Revisting the highway classification system

I agree that we ought to get rid of the number-based classification scheme. “Importance” as I used to say back in 2013, or “performance” as Dave put it not so long ago, is the main criterion elsewhere.

Speed limits, traffic lights, speed bumps, etc. cannot be detected from imagery. Since almost all our data are from imagery, most of the times such tags are missing (and by the way, speed limit tags are missing on most of tertiary and lower roads in Germany - i.e. a really well mapped country with a large local user base - too).

While in Thailand, I use to travel on bicycle (no motor). I try to avoid busy roads, and I normally plan my trip by looking at the map, and use the routing features only occasionally. Hence, rendering must provide that information (fortunately, I understand map creation for Garmin, and have some basic knowledge with map creation for Osmand).

In my experience, the lower classified roads (unclassified, residential, service) are far too often classified below their actual status. Often, main roads between cities are unclassified or even residential where I’d prefer at least secondary.

Long roads may have sections of different classification. Just look at K9652 near Karlsruhe (Germany):
trunk: https://www.openstreetmap.org/way/278159586
tertiary: https://www.openstreetmap.org/way/26661548

Currently, I do not touch the highway tag added by others anymore. Instead, I add temporarily a new un-official tag which I use to overwrite existing classification when creating the map.

I want to throw in another aspect of how road classification actually appears.

With Facebook we saw a massive increase of roads painted from aerial imagery. This happened before as well, but at a much smaller scope.

Now: From a satellite image you can’t see any road signs. You could guess that a road might connect two villages or stretches further and connects multiple cities.

But to match with our tagging, we gave such roads at most an “unclassified” tagging unless someone on the ground reviewed them and added the road number.
And this is the best-case. I changed multiple Facebook tracks in tertiary as they had nice paved surface and 4-digit numbers.

Some of such half-mapped roads are visible on http://compare.osm-tools.org/ in the lower zoom levels as the road is not in the vector tile set on that zoom and is displayed as unmapped before zooming in more.

This plays another role in the large overhead of the residential/unclassified set of roads. They simply did not yet get the attention of mappers to classify them correctly.

Certainly: The less room for interpretation and personal opinion we have on the tagging of a road the easier will it be to come to a common level.
I want to avoid edit-wars where one party upgrades a classification because “this is a road people frequently use as a shortcut” and others downgrade with “this road is not intended to be used as a shortcut and the traffic disturbs people”.

Getting rid of the number-based scheme entirely seems rather extreme to me. I think they’re still useful for at least some basic guidance. While there are cases where the ref poorly reflects a road’s importance, these are probably still the exception rather than the norm. At least with the lower levels (secondary and tertiary), I’ve found that most of the time in rural areas, four-digit national highways are more essential than rural roads in getting to a destination. I’d suggest not changing the default guideline on these, even if the distinction is primarily administrative. (There are thousands of such highways—it would be impractical to make a separate decision for each and every one of them.) Where exceptions are warranted, I’d suggest documenting them, as mentioned above.

As for the higher levels, I do think we should try to avoid having dangling ends. The current scheme essentially defines a trunk highway based on its capacity, but there’s also another aspect to a road’s “importance”: its place in the countrywide network. Take Hwy 11, for example. It’s the second-most important link into the North, but until only recently was a 2-lane road most of the way. The same goes for Hwy 12, which is currently undergoing expansion under the East-West Corridor project. I’m quite bothered by the abrupt changes in Highway 33 on OSM, not least because it’s also designated as Asian Highway 1. These routes’ roles in the wider network should warrant classification as trunk, even if they aren’t four-lane dual carriageways all the way. (Also, most of them are under planned or current expansion anyway.) This is where the highway numbers are more useful—a highway’s number doesn’t tell much about the amount of traffic it serves, but gives a pretty good idea of the distance it covers. I’d say it would be simpler to apply highway=trunk classification to all one and two-digit national highways, with exceptions noted on a case-by-case basis (such as the above Phang-nga example).

