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

Peter Miller peter.miller at itoworld.com
Fri Aug 31 11:29:02 BST 2007

Comments to 2 messages in-line

> Message: 1
> Date: Thu, 30 Aug 2007 10:43:05 +0100
> From: David Earl <david at frankieandshadow.com>
> Subject: Re: [OSM-talk] dangerous cycling lanes (was Re: A new
> 	highwaytaggingscheme - thinking about)
> To: Ian Sergeant <isergean at hih.com.au>
> Cc: talk at openstreetmap.org
> Message-ID: <46D69129.8050606 at frankieandshadow.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> On 29/08/2007 23:32, Ian Sergeant wrote:
> > We need to capture things that make cycling good or bad on a route.  We
> > need to capture the elements of what makes a good cycling route, traffic
> > volumes, rough surfaces, high pedestrian volumes on paths, squeeze
> points,
> > inclines, propensity for debris, lighting, speed limits, etc, rather
> than
> > an subjective assessment.
> As Peter pointed out, The Cambridge Cycling Campaign online planner
> offers you the choice of 'quickest', 'shortest' or 'quietest'. The
> result isn't derived from lanes/paths/roads etc being marked just quiet
> or fast, but by each route (there may be alternative routes on the same
> street - a cycle path on the pavement and the ordinary road, for example
> even sometimes more than two) having a single designation chosen from
> about 25 alternatives (such as 'segregated shared use path', 'main road'
> etc). Each of these has a weighting for speed and quietness and the
> route is the one with the best balance of speed vs distance or quietness
> vs distance. You can also say how fast you typically cycle.
> The key point is that a single (substantially objective) parameter is
> sufficient to provide a highly effective route plan. There aren't any
> significant hills in Cambridge though.
> David

Personally I think this may be a pretty good compromise between capturing
loads of different elements. The single tag can be interpreted by different
people for different purposes and is as objective as is practical. Fyi, here
is a list of current tags using in Cambridge.


>From this single tag a number of different detailed tags could be populated
(Speed Limit), Width, Traffic Level etc. Conversely, given the tag
'primatives' (Speed/width/traffic volume/classification) the CamCycle
combination tag could sometimes be deduced. 

Different renderers and routers would then use different lookup tables
between link type and rendering style/speed/quietness etc.

> ------------------------------
> Message: 2
> Date: Thu, 30 Aug 2007 10:49:26 +0100
> From: Stephen Gower <socks-openstreetmap.org at earth.li>
> Subject: Re: [OSM-talk] dangerous cycling lanes (was Re: A new
> 	highwaytagging scheme - thinking about)
> To: talk at openstreetmap.org
> Message-ID: <20070830104926.A16113 at calculus.wolf.ox.ac.uk>
> Content-Type: text/plain; charset=us-ascii
> On Wed, Aug 29, 2007 at 09:29:24AM +0100, Tom Chance wrote:
> >
> > On Wed, 29 Aug 2007 16:07:12 +0800, Cameron Patrick
> > <cameron at patrick.wattle.id.au> wrote:
> > >  I'd go for "cycling=poor|medium|good"
> >
> > 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.
>   I'd prefer easy|intermediate|advanced|dangerous or something
>   similar that doesn't mention "good". It's still going to be
>   subjective to a certain extent, but I think there would be
>   consensus on most roads once there was a feel for it.
> > 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.
>   Two problems with this that I see:
>   1) We're trying to make maps! Distilling four keys (each with
>   multiple possible values) into simple rendering rules, is going to
>   be a nightmare.  And personally, I don't think the four
>   "characteristic keys" listed on that page are enough to classify a
>   road's suitablity for cycling - making the job even more
>   complicated. Do we even have keys for saying "this road is often
>   used by HGVs", "this road has nasty traffic calming meaning cars
>   often vear suddenly to one side", etc?  A combination of
>   cycling-grade (easy, advanced, etc) along with a "description" (or
>   perhaps a specific cycling description) tag describing the factors
>   would be easy to render and therefore might actually happen!

Looking at it the other way round we are going to need to render maps for
many different users, not just car drivers and they will need a good way of
indicating to renders how to colour links for the group.

A map will normally use about four different colours/patterns to distinguish
different qualities of link (for car drivers it is motorway, trunk etc) and
there needs to be a logical relationship between the tags and the rendering.
The CamCycle approach seems to work for this.
>   2) I could probably assess from memory most of the likely cycling
>   roads in Oxford with a cycling grade.  It would take a few hours
>   sat in front of a computer to enter this, but nothing like the
>   weeks of work it would take to resurvey every street to the level
>   of detail needed by characteristics tags.

Agreed! I am also wanting to avoid lots of anal surveying when actually I
just want a few different coloured lines on a cycle map, but the complaints
about 'purely subjective' to 'good' poor' etc have merit. I assume that you
could also be able to pick from the CamCycle list from memory. Possibly it
would be better to get about 4-5 cyclists together who know different bits
of the town (and also have different opinions) and populate the tag across
the town.

The CamCycle tag list would allow different sorts of maps to be produced for
'advanced' cyclist and also for 'beginners' or racing cyclists and would
also provide more information for a router.

The CamCycle tags are really 'meta' tags containing lots of different
information about width, traffic Volume, surface etc. 



>   Just my 2p!
>   s

More information about the talk mailing list