Talk:Proposed features/Cycleworth

From OpenStreetMap Wiki
Jump to: navigation, search

I fear "cycleworth" is not "worth" the discussion … Because no one can follow the origination of the grades it has no worth for the users …

Finding a suitable rout for cycling is a VERY individual decision.

One cyclist may prefer tracktype=grade5, the other one only grade1
One cyclist may prefer cycleway (and footways, if no cycleway is present in this street …:-( ), the other one avoids cycleway on any circumstances because of security …
One cyclist cycles for experience nature, the other one wants to go quickly from A to B

If someone gaves a "worth" from 3 to a route, no one will know, if it is "3" for MY needs …

You may invent thousands of "worth"-tags to solve this problems …

... or use the thousands of existing tags:

surface=*
smoothness=*
(there is missing a tag for "the grass between two tracks of a highway=track" for useres of wheelchairs, three-wheel-bicycles, …)
cycleway=*
(not really good for all sorts of cycleway and cyclelanes in Germany obeying mandatority …)
(common lanes (not only for bicylces) and other features of the road should be tagged, too)
width=*
barrier=* for users of three-wheel-bicycles, trailers, …
and so on …
missing tags:
traffic intensity (cars and pedestrians …)
"touristic worth" of a route
the last one divided for (slow) non_motor_vehicles including pedestrians and (fast) motor_vehicles

--Mueck 13:12, 13 June 2009 (UTC)

Yes you may have thousands, but there are so many missing its not even funny. There is no traffic key, there is no scenery key, there is no key summarising how many cars are parked alongside a road, there is no key describing how rude cars drive on a road (and yes there can be many differences from street to street, there is no key, watch out cyclelane with many cars non watching coming in from the right hand side - danger to your life - 3 cyclists killed here last yers, on some streets cardrivers believe cyclist are obstacles....). Getting al theese tags would be a sheer nightmare. Therefore I propose this tag. Of course such a simple tag will not be perfect, but better 90% perfect than completly impossible to use at all like we have right now. Have a look at gps-tracks portals, everywhere voting the popularity of tracks is an essential and widely used feature. Maybe we need voting count on cycleworth? Before this is coded into OSM editors I prefer edit wars over nothing! --Extremecarver 14:02, 13 June 2009 (UTC)
Do you really want another edit war in the OSM wiki like the one about the smoothness tag? Because this tag has all the same problems as that one has. Probably the same people will come out of the woodwork to battle again. --Cartinus 15:56, 13 June 2009 (UTC)
Smoothness => access restrictions. Cycleworth => educated guesses on who might (dis)like the way when riding a bicycle. Alv 16:10, 13 June 2009 (UTC)
I agree that the plain cycleworth key might well be difficult to use... except for some negative values when a legal and usable cycleway has something distressing (relative to nearby ways) about it that can't be emphasized with tags depicting the physical state. People can make an assumption what an "average person cycling" (as in "where would your mom cycle") would rate it (for most ways it'd be zero). It should then be emphasized that one shouldn't tag all unpaved cycleways as cycleworth=-1 just because they are unpaved and there are paved ways nearby; preferring paved ways should be a separate choice.
With a little more definition the proposed usage classes seem quite good, though. A person wanting the fastest route would look at cycleworth_destination and the one looking for "travel in nature" at cycleworth_touring + scenery=-3..+3. (The last one is something I've contemplated on proposing; a road, say passing by any of the "wonders of the world", would be rated at +3 on that scale, a "breathtaking view" at +2 and a cycleway on the perimeter of an industrial waste dump -3. Yet it does have it's problems as a waste dump might be a breathtaking view... I haven't thought any more about the scale extents, either.)
I'd maybe add something like cycleworth_fast for experienced and fit cyclists that like to and can keep up a fast speed but aren't limited by having thin road racing tires. A way could be cycleworth=-2 cycleworth_mtb=1 or cycleworth=0 cycleworth_touring=-1.

