to do to turn miles of existing one-way-per-railway tagging into one-way-per-track.

When it comes to tramways and level crossings, whichever approach we take there are many counter-examples to be corrected, of course.

I would note that if we *did* adopt a separate tagging scheme for detailed mapping of tracks then at least one could simply re-tag the existing railway=rail as (e.g.) railway:track=rail and not destroy the existing detailed geometry. Of course, someone would have to draw in the basic route network too; I bet the data for that is somewhere buried in the OSM historical record, since a schematic view of the UK railway network was imported into OSM as a starting point many years ago and is still the basis of simple mapping in many rural areas, but admittedly I have no idea how or whether it could easily be recovered for situations like your Battersea example.


However what I do agree with you is that the rail guidelines should be more
detailed but I would go the other way with saying that all tracks should be
mapped not less for complete mapping. That is a common way of doing things
and going forward especially as we get more detailed mapping (it's slowly
coming to the US ;))


I'd be fine with that as long as the result includes details about how to connect the road and rail networks in a clear, unambiguous way at all levels of detail. It is the lack of definition around these interactions that causes the most difficulty, I think.

London isn't a great example of the problem since it has many, many railways but very few situations where railways connect with highways at grade.

Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.




_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging

Did you intentionally write "one way per rail", or did you intend to say "one way per track"? I had not seen anyone in the discussion, up until now, propose mapping each rail as a separate way. This would mean that a single-track right-of-way would require two ways, unless you were mapping a monorail.

--
John F. Eldredge -- [hidden email]
"Reserve your right to think, for it is better to think wrongly than not to think at all." -- Hypatia of Alexandria
_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

David Fisher
Sorry, yes, "one way per track", of course.  Was writing hurriedly.
Thanks,
David.


On Sat, Apr 13, 2013 at 9:06 PM, John F. Eldredge <[hidden email]> wrote:
David Fisher <[hidden email]> wrote:
> Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want
> to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.

As a Croydon mapper myself who has made changes to the tramlink network... firstly thanks for noticing that it looks OK!  It's by no means all my own work, but I like to think I've done my bit :)   There's actually only one stretch which is relevant to this debate: the Addiscombe Road section between East Croydon and Sandilands (I presume it is this section you refer to when you say "completely disconnected from the highway network"?).  All the other on-highway parts are only single-track, so are easily represented by a single way with nodes merged with the highway.

Anyway.  My two cents, for what it's worth:  I am strongly in favour of mapping highways and railways differently (one way per separated piece of tarmac for roads; one way per rail for railways).  One form of compromise, however, could be to treat specifically on-highway rail systems with the "highway" protocol.  Or, maybe for multi-rail on-highway sections, map them as separate ways (cf Addiscombe Road tramlink) and use a relation just to cover these sections?  I realise this is not ideal for cities with a large proportion of such sections, but long-term it may be a way to maintain detail whilst limiting complexity (since relations would not be needed for *every* section, just shared sections).

Cheers,

David Fisher.




On Sat, Apr 13, 2013 at 7:11 PM, Martin Atkins <[hidden email]> wrote:
On 04/13/2013 10:18 AM, Rovastar wrote:
Martin,

The example you gave for tunnels and bridges are the same for roads as well.
If you have a bridge or tunnel with 2 roads (one for each one-way) and a
train line(s) and footpath each will be a tagged with a separate bridge. So
in that regard rail is actually are consistent with the road network.

Point taken: it is a general problem with bridge tagging, not one with railway tagging. I found some discussions on the wiki about modelling bridges as areas that would address this, but I don't really have any interest in mapping bridges in particular, so example retracted for the purpose of this discussion.


Also you say you want it better for simple mapping and other can do more
detailed mapping if they want to. I see no part of your proposal to add
additional tracks like is now yet you imply in the posts here that it is. If
you do think this then it doesn't not help the crossing example you gave as
we will have the same problem again.

I have not described a way to describe the actual routes of tracks; I lack the expertise (or interest, frankly) to describe that next level of detail, I just propose that we separate that next level of detail from this simpler level of detail, rather than using the same tags for both.

I've seen the area:highway proposal for mapping the detailed shape of streets, sidewalks and footways. My assumption was that this proposal could be extrapolated to include a similar model for railways, modeling the precise shapes of the trackbed the rails run along, the positions of the individual tracks within that trackbed, etc.