This would narrow the area of uncertainty down to three-digit national highways. Keeping with the above, I’d suggest writing down the routes to be classified as trunk. In contrast to the current guideline, I think a highway’s relative prominence should also be taken into account, so not all dual carriageways need be upgraded to trunk (e.g. the Hwy 347 vs Hwy 32 case).

Maybe I could try digging into the current data and come up with a preliminary list. Not sure how hard this will be…

And yes, as Nitinatsangsit mentioned, urban roads should probably be covered in a different discussion.

I’m thinking about how to compromise between importance and performance (For me, in ideal situation both things should be the same.). to the main aspect that is to make the map easier to use.
Having the number-based scheme as a base point, with an exception, is quite acceptable for the purpose of consistency.

I agree that all 4-lane highway shouldn’t set to be trunk. For 1 and 2 digit highway, I think most of it can be trunk, except some section with too low performance, e.g. have only 2 lanes, or 4 lanes but with many traffic light and no island (whether physical or painted). These are roads that serve light traffic and people are using other roads instead to travel for long-distance, so give the trunk rank to that more important road. However, a short section that doesn’t meet the requirement can be ignored, and a short section that would be dangling ends should be avoided.
With these kind of guidelines:

  • Hwy 1 from Lop Buri to Chai Nat Interchange is to be degraded to primary, with Hwy 311 from Lop Buri to Sing Buri be trunk instead.
  • Hwy 11 from In Buri to Wang Thong is to be degraded to primary, with Hwy 117 from Nakhon Sawan to Phitsanulok be trunk instead. (This is the current situation in OSM.)

For 3 or more digit, or even with province prefix, it could be trunk if it is look like motorway, except the toll gate, and represent the countrywide main network, such as run between main cities, with no other trunk nearby and most used for long-distance traveling. (we can set a minimum length, like 50 or 100 km.).
With these kind of guidelines:

  • Hwy 314 then 304 from Bang Pakong to Nakhon Ratchasima should be trunk. (This is the current situation in OSM.)
  • Hwy 212 from Mukdahan to Ubon Ratchathani should be trunk.

By the way, I think current guidelines in wiki is quite acceptable, maybe we can discuss for a slightly change. But for urban roads, I think we should have a consensus then add it to wiki.

Actually, since there are just a handful of 1- and 2-digit highways, having a default classification wouldn’t really matter if we describe them all.

I looked through the extant highways, and here’s how they’re currently mapped:

• Hwy 1 and 2 are trunk all the way except when passing through Saraburi (bypassed by Hwy 362).
• Hwy 3 is primary in Bangkok and Samut Prakan, and trunk between Chonburi and Trat, after which it becomes primary the rest of the way to the Hat Lek border. Also primary in the cities of Chonburi (bypassed by 361) and Rayong (364 and 36).
• Hwy 4 is trunk except where bypassed around Hua Hin (Hwy 37), from Khuraburi to Khao Lak (no bypass, which makes it a hanging end), through Thai Mueang (a short missing segment), the approach to Phang-nga (bypassed by Hwy 4311), through Trang (Hwy 419) and Sadao (สข.1008)
• Hwy 11 is primary from Inburi to Wang Thong, and Trunk the rest of the way.
• Hwy 12 is trunk except through Sukhothai (bypassed by Hwy 125) and Phitsanulok (Hwy 126), and where the new segment is under construction south of Kalasin.
• Hwy 21 is trunk only up to Lom Kao, after which it becomes primary the rest of the way to Loei.
• Hwy 22 is primary the entire way from Udom Thani to Nakhon Phanom.
• Hwy 23 trunk up to Roi Et (meeting Ring Road 232), except through Maha Sarakham (bypassed by Hwy 291). It’s then primary the rest of the way to Ubon Ratchathani.
• Hwy 24 is trunk up to Det Udom, and primary from there to Ubon Ratchathani.
• Hwys 32, 34, 35, 36 & 37 are trunk in their entirety. Hwy 31 and 38 are in the Bangkok area, which I’ll exclude from this analysis.
• Hwy 33 is trunk except between the junction to Prachin Buri and Prachantakham (missing segment), and through Aranyaprathet (bypassed by สก.2089)
• Hwys 41, 43 & 44 are trunk.
• Hwy 42 is primary from its beginning at Khlong Ngae to Pattani, then trunk (after Hwy 418) to the junction to Tak Bai, then primary to Sungai Kolok and the border.