I'm not very concerned by this subject but I see a similar proposal for Mountainbike. Could you try to unify both proposals or use similar conventions (e.g. not using negative numbers) ? My 2 cents -- Pieren 13:47, 13 June 2009 (UTC)

Nope you haven't read carefully. Mountainbike page only describes difficulties for ways. Not scenery, traffic etc... It was mainly me setting up the mountainbike page. I prefer negative numbers, because they will degrade a way, whereas 0 would mean no better not worse. But of course let's listen to what other people think about it. In the end the value itself is only a preset in JOSM or other editors. --Extremecarver 14:02, 13 June 2009 (UTC)


cycleworth_destination

Maybe cycleworth_transit? Alv 14:13, 13 June 2009 (UTC)

Yeah, that sound much better.--Extremecarver 14:56, 13 June 2009 (UTC)



This has been extensively discussed on other pages (http://wiki.openstreetmap.org/wiki/Cheltenham_Standard, http://wiki.openstreetmap.org/wiki/Talk:Cheltenham_Standard, http://wiki.openstreetmap.org/wiki/Cycle_routes/cyclability) and on IRC. It is indeed an area deserving attention but we are nowhere near a consensus. Though I applaud your interest, I'd much rather you opened the discussion - maybe even encouraging a "cycling@lists.openstreetmap.org" list for discussion - rather than proposing something ill thought-out at this point in the process. This proposal is simply too soon.

One of the problems is, as �the posters above point out, that it's too individual a decision for objective tagging. See this forum discussion for an example. You simply cannot come up with anything that is at all meaningful for more than a subset of cyclists. I have had long arguments in the pub and on other Internet forums with cyclists who disagree with my views on what makes a cyclable road!

For what it's worth, "cycleworth" doesn't make any sense at all in English - sorry. 'cycleworthy' is a bit more correct but you still wouldn't say it.

--Richard 21:26, 13 June 2009 (UTC)

Well I think even if this is a tag that may cause edit wars, it is still much better than not having that key at all! Currently it is simply more or less impossible to know whether a way is suitable for cycling or not. If there is an edit war, hell yes we have won, cause the result wil still be much better than not knowing about it at all. Regarding the name, you're absolutely correct, I would have taken cyclability but that was already taken more technically. Off course this will be individual, but I think in most cases there will be more or less a consensus. And we need something to get the mappers in the cities happy, tagging housenumbers is nothing which makes you happy if you can't get out into the nature.....
I think it might be too early for many countries but definitely not for agglomerations in Germany and a few cities in Austria and Switzerland that are more or less 100% mapped. Sorry but I'm not interested at all in keys for correct car use, I don't have a car so I will map what is interesting to me. I know that many mappers in Vienna don't have a car, I would even guess that at least 50% of all OSM participants in Austria, Germany, Switzerland are cycling frequently.
Of course we would need some subsets, but I would be against making it too complicated, though of course with JOSM presets it's usually not too hard to implement. Maybe we should straight go for an array sheet?
The thinking whether a way is in the end +1 or +2 is completely unimportant, because better something than nothing. And solely relying on cycleroutes or mtbroutes simply does not cut it. If you think there is any other possbibility of classifying a ways suitability for cycling then propose it, simply saying it is too difficult or complex does not help. If the tag is correct in 90% of cases, then this is better than currently where educated guesses based on tags result maximum in 50% suitable ways.
Opening up a cycle related mailinglist is a great idea. Can you organise this?

--Extremecarver 23:02, 13 June 2009 (UTC)

Having now read that forum discussion it seems like a user group size thing. UK: mostly "active" cyclists hurdling down the roads and some of "only these sustrans routes are safe oh they're quite inconvenient/scenic". Netherlands, Germany, Finland: Anyone not with a disability might be encountered cycling just because the weather's fine and the destination is within half an hour. Lots, or at least many, of dedicated cycleways (actually, combined foot&cycle) to choose from, but with differences that can't yet be mapped comprehensively enough (blind corners, hills vs. level, tourists gawking at the churches etc.). The sporty guys speeding among the cars make up only a minor group of cyclists. I'm confident that we can come up with the other major user groups with significantly different needs. Alv 23:42, 13 June 2009 (UTC)
Yip, seems a bit different for UK. But download my openmtbmap from the netherlands, and you will see cyclways and cycleroutes EVERYWHERE (easy to see as i highlight them). UK in comparison seems like Christopher Columbus just set foot and cartographers have not yet landed (sorry not that bad but close..). I can't remember seeing many cyclists in UK when I was there. Cycling is much more popular in other countries, and that shows of in the mapping. I'm sure that if we have the tag cycleworth_mtb within 1 year 50.000km of mtbways in Germany, Switzerland, Austria alone will be tagged according to this scheme, probabley more twice or three times that many! Cycleworthy seen general would draw even more attention. And there is a big difference in approach to Cheltenham Standard. I want the tag to upgrade or degrade existing roads, which is much easier than searching for specific roads.--Extremecarver 00:38, 14 June 2009 (UTC)