If I were trying to define such a thing my first thought would be to define a new way tag to mean "the exact path of a track" and use separate ways from the simple route network. e.g. railway:track=rail . I've not spent nearly as much time pulling that idea apart as I have my simple route-oriented proposal, so I'm sure someone who knows more about railways than I do could find examples where that doesn't apply, but it's a strawman to start with.

I could also compromise on making the schematic network be the thing that gets new tags, but I think it's tough to say whether it's better to suggest re-tagging detail work in dense areas where there are evidently lots of avid mappers at work (and the re-tagging could thus happen relatively quickly), or to suggest re-tagging the basic network in areas where there is less detail and there are fewer (or no) active mappers.


Do you propose that we change *all* the currently mapped multi track rails
to conform to your new standard?
e.g. here there are hundreds of tracks/railways which IMHO accurately
reflects what is on the ground.
http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M

Under my proposal it wouldn't do any harm to leave existing detailed tagging in place where the railway doesn't cross the highway, since a railway way represents "one or more tracks", and that holds for the example you showed. Optionally one could add tracks=1 to the existing ways to make it very clear.

The converse is not true: if you define that a railway way represents exactly one track, then there's lots of work to do to turn miles of existing one-way-per-railway tagging into one-way-per-track.

When it comes to tramways and level crossings, whichever approach we take there are many counter-examples to be corrected, of course.

I would note that if we *did* adopt a separate tagging scheme for detailed mapping of tracks then at least one could simply re-tag the existing railway=rail as (e.g.) railway:track=rail and not destroy the existing detailed geometry. Of course, someone would have to draw in the basic route network too; I bet the data for that is somewhere buried in the OSM historical record, since a schematic view of the UK railway network was imported into OSM as a starting point many years ago and is still the basis of simple mapping in many rural areas, but admittedly I have no idea how or whether it could easily be recovered for situations like your Battersea example.


However what I do agree with you is that the rail guidelines should be more
detailed but I would go the other way with saying that all tracks should be
mapped not less for complete mapping. That is a common way of doing things
and going forward especially as we get more detailed mapping (it's slowly
coming to the US ;))


I'd be fine with that as long as the result includes details about how to connect the road and rail networks in a clear, unambiguous way at all levels of detail. It is the lack of definition around these interactions that causes the most difficulty, I think.

London isn't a great example of the problem since it has many, many railways but very few situations where railways connect with highways at grade.

Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.




_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging

Did you intentionally write "one way per rail", or did you intend to say "one way per track"? I had not seen anyone in the discussion, up until now, propose mapping each rail as a separate way. This would mean that a single-track right-of-way would require two ways, unless you were mapping a monorail.

--
John F. Eldredge -- [hidden email]
"Reserve your right to think, for it is better to think wrongly than not to think at all." -- Hypatia of Alexandria

_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Jo-2
FWIW I also prefer one way per track. When the tracks are embedded in the tarmac of a normal 2 way street, I'll draw 3 ways (2 railways and 1 highway), each with the name included. When I bother to add associatedStreet relations I'd add all of them into it.
When it's 1 track on a one way street (or half a dual carriage road), I'd be inclined to simply add a railway=tram to the highway=...
But usually the tracks happen to be in the middle between the carriage roads in that case, so 4 ways. If buses also us this middle part, this may mean an additional 5th highway=service could be added. Although I'm usually simply adding highway=service, psv=yes to both of those railways. Technically the bus could overtake another bus or a tram, but that seems like an unwise thing to do, unless it's defective. Maybe I'll start adding that 5th way...

Anyway, the tendency is towards mapping more detail, not less, so mapping tracks as separate ways makes the most sense to me, especially since aerial imagery allows us to see the actual tracks. I certainly prefer that over breaking trams by having them take 90 degrees turns.

Polyglot



2013/4/13 David Fisher <[hidden email]>
Sorry, yes, "one way per track", of course.  Was writing hurriedly.
Thanks,
David.


On Sat, Apr 13, 2013 at 9:06 PM, John F. Eldredge <[hidden email]> wrote:
David Fisher <[hidden email]> wrote:
> Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want
> to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.

