Even if we choose the surface tag, I believe new values for tracktype can be pushed no matter what we decide in this discussion, it seems a lot more reasonable than using the 4wd_only=yes tag (so don't give up yet :D ). I was going to naively propose tagging more measurable qualities of the road, but I wonder how easy and accurate it could be: http://en.wikipedia.org/wiki/International_Roughness_Index
>From a Brazilian user perspective, I see flaws on both tags. For instance, the only way I currently have to represent that an asphalt road has bumps is using the (infamous) smoothness tag. I thought again about the unpaved rendering question and I'm beginning to lean away from the tracktype tag, but let me explain how I'm reasoning. Let me first say that, being a programmer, I'd definitely prefer to use the tracktype tag (it's so much easier to implement and maintain), and I think Matthijs' position in this discussion is similar (prefers tracktype for easy maintenance, but feels like it's not quite adequate for everyone). So, which approach is most convenient for the rendering app? Definitely it's using the tracktype tag, there's only 1 value to be ever considered. Which approach is most convenient for mappers? Probably it's using the surface tag, since the tracktype tag requires knowledge that is not so evident from the names in tag values. Mappers could, however, learn it once and apply everywhere. So it's perhaps a "minor" inconvenience (that may not be so "minor" when talking about novice users). Which tag is most meaningful and to which users? I think tracktype means a lot to car drivers, but surface is more meaningful for other types of users. On the other hand, a surface tag may not be so descriptive for drivers but it's not unusable either (it just requires mappers to choose more specific values), whereas I believe many kinds of users (like cyclists, wheelchair users, pedestrians) view tracktypes 2-5 as quite similar, even after heavy rain. For a variety of users, a more varied description of the surface would work better. But is it descriptive enough for all situations? No, and that's why you and others have been advertising the tracktype tag. How about in emergency situations, like crisis mapping? They need a simple system to work quickly. Maybe this is why the HOT team chose the surface tag in their map style, because it's more obvious than tracktype, allowing new mapping workforce to be easily integrated. I remember having tagged a few tracks within parks here in my city. It took me a while to grasp the exact differences between each track grade. So should we abandon tracktype in this debate? I don't think so. If "surface" makes more sense, but "tracktype" is easy to code, I don't see why we can't combine both approaches. Our recommendation could simply be: use the tag you prefer. Both say something about the surface. Let me try to suggest something new. Reading the descriptions in the wiki, I think one could say that: - surface=paved/compacted corresponds to tracktype=1 - surface=gravel/fine_gravel corresponds to tracktype=2 - surface=grass corresponds to tracktype=5 - surface=dirt/earth/sand is detailed by tracktypes 3-5 and by the proposed tracktypes 6-8 - 4wd_only=yes corresponds to tracktypes 6-8 Perhaps an alternative to describe all these situations can be: - replacing surface=dirt/earth/sand with surface=soil + soil=dirt/earth/sand + soil:grade:car=1/2/3 - sticking with the 4wd_only tag or maybe adopting a minrideheight tag On Thu, Jan 2, 2014 at 4:05 AM, David Bannon <dban...@internode.on.net> wrote: > On Wed, 2014-01-01 at 22:57 -0200, Fernando Trebien wrote: >> Welcome, David. If you've just been advised about this discussion, you >> may wish to read it from the start: >> http://gis.19327.n5.nabble.com/Tags-useful-for-rendering-of-roads-in-poor-conditions-td5791303.html >> > Actually, the particular issue has been on the table for five years ! > Please see https://trac.openstreetmap.org/ticket/1447 > We did not get anywhere in the life of mapnik .... > >> .... classifying roads as highway=track simply >> because they are unpaved > > Yep, thats a bad thing. Its important that the highway= tag is not > intended to address the state or condition, its meant to describe what > the road is intended for. But the special case of highway=track messes > that up ! > > There is no reason to assume tracktype= applies only when highway=track. > An unfortunate semantic link, thats all. > >> I'll try to summarise: I'm pretty much open to using either the >> surface or the tracktype tag, > > Agree. What is important is that we come up with one recommendation so > people know what to put into OSM and renderers have a reasonable > understanding of it. Tracktype= outnumbers surface=unpaved 2:1 in the > database. But there are a lot of other surface= values. In some cases > surface= values are subsets of other values and sometimes they are > mutually exclusive. So people are sometimes unsure what to use. So in > the Aus guidelines, we encourage people to just use surface=unpaved. We > also advocate use of the 4wd_only= where necessary. > >> .... To make >> tracktype popular with highways that are not highway=track, I believe >> we'd need to request JOSM's developers to add a tracktype field to >> many presets, so it's extra trouble. > > I think the JOSM developers are more likely to follow common usage. If > enough people ask for that then the developers will implement it. But > its quite easy to do in JOSM 'manually'. I routinely do so. > >> .... that tracktype=grade1 can >> be considered "paved" (for rendering) and any other tracktypes can't. >> But in the absence of that tag, we can then use the surface tag to >> decide. > > That may be more complicated than needed IMHO. In practice, few if any > rendering engines look at surface=. Lets leave it that way. If > tracktype= is not asserted, we already assume its paved or similar. > > I guess my point is that to 'fix' the problem, all that needs to happen > is the renderers observe tracktype whenever its there, not just when > highway=track. > >>If I were mapping in Brazil, I'd certainly prefer "paved" as default, > > Agreed, thats far too much the default to dream of changing it. > AlaskaDave said the same thing. > > Anyway, if we all agree to recommend "surface=" (rather than tracktype=) > as the trigger for 'different' rendering, I would be willing to go > along. Unhappy but willing. I guess all that would need to happen is we > add a note to the surface= wiki page to say values listed under > 'Unpaved' be shown differently. > > I would be a bit sad if we went that way as my dream of adding three > extra grades to tracktype= would be a bit harder but at least the more > common cases would be dealt with. And that is important. > > David > > > > -- Fernando Trebien +55 (51) 9962-5409 "The speed of computer chips doubles every 18 months." (Moore's law) "The speed of software halves every 18 months." (Gates' law) _______________________________________________ Tagging mailing list Tagging@openstreetmap.org https://lists.openstreetmap.org/listinfo/tagging