not the solution

I see the problem you want to highlite, but this is not the solution. If this key is passed, than we open for tags such as "wheelchairworth", "skateworth", "skateboardworth", "carworth" and so. It is better to identify the individual problems (such as traffic density, car parking, etc) and tag this. I know for example that there is a proposal for tagging of in street parking lanes, which will help a lot for all of this. If too many specialized tags exists we only run in the danger of "areas not being properly tagged". All of the *worth tags I have mentioned can take use of all the existing tags, but might be helped by a few additional tags, besides, each user/those generating maps, can themselves decide how these tags should be weighed. An mtb cyclist knows a 100 times more about how he likes the tracks I use whan I go in the forest, meaning I have no chance of tagging how worthy my tracks are for mtb, same, I do not skateboard either, so I do not know how good the local beach walk is for skateboarders. But factors like surface, smoothness (I really dislike that tag, but so be it), light, physical features, steepness, are all things I can relate to, and that way tag. IMO *worth tags are not the solution. --Skippern 09:33, 14 June 2009 (UTC)

80% (off the top my hat) of casual cyclists would consider the same "worthyness" for a given cycleway/normal road. Active cyclists can answer a lot of questions/set options to choose their preferred routes (parking, lit, width, segregated, traffic (in vehicles/hour or day), maximum/average incline, etc.) and this is in no way a deterrent to tagging those separately. Getting a good coverage of the inclines will take a whole lot longer than even the house numbers... The forest trails for mtb training have their own tag; the one mentioned in this is presumably for distinguishing the "very poor for all" cycleways from those which a mountainbiker wouldn't consider significantly worse from the nearby road. Osm mappers have valuable local knowledge that's difficult to present thoroughly with physical properties only. Alv 10:30, 14 June 2009 (UTC)
Hence also 80% of OSM contributors will not know how to use this tag, so in theory only 20% will have a good coverage of this tag. To me that sounds prety useless. On the other hand, the few who generates premade cyclemaps should have this knowledge, and therefor should be able to estimate this depending on existing tags. When the tag is suffering the same weaknesses and objectivity as smoothness, we will probably rather end up with another edit war instead of a tag. --Skippern 02:26, 16 June 2009 (UTC)
No, the 80% will know looking at a way "i'd like to cycle down this one and can presume other casual cyclists would also" or "i'd avoid this way" ("since i'm not a active cyclist speeding down the car lanes"). The last 20% are either active cyclists who then tag for the group they identify to (racing bicycle, touring, maybe others). If you ask yourself for every cyclable way: "if my mother were cycling, would she consider this way average/worse/better than average for cycling?" you'll end up with a good value. People can empathize and estimate others' opinions.
As to generating these from tagged physical features, there won't be sufficient data in decades for most parts of the world to be anywhere near exact. Would it be better to start with
way 1 near a road way 2 in the forest comment
surface=paved surface=unpaved
unpaved=compacted
surface:winter=paved surface:winter=unknown
lit=yes lit=yes
average_noise_level=65dB average_noise_level=45dB who's gonna measure these
average_noise_level:rushhour=70dB average_noise_level:rushhour=55dB
wind_cover=none wind_cover=wood
wind_cover_effect=0% wind_cover_effect=70% how could we measure these?
maximum_altitude_difference=25 maximum_altitude_difference=25
ele=* on nodes on every peak and valley 5 times as many nodes with ele=* who's gonna measure these - srtm isn't detailed enough and doesn't cover the whole populated world.
winterplowclass=1st winterplowclass=4th only deduced from local knowledge, not available from the authority
seasonal:winter:gritted=yes
moped=yes
segregated=yes segregated=no
width=4 width=3
free_visibility:average=400 free_visibility:average=50
scenery=highway scenery=wood
or start with or start with
surface=paved surface=unpaved
cycleworth=0 cycleworth=1
cycleworth_touring=-1
cycleworth_transit=1 cycleworth_transit=0
cycleworth_rr=2 cycleworth_rr=0
seasonal:winter:cycleworth=1 seasonal:winter:cycleworth=-2
seasonal:winter:cycleworth_rr=-3
If a user knows what boundary conditions he has (say "I want to avoid possibly windy sections") he can code a navigator to use that information, if it's present. Many casual cyclist would likely find it hard to accurately describe their preferences and their internal relations (for example is a windy stretch worse than an unpaved but smooth way). Alv 11:16, 16 June 2009 (UTC)