Here are my comments:

• I don’t quite agree with downgrading Hwys 1 & 2 around Mittraphap Junction. The junction is outside the town of Saraburi, so traffic isn’t usually an issue, and the bypass takes quite a long detour and has traffic lights. If you’re driving from Bangkok to the Northeast, it makes more sense to take the junction, and vice versa. (You’d take the bypass if headed to Lop Buri.) I suggest that these be trunk except Hwy 1 between Mittraphap Junction (where it meets 2) and km 111, where it meets the bypass on the northern side.

• Perhaps Hwy 2 should be downgraded where it passes through Nakhon Ratchasima (easily bypassed by Hwy 204). Not sure about Khon Kaen and Udon Thani.

• The hanging end of Hwy 4 along the Phang-nga coast is not okay at all. It should either be trunk all the way from the Chumphon junction to Khok Kloi (the junction to Phuket), primary all the way, or Trunk until Ranong, which is the most major town along this segment.

• The bypass around Sadao seems tiny compared to the through route. Not quite convinced this segment shouldn’t be trunk.

• Lom Kao is a tiny town. Hwy 21 should be either trunk or primary all the way.

• There really should be a trunk road to Ubon Ratchathani. Both 23 and 24 should probably be made trunk all the way.

• Make Hwy 33 trunk north of Prachin Buri. Not sure about Aranyaprathet; one would take the bypass if headed for the border, but the main route is the designated AH1.

• Hwy 42 should probably be trunk to Sungai Kolok. (It’s already dual carriageway.) Probably also the part between Nong Chik and Pattani, since it carries AH18.

PS I’m not familiar with the Lop Buri - Chai Nat section of Hwy 1, but Nitinatsangsit’s suggestion does seem reasonable.

There’s a lot of construction being done on road 4 between Chumphon and Phuket. The section from Phuket to Khao Lak was almost complete last December. In Khao Lak, one could see that the road was soon to be built to dual carriage or 4 lanes.
North of Trang, road 403 might perhaps be the better choice than road 4 - but I do not know that section closer.

News reports indicate the work is complete up to Takua Pa, the section from Chumphon to Ranong is expected to be finished next year, and the remaining section between Takua Pa and Ranong to commence after. I think we can go ahead and make these sections trunk. Hwy 403 does seem like it would supersede Hwy 4 for most traffic. As for the others…

Three-digit highways that are currently tagged highway=trunk:

• 117 (partial) from Nakhon Sawan to Phitsanulok bypass (126)
• 122 (all) Nakhon Sawan bypass (Hwy 1 still trunk)
• 125 (all) Sukhothai bypass (superseding 12)
• 126 (partial) Phitsanulok northern bypass (superseding 12)
• 130 (all) Mae Sot bypass (Hwy 1 still trunk)

• 204 (all) Nakhon Ratchasima bypass (Hwy 2 still trunk)
• 212 (partial) from Nakhon Phanom to That Phanom (unconnected island)
• 232 (partial) Roi Et ring road southern section (hanging end, unless 23 is upgraded)
• 238 (partial) Mukdahan bypass (hanging ends at 212)
• 290 (all) Nakhon Ratchasima southern bypass (parts still under construction, 303 still trunk, hanging end at 224)
• 291 (all) Maha Sarakham bypass (hanging end at 213)

