fernando.trebien at gmail.com
Thu Mar 20 05:53:44 UTC 2014
Mapping has a conflict: we want to be precise enough to make a useful
map (more tags), but we also want to map quickly (less tags).
Describing the surface probably is one of those problems that lies
near the middle of these opposing goals, and finding the perfect
balance is the challenge. For now, a recommendation of two tags may
not seem ideal (as would a single tag) to the mapper, but it could be
I'm entirely convinced of several things you mentioned:
- "We must use all the tags we've been discussing to describe a given
highway whenever possible"
- "In looking at our "standard" surface tags for highways, I'm
thinking now that they don't actually tell us much"
"Let's forget about surface composition and concentrate on those other
tags." My view on this is that mappers who care about routing would be
looking primarily at tracktype and smoothness (hopefully both, but
this may change per mapper and per region). But if they're going
through that trouble, it is a good opportunity to also provide the
surface tag as well, since they'll probably be looking at the surface.
(The wiki can suggest that to the mapper.)
The "surface" tag as defined today would be important:
- for studies on urban planning
- as a rough guess in routing when both tracktype and smoothness are missing
I don't see any other scenarios where the surface tag would be useful
for applications today.
"The tags surface=asphalt, surface=concrete, surface=paved all bring
certain characteristics to my mind, different ones to yours possibly"
[not really] "but they are actually useless for making routing
decisions." [mostly true]
I would change "useless" with "unreliable", but more so on other
values of surface (such as surface=dirt). Asphalt and concrete almost
always imply smoothness=excellent in developed countries, and I would
say the same is valid for Brazil, even if we have one of the worst
road infrastructures in the world (not so evident if you're in the
South, much more evident going to the countryside towards the Amazon
rainforest). Excellent smoothness is, after all, one of the goals of
any honest road construction project. Here we expect to be warned by
friends and/or the media when the pavement has a problem. We surely
would expect to be warned by a digital system that contained such
information (it is important and very unpredictable). OTOH,
surface=paved really is almost completely unreliable, no matter where
you are on the planet. It could mean we have paving stones,
cobblestones (not all countries though), or (as I remember somebody
mentioning here) even surface=compacted. Those would hardly ever
present smoothness=excellent (smooth for thin rollers, as the wiki
describes it), but they could vary from smoothness=bad to
smoothness=good most times, and that's a wide range. Material
firmness/durability would also vary quite a lot for paved, but almost
never for asphalt and concrete (which should be quite durable even
with potholes, which often just implies that nobody is fixing the
surface regularly; one thing is its current state, another is its rate
Wondering if any country would be doing worse than Brazil in terms of
road infrastructure, I found this:
And then I went searching for images of roads in these countries. Most
of the roads looked just fine (solid and smooth). The worst ones I
- in Eritrea: http://cdn.c.photoshelter.com/img-get/I0000LEZumn6Zhqo/s/900/900/ERITREA-00085-BC3.jpg
- in Mongolia: http://www.mongolia-travel-guide.com/image-files/mrm-mongolia-main-road.jpg
- in Suriname (just like some of Brazil's worst roads):
- in Angola: http://farm3.staticflickr.com/2434/3897036578_12207239a3_o.jpg
- in Angola again: http://i37.tinypic.com/r74ro0.jpg
- somewhere between Angola and Namibia:
- in Congo: http://www.orderofmalta.int/wp-content/uploads/2011/07/Congo-new-road-MI-small.jpg
- in Congo again:
(I learned yet another aspect of surface: quality of painted markings.
It seems to be a problem in Thailand, and it also is in Brazil in many
places. Should this ever affect routing decisions somehow? Should this
be mapped one day?)
How they compare to Brazil:
- typical road: http://www.halcrow.com/Global/Images/highways/brazil_roads_1.jpg
- how bad it can get sometimes (varies by region):
Even though the former is below developed world standards, I don't
think a person from such countries would rate it worse than
smoothness=good. And so wouldn't a Brazilian person, who wouldn't rate
it excellent either, due to the soft patches.
I'm quite positive people would also rate the latter with
smoothness=bad/robust_wheels or smoothness=very_bad/high_clearance.
Here the alternative names for smoothness really help making an
informed choice, almost independently from personal experience. Even
for Brazilians, some could lightheartedly describe it as "very
horrible" (if not "impassable") or just "bad" in case they don't
really think of what kinds of vehicles can pass safely.