As a Croydon mapper myself who has made changes to the tramlink network... firstly thanks for noticing that it looks OK!  It's by no means all my own work, but I like to think I've done my bit :)   There's actually only one stretch which is relevant to this debate: the Addiscombe Road section between East Croydon and Sandilands (I presume it is this section you refer to when you say "completely disconnected from the highway network"?).  All the other on-highway parts are only single-track, so are easily represented by a single way with nodes merged with the highway.

Anyway.  My two cents, for what it's worth:  I am strongly in favour of mapping highways and railways differently (one way per separated piece of tarmac for roads; one way per rail for railways).  One form of compromise, however, could be to treat specifically on-highway rail systems with the "highway" protocol.  Or, maybe for multi-rail on-highway sections, map them as separate ways (cf Addiscombe Road tramlink) and use a relation just to cover these sections?  I realise this is not ideal for cities with a large proportion of such sections, but long-term it may be a way to maintain detail whilst limiting complexity (since relations would not be needed for *every* section, just shared sections).

Cheers,

David Fisher.




On Sat, Apr 13, 2013 at 7:11 PM, Martin Atkins <[hidden email]> wrote:
On 04/13/2013 10:18 AM, Rovastar wrote:
Martin,

The example you gave for tunnels and bridges are the same for roads as well.
If you have a bridge or tunnel with 2 roads (one for each one-way) and a
train line(s) and footpath each will be a tagged with a separate bridge. So
in that regard rail is actually are consistent with the road network.

Point taken: it is a general problem with bridge tagging, not one with railway tagging. I found some discussions on the wiki about modelling bridges as areas that would address this, but I don't really have any interest in mapping bridges in particular, so example retracted for the purpose of this discussion.


Also you say you want it better for simple mapping and other can do more
detailed mapping if they want to. I see no part of your proposal to add
additional tracks like is now yet you imply in the posts here that it is. If
you do think this then it doesn't not help the crossing example you gave as
we will have the same problem again.

I have not described a way to describe the actual routes of tracks; I lack the expertise (or interest, frankly) to describe that next level of detail, I just propose that we separate that next level of detail from this simpler level of detail, rather than using the same tags for both.

I've seen the area:highway proposal for mapping the detailed shape of streets, sidewalks and footways. My assumption was that this proposal could be extrapolated to include a similar model for railways, modeling the precise shapes of the trackbed the rails run along, the positions of the individual tracks within that trackbed, etc.

If I were trying to define such a thing my first thought would be to define a new way tag to mean "the exact path of a track" and use separate ways from the simple route network. e.g. railway:track=rail . I've not spent nearly as much time pulling that idea apart as I have my simple route-oriented proposal, so I'm sure someone who knows more about railways than I do could find examples where that doesn't apply, but it's a strawman to start with.

I could also compromise on making the schematic network be the thing that gets new tags, but I think it's tough to say whether it's better to suggest re-tagging detail work in dense areas where there are evidently lots of avid mappers at work (and the re-tagging could thus happen relatively quickly), or to suggest re-tagging the basic network in areas where there is less detail and there are fewer (or no) active mappers.


Do you propose that we change *all* the currently mapped multi track rails
to conform to your new standard?
e.g. here there are hundreds of tracks/railways which IMHO accurately
reflects what is on the ground.
http://www.openstreetmap.org/?lat=51.47119&lon=-0.14847&zoom=15&layers=M

Under my proposal it wouldn't do any harm to leave existing detailed tagging in place where the railway doesn't cross the highway, since a railway way represents "one or more tracks", and that holds for the example you showed. Optionally one could add tracks=1 to the existing ways to make it very clear.

The converse is not true: if you define that a railway way represents exactly one track, then there's lots of work to do to turn miles of existing one-way-per-railway tagging into one-way-per-track.

When it comes to tramways and level crossings, whichever approach we take there are many counter-examples to be corrected, of course.

I would note that if we *did* adopt a separate tagging scheme for detailed mapping of tracks then at least one could simply re-tag the existing railway=rail as (e.g.) railway:track=rail and not destroy the existing detailed geometry. Of course, someone would have to draw in the basic route network too; I bet the data for that is somewhere buried in the OSM historical record, since a schematic view of the UK railway network was imported into OSM as a starting point many years ago and is still the basis of simple mapping in many rural areas, but admittedly I have no idea how or whether it could easily be recovered for situations like your Battersea example.