• 302, 338 - to be discussed as part of Greater Bangkok
• 304 (partial) from Chachoengsao to Nakhon Ratchasima
• 314 (partial) from Bang Pakong to Chachoengsao (missing the early part where it meets Hwy 34)
• 317 (partial) from Chanthaburi to Khao Khitchakut (hanging end)
• 331 (partial) from Khao Hin Son (304) to Si Racha
• 340 (all) from Bang Buathong through Suphan Buri to Chai Nat
• 344 (all) from Chonburi to Klaeng
• 347 (partial) from Pathum Thani to Bang Pahan
• 359 (all) from Khlong Yang (304) to Khao Hin Son (33)
• 361 (all) Chonburi bypass (supersedes 3)
• 362 (all) Saraburi bypass (supersedes 1, 2)
• 364 (all) Rayong bypass (supersedes 3)
• 365 (partial) Chachoengsao bypass (304 also Trunk)

• 401 (partial) several segments near Surat Thani (hanging end towards Phang-nga), and from there to Nakhon Si Thammarat (also a hanging end)
• 402 (all) to Phuket
• 403 (partial) from Nakhon Si Thammarat to Kantang, minus a small segment inside Trang ring road (419) (hanging ends at Kantang, and Thung Song)
• 404 (partial) from Trang to Yan Ta Khao (hanging end)
• 407 (all) Hat Yai to Songkhla (hanging end past 408)
• 408 (partial) from Nakhon Si Thammarat to Songkhla (missing segment between Ranot and Sathing Phra)
• 409 (partial) only the segment near Yala (hanging end)
• 414 (all) Hat Yai bypass (407 still trunk)
• 417 (all) to Surat Thani
• 419 (all) Trang ring road (supersedes 4, 403)
• 418 (all) Pattani to Yala
• 420 (all) Surat Thani ring road (supersedes 401)

Four digits and rural roads:

• 3298 (all) short link at Bang Pahan
• 3560 (all) short link at Saraburi
• 4008 and 4153 (all) to Tha Kham and Surat Thani
• 4302 (all) Sarasin Bridge outbound
• 4311 (all) to Phang Nga (supersedes 4)
• 4303 (all) Khok Kloi bypass
• 4103 (partial) Nakhon Si Thammarat bypass (401 still trunk)
• 4287 (partial) short segments near Hat Yai (hanging end and island)
• 4054 (partial) from Saodao (bypassed by สข.1008) to Padang Besar border
• นบ.1020 and นบ.3021 - discuss with Greater Bangkok
• ชบ.1073 - Motorway 7 to Bang Saen
• สก.2089 - Aranyaprathet bypass

For bypasses and ring roads, when should the bypassed through route be or not be downgraded? Currently, Nakhon Sawan, Mae Sot and Nakhon Ratchasima are not. I’d mentioned Nakhon Ratchasima above, but am not sure about the rest. (Conversely, there’s also Saraburi and Aranyaprathet, mentioned above.)

The bypasses that leave unconnected ends should be trimmed. (238, 290, 291) I’d also downgrade the end of 401 near Nakhon Si Thammarat and 403 at Thung Song.

The Isan region is generally underconnected. At least, Sakon Nakhon is a city municipality and should be linked by a trunk, so make Hwy 22 trunk, either to Sakon Nakhon or all the way to Nakhon Phanom. Either downgrade 212 to primary, or make it trunk from Nakhon Phanom to Ubon Ratchathani.

I know 304 and 344 are major routes in the East, but I’m not sure about the rest. Should 304 be trunk from where it crosses the Bangkok Ring Road (Motorway 9)? If not, the loop through Chachoengsao should be downgraded. North of Bangkok, as mentioned above, I think 347 should be downgraded.

In the South, I suggest getting rid of the hanging western end of 401 (at least until it and 415 become the main route to Phuket). Is the 4153-4008-401 route to Surat Thani as prominent as 417? If not, I’d say it’s primary at most. 404 south of Trang doesn’t quite lead anywhere, and is Kantang major enough to warrant 403 being trunk here? 408 should be either trunk or primary all the way (though I remember it as more of a scenic route than a major thoroughfare).

And finally, the missing end of 314 needs fixing, and I think the stubs that lead nowhere (317, 409 & 4287) can be safely downgraded.