How about tracktype in this case? If we think of durability, maybe
this may lead to some doubts: are we talking about the asphalt itself
or of the road as a whole? Do we expect the road to get fixed quickly?
Do we expect the potholes to reappear quickly after they have been
fixed? Is this predictable at all? If we think of firmness, well, it
is firm, and should probably stay firm for quite a while. Smoothness
is the limiting quality affecting speed, so both tags would be needed
for a more complete description.
The same can't be said of this road:
Here, smoothness is close to horrible/off_road_wheels. The ground is
not so firm, but durability seems to be lower than firmness would
suggest, so what should tracktype be? On firmness, I'd go with
tracktype=grade4. On durability, tracktype=grade5. Smoothness=horrible
anyway, so that aspect shouldn't affect my choice of tracktype, right?
On Thu, Mar 20, 2014 at 12:03 AM, Dave Swarthout
<daveswarthout at gmail.com> wrote:
> I just read (almost) the entire thread about smoothness Fernando mentioned
> and I must say, it looks like an uphill battle to make substantial changes
> in any definition of a road's usability for routing purposes. I reckon we
> all agree that several tags are necessary to fully characterize a highway.
> Furthermore, the choice of any of those tags will require some degree of
> subjective assessment by the mapper -- that's because people are different,
> road conditions vary by country, even to the point of saying a primary
> highway paved with asphalt in the U.S. is generally superior to the same
> sort of road here in Thailand. Can smoothness be "excellent" in both cases
> when the best highways in the U.S. are "smooth as glass" while the Thai
> counterparts are slightly less than that? (More entirely subjective
> wanderings can follow here...) It all depends on who's talking and what
> their experiences are in their own locality.
> IMO, we must use all the tags we've been discussing to describe a given
> highway whenever possible. The definitions for tracktype grades might be
> fine tuned to include the concept of durability, especially as regards
> weather resistance, and it should be somehow said up front that the various
> grades represent a scale of usability by wheeled vehicles with grade1 at the
> top. Maybe that's obvious but it should be said nevertheless. All these
> (subjectively determined) characteristics occupy positions on a continuum --
> there is no "right" choice.
> It might also simplify things if we were to somehow rate their usability for
> bicycles and roller blades separately or at least with additional tags.
> Although I like the smoothness definitions using wheels as a guide trying to
> define usability for such a wide range of purposes with the same set of tags
> is perhaps expecting too much. Using an mtb_scale or sac_scale tag could
> extend the characterization when needed as the examples indicate.
> Getting back to your question. I would say, it all depends. When I first
> read it, I thought of a stone road, sort of like a walkway, that is, flat,
> smooth stones set in a sand base. Such a "track" might be very smooth. But
> if you're referring to "crushed stone" (like railroad ballast) the situation
> could be quite different. We need more detailed information so we can employ
> more tags in such situations. That said, I would not know how to tag a stone
> road like I just described. While surface=stone and smoothness=good would
> work, none of the tracktypes fit well. If you're describing a road that runs
> over a stretch of natural stone, say in a mountain pass, then tracktype
> comes into play again.
> I assume what we're looking for is a way to assign an overall value to the
> highway so a routing engine can make correct decisions. In looking at our
> "standard" surface tags for highways, I'm thinking now that they don't
> actually tell us much. The tags surface=asphalt, surface=concrete,
> surface=paved all bring certain characteristics to my mind, different ones
> to yours possibly, but they are actually useless for making routing
> It appears to me then that we should be looking for is a way to standardize
> the smoothness and tracktype tags so a combined numerical value can be
> manipulated within a routing algorithm. Let's forget about surface
> composition and concentrate on those other tags...
> I don't know where to go from here. Anyone else?
> On Thu, Mar 20, 2014 at 6:16 AM, Fernando Trebien
> <fernando.trebien at gmail.com> wrote:
>> I think that adding the idea of "risk of degradation" is very
>> enriching to the article.
>> Just to test the concept: if tracktype means durability/endurance more
>> than firmness, what tracktype would you (and others) expect to see
>> alongside with surface=stone?
>> On Tue, Mar 18, 2014 at 2:14 AM, Dave Swarthout <daveswarthout at gmail.com>
>> > Yes, I agree firmness works better than stiffness for describing a
>> > surface.