I see you are able to define what you want in the various cycleworth values, I can't, and therefor the areas where I map will lack cycleworth values (unless somebody else put it there). I know that I am not the only one, in fact, most OSM'ers tag for their need and interest. This is perfectly fine by me, but when it comes to such tagging schemes as this cycleworth where you can derive the values from already existing tags, than we should rather focus on improving the existing tags (for example, double check that all roads in your area have surface and lit tags in place). We could probably make houndreds of *worth tags without doing any improvements to the map, and without getting one useful *worth tag. Besides, most people who want a cycle map does not generate them theirselves, but looks for pre-generated maps on the net. That way, those who generates the map should make the rules for their maps, independent of a cycleworth tag. --Skippern 07:50, 18 June 2009 (UTC)
Off course we could make 100s of worth tags, and I think that in some years consensus will be that we actually do need them. Simply to difficult to tag all those factors individually, and more or less impossbile for any autorouting to actually decipher them and put them into context, as long as a tag is applied to a line it's easy, but I don't know of any routing engine that right now can handle POIS for autorouting information, let allone classification. Making a cyclemap route well without cyclewort tag, will be more or less impossible for the next few years. 20 different tags means 2°19 combinations to check for to get the information sorted, simply impossible. Thinking factors is theory, adding worth tags is practical. tracktype is also highly objective, we could also say don't tag it but better analyse granularity of surface you are on, include chemical analysis and so on.... There is no way around simplification.--Extremecarver 11:20, 22 June 2009 (UTC)
Extremecarver, you do not need 20 different tags, but more like 3-6 of them. And even if they have ratings 1-10, you would not need to evaluate 10^n possibilities... Your preference would be a simple formula giving preference to that few tags. For example, if tags are "scenic", "condition", "shortness" you could enter just 3 numbers and make your preference be = 7*scenic -3*condition -2*shortness if you are for an easy scenic ride, or 2*scenic+7*condition-5*shortness if you are in for workout, or 0*scenic-5*condition+9*shortness if you're just in hurry to get somewhere, etc... and pick the route that computer calculate (very quickly) scores highest. It is simply not possible to get anywhere near that with a single (or few of them, or even a few hundred of them) "cycleworthness" value(s). --mnalis 23:06, 11 October 2009 (UTC)
Maybe we should stick for the content of the main OSM database to store only verifiable "facts", so no worth tags. But we could maybe think about a system of storing subjective information in a standard way externally to the main OSM database. (There is of course the issue of instable OSM id's, so I am not sure how to link the subjective information to the OSM elements, some work in the editors would then be needed to update as well the external databases) Or we agree on a new namespace for tags such as "subjective:" which should be considered as "write once". If somebody does not agree with the value, he should add an other tag with his opinion (instead of correcting it). Like this a smart algorithm can use the "average" to make a cyclepath proposal. See my proposal for "voting" below--Behrica 17:13, 22 June 2009 (UTC)
I completely agree with sticking to basic OSM principle of Verifiability. This proposal is so subjective it simply couldn't work as proposed. It is like trying to tag food with 'food:tastiness=1-9' or 'book:readability=1-9' or 'joke:funniness=1-9'. What is good for one, will be totally wrong for another, even if you try to lump then in few different groups like "mtb", "touring", etc. Additionaly, even if you have to get subjective, do it by some tangible effect (like: "is it scenic?" or "is it hard to ride?" or "do you get muddy" etc) and not by coming up with some imaginery group "global happines index" (like "would an average MTBer like this?" -- there is no such thing as an average MTBer!!). Have a look how a more specialized groups make that at gspies.com. In addition to more or less verifiable facts (like elevation maps, ride length, photos, terrain type and condition etc), even when they have to be subjective, they implement per-track voting for different values (like landscape, condition required, fun factor, linking to track reviews etc), so one can choose if the track is just out of his league due to poor physical condition, or if you look for physical activity if it is goind to be too undemanding; are going to be OK with fighting with cars for piece of the road in order to make it somewhere in time or would you rather take your time and enjoying the scenic views instead etc. Even the same person on the same bike on different day of the week can love one set of features, and hate it one the other day -- single "cycleworth" value couldn't make it. To make a point again, even if you must go into subjective waters, do not use "food:averege_american:tastiness=1-9" but instead a combination of "food:hotness=1-9", "food:spices=chinese/mediteranian/mexican/...", "food:satiating=1-9", "food:visual=1-9" etc. --mnalis 23:06, 11 October 2009 (UTC)