Can anyone confirm if the new section of Hwy 12 from Kalasin to the Thai Border, is now fully open ? I drew the small section “under” Kalasin as Primary when it was single carriage, but I know they were building a second carriage back in 2019. Where bridges were not finished, I left sections as construction.
Once all open, I’ll change the entire length to trunk, but do need confirmation as I won’t be going back there for some time :frowning:

I agree with Paul. Moreover, Hwy 1 from Junction to Km. 111 should be primary, not secondary, to show that it’s importance is just lower than the bypass road, not as poor as secondary.

I’m not sure about this, since most part of it is packed with frontage road, u-turn bridge, and only 3 traffic light. But I think primary is also acceptable, like main street in Bangkok.

Agree. From Ranong to Khok Kloi, I think it should be primary, since most of it is only 2 lanes. From Chumphon to Ranong I can accept both primary or trunk.

Bypass around Sadao is look like just normal road run around the town (4-lane without island), not that type of bypass road for travel through traffic. This bypass road shouldn’t be higher than secondary. And for sure, Hwy 4 through the town should not be downgraded.

I think it shouldn’t be trunk, since it is only short distance. From Lom Kao to Loei is mostly 2 lanes. Moreover, around Lom Sak, Hwy 2466 from South Bypass Road Int. to Pho Khun Pha Mueang Int. should be upgraded to trunk, since it’s performance is not lower than bypass road. Also, bypass road from West Bypass Road Int. to North Bypass Road Int. should be downgraded to primary.

Agreed, since Hwy 23 and 24 is under construction to be 4 lanes all the way, only few part is not finished yet. Moreover I would suggest upgrade part of Hwy 2085 then 2178 from Ban Chan Int. to Ubon Ratchathani to primary, since it is on plan to upgrade to 4 lanes. It is the main road that people use to travel from Bangkok to Ubon Ratchathani. Another is Hwy 2050, main long-distance road that connect Ubon Ratchathani and Khemmarat, to be primary too, because most part is already 4 lanes with wide painted island, and little part is under construction.

For Prachin Buri, new way for Hwy 33 around Prachin Buri is under construction. I think we can make the old 2-lane road to be trunk, until the new way construction is finished.
For Aranyaprathet, I’m ok with primary through the city, but not secondary as it still be for half of it.

Agreed both.

If the through city section is still very wide road, with little traffic light, I vote for not be downgraded.

Agreed all.

For Hwy 22, I absolutely agree to upgraded to trunk from Udon Thani to Nakhon Phanom. Because most part of it is already 4 lanes with physical island. A few part is under construction to make them 4 lanes all the way.
For Hwy 212, I strongly vote for upgraded to trunk, since all the way from Nakhon Phanom to Ubon Ratchathani is already 4 lanes with physical island.

For me, I suggest no, because in Bangkok between Motorway 9 and Min Buri the road is very busy.

I’m not sure about this since Hwy 347 is very high performance highway with only 2 traffic light along the way (from Pathum Thani to Bang Pahan Int.). But in compare to Hwy 32, maybe downgrade Hwy 347 is more useful for route selection, although the performance of these 2 highway is quite similar.

Agreed.

It’s quite high performance. But in compared to Hwy 417, maybe downgrade to primary is more useful.

I vote for downgrade 404 to primary. For 403, I’m not sure since Kantang is middle size town.

Since it’s 4 lanes with physical island all the way already and very few traffic light, I vote for trunk.

Yes!

Yes, according to this news https://www.prachachat.net/property/news-396635

Ps. I would recommend looking at https://gis.drr.go.th/drr/Map2.aspx
National Highway with 3 or more lanes is showed yellow (most of it is 4 or more lanes). National Highway with 2 lanes is showed red.
Although this is not that very recently updated, it’s useful for look at overall network of highway to decide about tagging them.