However what I do agree with you is that the rail guidelines should be more
detailed but I would go the other way with saying that all tracks should be
mapped not less for complete mapping. That is a common way of doing things
and going forward especially as we get more detailed mapping (it's slowly
coming to the US ;))


I'd be fine with that as long as the result includes details about how to connect the road and rail networks in a clear, unambiguous way at all levels of detail. It is the lack of definition around these interactions that causes the most difficulty, I think.

London isn't a great example of the problem since it has many, many railways but very few situations where railways connect with highways at grade.

Croydon Tramlink is one counter-example, and I'd concede that someone has done great work in accurately mapping the path of it that I wouldn't want to destroy, but it is sadly completely disconnected from the highway network, and that is what I'd like to address.




_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging

Did you intentionally write "one way per rail", or did you intend to say "one way per track"? I had not seen anyone in the discussion, up until now, propose mapping each rail as a separate way. This would mean that a single-track right-of-way would require two ways, unless you were mapping a monorail.

--
John F. Eldredge -- [hidden email]
"Reserve your right to think, for it is better to think wrongly than not to think at all." -- Hypatia of Alexandria

_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Martin Atkins
In reply to this post by David Fisher
On 04/13/2013 12:52 PM, David Fisher wrote:

>
> Anyway.  My two cents, for what it's worth:  I am strongly in favour of
> mapping highways and railways differently (one way per separated piece
> of tarmac for roads; one way per rail for railways).  One form of
> compromise, however, could be to treat specifically on-highway rail
> systems with the "highway" protocol.  Or, maybe for multi-rail
> on-highway sections, map them as separate ways (cf Addiscombe Road
> tramlink) and use a relation just to cover these sections?  I realise
> this is not ideal for cities with a large proportion of such sections,
> but long-term it may be a way to maintain detail whilst limiting
> complexity (since relations would not be needed for *every* section,
> just shared sections).
>

This is actually pretty insightful, and lead me to the following thought:

Really my proposal is all about trams. The problems with level crossings
and bridges tricked me into thinking it was a general rail problem, but
really I think the core issue is with these special characteristics of
trams as compared to full-blown trains:

- They have a small form factor like a bus, and unlike a train. This
allows them to behave more like "normal" traffic, with the exception of
not being able to change lanes. For example, they can make tight turns
and can stop quickly when traffic conditions require it. (Trolley buses
are similar in that they are tethered to special infrastructure, but
they have a little more flexibility because of the trolley poles.)

- They usually participate as full members of the roadway... that is,
they often stop at the same traffic signals, the same stop signs, get
stuck in the same traffic queues. On the other hand, when full railways
interact with roads they generally get automatic priority, stopping all
other traffic that might otherwise be using the right of way.

- They often have boarding areas that are more like bus stops than
platforms, and depending on where in the roadway the track runs there
may not be any special boarding facilities at all.

Some of this applies to a lesser extent to light rail in some areas.

So with all of that said, I'm considering reigning in my proposal to
discuss trams only; would one-way-per-tramway but one-way-per-rail-track
be an acceptable compromise? Again I'm really only personally interested
in a comparable amount of detail as trolley_wire=yes, but if it will
help to convince I can again try to create a strawman detailed mapping
proposal.

There is still the problem with level crossings to deal with, but that's
minor and can be dealt with by a separate proposal (tramways don't
usually have level crossings in the same sense as railways anyway) and
otherwise the full detailed mapping of railways is harmless.



_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Paul Johnson-3
I'd still stick to one-way-per-track for consistency, especially given many cities (London, Portland) have 100% complete systems mapped already using that method, and it's not just torturing the data.  Your assumptions of how trams interact with traffic are also pretty far from universal, with Portland and Phoenix being examples that pretty much flies in the face of what you suggested (Portland's newest trams are actually slightly longer than the shortest platforms, which are a city block long, frequently leaving the tail end of the train in the intersection, with most trains being only about 10 feet shorter; priority signaling is the norm; 480 tons doesn't stop on a dime in any sense of the word).


