[OSM-talk] dangerous cycling lanes (was Re: A new highwaytagging scheme - thinking about)

Peter Miller peter.miller at itoworld.com
Wed Aug 29 10:52:51 BST 2007

> Message: 4
> Date: Wed, 29 Aug 2007 09:29:24 +0100
> From: Tom Chance <tom at acrewoods.net>
> Subject: Re: [OSM-talk] dangerous cycling lanes (was Re: A new
> 	highwaytagging scheme - thinking about)
> To: Cameron Patrick <cameron at patrick.wattle.id.au>
> Cc: talk at openstreetmap.org
> Message-ID: <8cb4814066f762c81bc18d9b2bfe246b at acrewoods.net>
> Content-Type: text/plain; charset="UTF-8"
> On Wed, 29 Aug 2007 16:07:12 +0800, Cameron Patrick
> <cameron at patrick.wattle.id.au> wrote:
> > I got the impression that "lcn" was reserved for more formal cycle
> > networks?  I'd go for "cycling=poor|medium|good" (note that this is a
> > new tag, as distinct from the bike=yes or whatever the current one is
> > called).

TO be clear, lcn_ref is used to identify a signed route (which may or may
not be good to cycle on) and lcn=yes is used to show other useful local
cycle route and is open to ideas on how to extend its use.

> But that's so subjective! Several of my colleagues refuse to cycle down
> Streatham High St to work because it's a busy trunk road with trucks and
> buses, but I prefer it because it's quick and saves me going up and down
> too many hills.

Personally I think we need to support many different users and user groups,
from 8yo kids to fit testosterone fuels adult males. We also need to cater
for all the other users, walkers, buggies, electric wheelchairs etc but I
don't hear people wanting to extend the model out that far yet.

I imagine that the Perth maps are considering the less confident cyclist and
are assuming that the confident cyclist will ignore the warnings as
confident cyclists do with all guidance!

It seems reasonably to provide more information focused at young and
cautious cyclists and let the more confident ones push out from there.

> As I wrote in my other email, and as is indicated on this rather nice page
> pointed out to me, it's much better to focus on characteristics that
> physical, administrative and safety related:
> http://wiki.openstreetmap.org/index.php/Cycleway
> You can then determine what is "poor" or "good" based on your preferences.

I have no problem with tagging physical characteristics, but a router will
also needs to be able to determine whether it is poor, good or whatever and
to achieve that the routing algorithm probably needs to end up with a single
number for each link.

I do agree that the objective/subjective argument will run and run, but
personally I think the router will have more chance to get good routes with
a subjective approach. In the absence of the lcn tag the router can guess
from other attributes (maxspeed, classification etc) but it is probably
still missing loads of data to get it right. What should it do with 'width'?
Surely the road width matters much less than how wide the available
carriageway is when parked cars are also considered, together with the
amount of traffic, the type of traffic (wide lorries or small cars) together
with how often people are getting in and out of the cars together with the
steepness of the slope and what happens at the bottom of the slope.

Personally, I am marking the quieter streets as 'lcn=yes' and the busier
streets as 'lcn=poor'. If you are going for looking for quietness then the
router will prefer lcn=yes and try to avoid lcn=poor. If you are going for
speed it will not pay so much attention to the lcn tagging at all except
possibly for lcn=no. I am not tagging signed routes with lcn=yes, the router
can assume that unless I consider them dangerous in which case I tag it
lcn=poor. I would suggest that a router should be able to take quiet/fast as
a user input.

Sorry to labour the point, but for a basic map or for a router I really
think the router and the renderer are going to need a hand!


> Kind regards,
> Tom

More information about the talk mailing list