[HOT] Trunks in Highway Tag Africa

Ralf Bernhardt raabee at gmx.de
Tue Jan 26 14:29:51 UTC 2021

South American communities have established their own rules.
What if Highway Tag Africa conflicts with national tagging conventions?


Am 26.01.21 um 14:02 schrieb Fernando Trebien:
> On Tue, Jan 26, 2021 at 7:51 AM Jean-Marc Liotier <jm at liotier.org> wrote:
>> With primary currently being the top of the topological hierarchy, what
>> objective criteria would differentiate it from trunk if trunk becomes
>> defined as "even more important than primary" ?
> As in Europe and in the Anglosphere, objective criteria probably
> should vary by country as conditions are not uniform across the
> continent. I mention the Anglosphere because the communities of
> countries like Canada [1] and Australia [2] have developed their own
> set of rules based on the more general descriptions of the English
> wiki.
> Recently in Brazil, [3] we adopted the rule "the main paved route
> between any two cities/towns with more than 200k people is trunk, the
> main paved route between any two cities/towns with more than 20k
> people is primary." This mainly assigns trunk to national routes and
> primary to regional routes in developed areas while working around
> oddities (when a regional route was built and a nearby national route
> was not and the regional route currently replaces its function and the
> national route will not be built soon and people use the regional
> route heavily because it is the best option locally) and without
> leaving large areas of the territory without a trunk network (which is
> the result I would expect in Africa if the current definition in
> Highway Tag Africa is applied to all places on the map in Africa). In
> Brazil, the resulting network is also highly correlated to maintenance
> frequency (which is strongly correlated with traffic intensity), which
> is a desirable property for optimal and safe driving. Two mappers
> independently assessed the resulting trunk network (if we had a third
> person, it would be even better), checking each other's work to ensure
> that main routes are chosen correctly and that none are missing and
> that none are mistakenly included unnecessarily, then they presented
> the result to the local communities, who suggested some adjustments
> (mainly replacing short stretches with poor infrastructure with better
> alternatives nearby). This will likely require continuous fine-tuning.
> Something similar happened independently in Argentina. [4] The idea
> for this was based mainly on descriptions on the wiki of other
> communities and on the idea that classification should be done mainly
> by importance instead of physical attributes, [5] and it also
> corresponds mainly to the functional classification manual provided by
> the Brazilian national infrastructure authority (which does not
> publish the resulting classification). The main point is that a
> divided "motorway-like" road is not necessary where traffic is not
> intense, and even a 2-lane single carriageway (1 lane per direction)
> may be the best road available for hundreds of kilometres in some
> places and vital for the region they cross.
> In Bolivia, the local community applies highway=trunk to any national
> route, regardless of whether it is paved or not. You can check the
> unpaved ones in this image [6] or using an Overpass query. [7] The
> inclusion of unpaved ways in the trunk network was rejected in Brazil
> and Argentina, which managed to pave their most important roads (with
> some exceptions in the North Region of Brazil). It makes some sense
> that the Bolivians want to have their highway classification this way.
> Since the "problem" with these ways is described on the map using
> surface=*, applications should be able to make sense of them.
> In Uruguay, the government provides the classification of roads. [8]
> AFAIK, none of the trunk roads there are officially restricted to
> motor vehicles [9] and most are only 2-lane single carriageways (1
> lane per direction). Even so, they are very safe and offer excellent
> mobility. The population density is low, so traffic intensity is low,
> so there is no need to build expensive roads beyond this standard,
> except near the metropolitan area of Montevideo. This is similar to
> the situation of the trunk roads that cross the Outback in Australia,
> such as Stuart Highway. [10]
> In the UK, route A537 [11] is considered one of the most dangerous,
> [12] not being restricted to motor vehicles and its characteristics
> [13] may be below those required for highway=trunk on Highway Tag
> Africa, but it has been highway=trunk in the UK for years. So, again,
> Highway Tag Africa seems to promote the high standards of some
> countries that are not even adopted by some other developed countries
> with abundant infrastructure.
> [1] https://wiki.openstreetmap.org/wiki/Canadian_tagging_guidelines#Trunk
> [2] https://wiki.openstreetmap.org/wiki/Australian_Tagging_Guidelines#Unsealed_and_4wd_Roads_.28Dirt.2C_Gravel.2C_Formed.2C_etc.29
> [3] https://wiki.openstreetmap.org/wiki/Brazil/Classifica%C3%A7%C3%A3o_das_rodovias_do_Brasil
> [4] https://forum.openstreetmap.org/viewtopic.php?pid=711105#p711105
> [5] https://wiki.openstreetmap.org/wiki/Proposed_features/Highway_key_voting_importance
> [6] https://i.imgur.com/58YMeFA.png
> [7] https://overpass-turbo.eu/s/12NM
> [8] https://wiki.openstreetmap.org/wiki/ES:Uruguay#Correspondencia_DNV_-_OSM
> [9] https://www.montevideo.com.uy/Noticias/-Pueden-los-ciclistas-circular-por-las-rutas-nacionales-Hay-un-decreto-pero-no-se-fiscaliza-uc326377
> [10] https://www.openstreetmap.org/?mlat=-23.02974&mlon=133.61183
> [11] https://www.openstreetmap.org/?mlat=53.2508&mlon=-2.0108
> [12] https://www.dailymail.co.uk/news/article-1290797/Britains-dangerous-road-named-A537.html
> [13] https://goo.gl/maps/Ag68KgNEGd6sPdTq7

More information about the HOT mailing list