>> > I still would prefer a term that better characterizes what Fernando said
>> > above: "To me, the idea [of] a firm/soft mixture seems closely related
>> > to
>> > "how well maintained" the track/road is, as mixtures that are not so
>> > durable/steady/firm quickly wear down and look 'poorly maintained'."
>> > A poorly maintained road, or one that is not well engineered, or one
>> > composed of loose, uncompacted materials will be much less durable than
>> > one
>> > that has those characteristics. Consequently, I still think durability
>> > fits
>> > the bill. I hesitate to bring this up but the discussion about
>> > trafficability tried to rationalize the relationship between a highway's
>> > surface, hardness, composition and smoothness and ran into similar
>> > problems
>> > (David Bannon?)
>> > FWIW, borrowing again from Fernando above I would reword the definitions
>> > as
>> > so:
>> > grade1: "heavily compacted hardcore"
>> > grade1: [Usually paved. If unpaved then a heavily compacted mixture of
>> > materials (gravel, sand, earth, clay) that provide a fairly smooth,
>> > durable
>> > and relatively weather-resistant surface.]
>> > grade2: "unpaved (...) surface of gravel [a hard material] mixed with
>> > varying amount of [soft materials] sand, silt and clay"
>> > grade2: [Unpaved (...) surface of gravel mixed with a varying amount of
>> > other materials and lightly compacted or rolled to provide a good
>> > surface.
>> > Less durable or weather resistant than a grade1 track.]
>> > grade3: "even mixture of hard and soft materials"
>> > grade3" [Almost always an unpaved dirt road. A mixture of uncompacted
>> > hard
>> > and soft materials providing a reasonable surface. Subject to moderate
>> > degradation in bad weather. ]
>> > grade4: "prominently with soil/sand/grass [soft materials], but with
>> > some
>> > hard materials"
>> > grade4: [A rougher unpaved dirt road with a mostly soft surface, poorly
>> > maintained and not very durable. Rain and other bad weather degrade this
>> > type of track rapidly.]
>> > grade5: "lacking hard materials"
>> > grade5: [A very rough unpaved track composed of loose, uncompacted, soft
>> > materials often having a surface of grass and dirt, or, in wet weather,
>> > mud.
>> > Not very durable -- easily eroded.]
>> > Other OSMers have amended this list to include grade6 and even grade7
>> > for
>> > tracks passable by 4WD or ATV only. What about those?
>> > On Tue, Mar 18, 2014 at 8:57 AM, Fernando Trebien
>> > <fernando.trebien at gmail.com> wrote:
>> >> "Firmness" sounds good to me:
>> >> http://www.merriam-webster.com/dictionary/firmness
>> >> I know that "soundness" means the same but has some additional
>> >> meanings (http://www.merriam-webster.com/dictionary/soundness),
>> >> "firmness" is more specific.
>> >> On Mon, Mar 17, 2014 at 9:09 PM, johnw <johnw at mac.com> wrote:
>> >> >
>> >> >
>> >> >
>> >> > On Mar 18, 2014, at 1:35 AM, Fernando Trebien
>> >> > <fernando.trebien at gmail.com>
>> >> > wrote:
>> >> >
>> >> > Replacing 'stiffness'
>> >> > with something else is absolutely fine with me.
>> >> >
>> >> >
>> >> >
>> >> > What about firmness? soundness?
>> >> >
>> >> >
>> >> > Javbw
>> >> >
>> >> > _______________________________________________
>> >> > Tagging mailing list
>> >> > Tagging at openstreetmap.org
>> >> > https://lists.openstreetmap.org/listinfo/tagging
>> >> >
>> >> --
>> >> 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 at openstreetmap.org
>> >> https://lists.openstreetmap.org/listinfo/tagging
>> > --
>> > Dave Swarthout
>> > Homer, Alaska
>> > Chiang Mai, Thailand
>> > Travel Blog at http://dswarthout.blogspot.com
>> > _______________________________________________
>> > Tagging mailing list
>> > Tagging at openstreetmap.org
>> > https://lists.openstreetmap.org/listinfo/tagging
>> 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)
> Dave Swarthout
> Homer, Alaska
> Chiang Mai, Thailand
> Travel Blog at http://dswarthout.blogspot.com
+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)
More information about the Tagging