Seasonal variations

It's been 10+ years from the last time I cycled through the snow with studded tires... There's the proposed seasonal:winter=* for basic "not available in winter/other season", but how'd this be used relate to ways that are perfect for cycling in the dry season but most would avoid say, when the snow is still melting? Nearby ways would have been plowed all along the winter and already dry while the ones in parks turn wet and icy, regardless of their surface. Likewise in winter the main routes are kept in drivable condition ( =0 ) and some cycleways near them are likely possible but the casual cyclists would rate them a -2 in winter.

Or this just needs to be acknowledged as "weather or season variation might invalidate these basic values".Alv 17:55, 14 June 2009 (UTC)


use votings tags

--Behrica 09:51, 15 June 2009 (UTC)

Maybe a kind of voting system could work. Something such as
"votes_cycleworth=10" (where each mapper which "likes" the way for his purpose adds +1 to the number (or -1, if he disagrees).
This could over time give a "correct" (= a majority of people like a way for a given purpose). This could be maybe supported by creating a special web-tool for this kind of voting.
This could give a new cycler of that area at least a first hint, where it is worth to cycle.

objective tags

I think OSM tags should be as abjective as possible. So better go for some additional tags to describe a way, then to combine different tags into a *worth tag. Since there are too many different types of cyclists (even if you stay in the same "class", e.g. mtb or rr), a *worth tag won't be of too much help for many of us, since the requirements (or definition of a good way for cycling) are very different. It won't help me to decide, which way to go.

Perhaps better define a basic set of tags, which should be used for tagging ways (e.g. for mtb I made a JOSM preset, available on File:Mtbpresets en.odt, where the basic information is in) and adding additional tags where necessary or helpful. --Kaivi 10:39, 30 June 2009 (UTC)

And especially for mtb, a *worth tag must be dependent on the direction of travel. Many ways are great for downhill, but unridable uphill. --Kaivi 18:24, 2 July 2009 (UTC)

cyclist groups

key description requires common preferences variable preferences
The original proposed groups are
cycleworth casual cyclist (as in "could be your mother") avoid steep hills

avoid high density traffic

cycleworth_mtb on a mountain bike but not off road scenic singletrails mtb:scale=1-2

mtb:scale:uphill=0-2

highway=path

scenery
cycleworth_rr racing bike user reasonably smooth surface (so can't be surface=unpaved) prefer roads where it's possible/safe to drive fast

sloped curbs

dedicated cycleways vs. roadside cycling
cycleworth_transit commuter, cycles often avoid detours

avoid very steep hills

avoid unsafe cycleways

avoid (very) high traffic

avoid routes shared with pedestrians

cycleworth_touring longer distances avoid high traffic roads scenery

low traffic

Other groups could be
cycleworth_mtbdownhill downhill and freeride biking (sorry, this is not meant too seriously, but comparing this to cycleworth_mtb describes the problem of mixing up too much tags in one...) steep hills and very difficult trails avoid going uphill

mtb:scale:uphill=0 go down every very steep hill

mtb:scale=3-5

avoid highway=track

use chairlifts

Feel free to add in preferences (or even lines) above. Alv 09:55, 2 July 2009 (UTC)

This is getting dafter by the minute. If the mapper is going to have to make 6+ subjective ratings like this, it would be a whole sight easier and more useful to tag 6 objective factors and let the client decide. I have no idea into which of the first five categories I fall. --Richard 07:12, 3 July 2009 (UTC)

Not "have to", but "can" when they see something especially nice or worth avoiding - most ways are the same as other ways around - and it's likely that people can't make a difference unless they know the area thoroughly already. Likely your category varies with every trip. Alv 12:11, 3 July 2009 (UTC)
Yeah, but the first point is true about objective factors too. You don't have to add an express gradient tag unless the gradient is remarkable, for example (and not SRTM-derivable). Second point - no, that's the point: I don't fit into any one of those categories even on a given trip. I'm going cycling this afternoon, for example, to Oxford (our nearest city) on a short ride in which I'm equally "casual cyclist", mtb and touring. If the cycleworth tags can't even cope with that... --Richard 09:27, 4 July 2009 (UTC)
There's always a group "other" but I'm hoping that with discussion we could get a 80-90% coverage. Alv 09:59, 4 July 2009 (UTC)