On Sat, Apr 13, 2013 at 11:34 PM, Martin Atkins <[hidden email]> wrote:
On 04/13/2013 12:52 PM, David Fisher wrote:

Anyway.  My two cents, for what it's worth:  I am strongly in favour of
mapping highways and railways differently (one way per separated piece
of tarmac for roads; one way per rail for railways).  One form of
compromise, however, could be to treat specifically on-highway rail
systems with the "highway" protocol.  Or, maybe for multi-rail
on-highway sections, map them as separate ways (cf Addiscombe Road
tramlink) and use a relation just to cover these sections?  I realise
this is not ideal for cities with a large proportion of such sections,
but long-term it may be a way to maintain detail whilst limiting
complexity (since relations would not be needed for *every* section,
just shared sections).


This is actually pretty insightful, and lead me to the following thought:

Really my proposal is all about trams. The problems with level crossings and bridges tricked me into thinking it was a general rail problem, but really I think the core issue is with these special characteristics of trams as compared to full-blown trains:

- They have a small form factor like a bus, and unlike a train. This allows them to behave more like "normal" traffic, with the exception of not being able to change lanes. For example, they can make tight turns and can stop quickly when traffic conditions require it. (Trolley buses are similar in that they are tethered to special infrastructure, but they have a little more flexibility because of the trolley poles.)

- They usually participate as full members of the roadway... that is, they often stop at the same traffic signals, the same stop signs, get stuck in the same traffic queues. On the other hand, when full railways interact with roads they generally get automatic priority, stopping all other traffic that might otherwise be using the right of way.

- They often have boarding areas that are more like bus stops than platforms, and depending on where in the roadway the track runs there may not be any special boarding facilities at all.

Some of this applies to a lesser extent to light rail in some areas.

So with all of that said, I'm considering reigning in my proposal to discuss trams only; would one-way-per-tramway but one-way-per-rail-track be an acceptable compromise? Again I'm really only personally interested in a comparable amount of detail as trolley_wire=yes, but if it will help to convince I can again try to create a strawman detailed mapping proposal.

There is still the problem with level crossings to deal with, but that's minor and can be dealt with by a separate proposal (tramways don't usually have level crossings in the same sense as railways anyway) and otherwise the full detailed mapping of railways is harmless.




_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging


_______________________________________________
Tagging mailing list
[hidden email]
http://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Rovastar
In reply to this post by Martin Atkins
I think Martin it is best to focus your proposal on the area that is manageable, smaller in scale and something that your know and have passion about.

So in this case I would just do a something for trams.

I think the initial problems with the proposal was it was radical and effected millions of ways as it was all railways, and that is a huge and fairly well established area.

Reading my comments back I didn't want/mean to sound aggressive I was just an understanding of what you were trying to do and the problems that it can cause.

I have never mapped trams so it is not my area of interest/expertise but common it seem to be going forward that many map 1 per track. I know that is not something you (initially) wanted but maybe something can come from it.

I am unclear on the whole proposal thing and if is need to clarify and add details to the wiki. I have added details and clarification to the wiki pages (Stadiums, golf courses (although I have never played or even like golf :/) in the past and I have never done a proposal. And I want to at some point add clarification and examples to the rail pages that we are striving to map all tracks if possible.

I think a lot can be done in relations for the routes of the trams. I notice that for your San Fran example there are no relations attached to these trams lines.
That will tell the people where they go and the routes they take. It is up to the mapping render/software to decide what is displayed. Transport map I think will show the relations and routes of the trams.
With relations they should be easier to follow and I think will led to better understanding of what can be mapped in the future.

I would look at other examples of trams around the world, find what you consider the best examples and discuss them here. Update the wiki with good examples. I find that is what is lacking in the wiki is good examples of good practice. Generally I think people will follow good examples of good practice.

Then you can be our tram tzar. :)
Reply | Threaded
Open this post in threaded view
|

Re: Feature Proposal - RFC - More Consistency in Railway Tagging

Rovastar
Apologies I looked a little more at the San Fran area link you post you do have relations for the tram (which you added as tram yesterday...) as a road combo just not for each line of the trams which is tagged as light_rail. Which I presume it is just a tram line and therefore wrong anyway.

I can see more clearly the problem you are having there now.

I think once it is all per track it will become clearer for you.
12