[Tagging] Feature Proposal - RFC - Public Transport v3

[Paul Allen]

On Mon, 9 Mar 2020 at 00:18, Joseph Eisenberg <joseph.eisenberg at gmail.com>
wrote:

> > editors would have to take similar precautions with nodes.  Not
> impossible, but it would take time to appear.  Your scheme will be more
> fragile than the existing one, at least for a while.
>
> Well yes, any change will take some work by the maintainers of editor
> applications, so we should not make changes lightly.
>

Given that iD still seems to scramble sorted routes in some circumstances,
one
should not assume that editors will correctly handle any changes we make.  I
might be being unfair to iD here, I didn't check the route was still sorted
before
I added a spur, so maybe somebody else doing something that didn't directly
affect the route using another editor scrambled it.

>
> But this should not prevent us from making a change which would make
> things easier for mappers and might better represent reality.
>

Given recent comments on the proposal's talk page, I doubt it will better
represent reality because it's the router that determines the fine details.
Maybe it will go for shortest path, or fastest path, or something else.
Maybe at some future date it will switch from shortest path to fastest
path.  Yes, there are vias, but read on.

For a long-distance inter-city route where the bus only stops at official
stops,
this might be good enough for many people.  The router may or may not
guess the exact route between two stops, but it's close enough for some
people.  Not for me, because I worry when a bus I'm on isn't following
the route I thought it would.  Maybe I'm on the wrong bus.  Maybe the
route has changed and it's not going to stop where I want to get off.  I
want to see exact routes, not close approximations.

The county I live in is the second-most sparsely populated county in
Wales and has an area of 1,795 square km (693 sq mi).  Just about
every route is hail and ride.  Even the routes in the towns (and the
one city) are hail and ride except in small sections where the road
configuration would make it unsafe to stop.  In these cases,
knowing the exact route is vital: otherwise you'll be waiting
for a bus where the router thinks it goes instead of where it
actually goes.  Yes, there are vias, but read on.

>
> Greater ease of mapping is a strong reason to consider changes to
> editor applications.
>

Is it actually easier?  For the "long-distance, few official stops, only
stops
at official stops" routes, maybe.  For any route in my county I'll have
to insert vias to make sure hail and ride buses (almost all of them) are
mapped as going where they actually go.  It means I have to spend time
thinking like a router, anticipating alternate routes between any adjacent
junctions and insert a via where a router might possibly come up with
the non-factual answer.  Sorting additional vias into place is going to
be almost as tedious as re-sorting existing relations of ways that
get scrambled.

However carefully I place the vias (and check the calculated route to see if
it's right, some change to the roads which doesn't affect the route the bus
actually takes may cause the mapped route to be wrong until one or more
new vias are inserted.  A new housing estate with two connections to
the existing road system could cause the bus to be re-routed.  Fine if
the person adding the estate takes care to check nearby bus routes.
A change to the speed limit on a way may cause the router to decide
on a different route.  Etc.

Adding ways to create a relation is mechanical and tedious but requires
little thought.  Guessing where one might have to insert vias that will cope
not only with whatever router is used to generate the carto but all future
algorithms that router might use requires a lot of thoughtful consideration
of
potential alternative routes.  Or I can avoid thought by placing a via at
every node along the route, which is more tedious than just adding ways.

In some cases this proposal might be easier, in others (the ones I deal
with) a
lot harder.  Which is fine if they're alternative ways of mapping bus
routes.  But
the proposers seem to want to make this the one and only way of doing
things.

-- 
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20200309/3ee1c404/attachment.htm>