I do not understand that. What do you mean - keep it as primary, or change that to trunk?
That western end of 401 leads through a “nowhere” area near Khao Sok national park towards Khao Lak. It has some importance, but not too much. Performance is not bad either, except when the hills west of Khao Sok village are crossed. A good 2 lanes highway. Hence I’d prefer primary.

Also in other countries, trunk roads lead to minor places or become primary somewhere. E.g. just near my home road B42: https://www.openstreetmap.org/way/319798337

Road 404 joins road 416, leading to Satun. 416 is a 4 lanes highway in some places.

Ah, sorry. I was writing those comments while looking at the current tagging on the map. The references to dangling ends leading nowhere were about roads being tagged as trunk only partially and abruptly changing to primary partway through.

Let’s take a more radical view: do not mention any administrative features when explaining our tagging scheme! Get rid of any numbers and digit counting.
Though the highway administration often follows our criteria (role in the road network, importance, performance, …), it has also some other points on their agenda. History: all those formerly major road which connected Bangkok with the provinces still have a low number (and most of them are still important, but not all, or over their total length). And financial aspects: it determines the administrative body which has to pay for the road.
Neither history nor financial aspects are relevant criteria for OSM highway tagging.
After all, I think we just remove those administrative hints from our current description.
We may add some chapter on the administrative numbering scheme, but we must not mention there any relation to our tagging scheme.
Thus we can get to a tagging scheme which describes what the road is, instead of a mere administrative number. We could thus get rid of these always returning discussions which focus on administrative numbers despite clear evidence of a different role of the road.

Totally disagree, Bernhard… so we recently had a hospital access road tagged as tertiary, because the user thought it was “important”… and Kaart retagged a load of tiny CM Sois as tertiary, because they though they were “important”… I am even arguing with Dave S. right now over where the Hwy 106 should change from its Primary tagging as it enters Chiang Mai…but as the road winds down from 4 lane to two lane, Dave thinks it should remain primary. I’m not knocking Dave but he thinks its an “important road”, I don’t… it shows how opinions can differ, and we are probably some of the most seasoned Thai mappers about.
All because each User has a different take on “importance”, and sure as heck, any description of a road that involves a judgement call, is going to result in huge vagaries across the Thai map, not to mention edit wars.
I think the idea of continued tagging based on the Hwy number, but with voted on/documented exceptions, was the way forward. Has that idea been abandoned?

And as Russ indicated Bernhard, I think your suggestion makes a lot of sense. I was going to raise Russ’s demotion of the Chiang Mai-Lamphun road (R 106) as a separate topic but it falls into this discussion perfectly. Russ changed the classification of this major north-south highway from primary to tertiary based on his reading of an administrative database that indicates that the end of the route, R106 that is, is at a certain mile marker. The road does not change from 4-lane to 2-lane at that point, as he asserts, but continues as a heavily used 2-lane road all the way to its intersection with Mahidol Road (R1141) and beyond. His reasoning makes no sense to me because it is the same road in almost every characteristic before his break-point and after. It’s a break point that is artificial and based purely on administrative considerations rather than reality.

Here is our changeset commentary about R106. https://www.openstreetmap.org/changeset/77188198

I don’t always take that side of this argument. It all depends on the reality on the ground. I had a discussion with some OSM mappers who raised the classification of many Alaskan highways from primary to trunk based on the Wiki’s definition of what constitutes a trunk highway in terms of its importance rather than the physical layout of the highways. Those highways are normal 2-lane highways with no limited acess whatsoever. There are RR crossings, driveways and residential intersections all along it yet because they are major connectors, usually the ONLY connectors, between cities they made them trunk highways. I disagree with that reasoning but never pushed the issue because the arguments get intense and people defend their own ideas regardless of what you might say or think.

Returning to my disagreement with Russ: Each user may indeed have a different take on importance. Like so many other things, smoothness and tracktype come to mind right away, these are judgment calls and we as experienced mappers make those decisions all the time. We also expect those decisions to be realistic and based primarily on the Wiki definition. However, the definitions in the Wiki aren’t always clear and cannot possinbly cover every situation so we are forced to make some decisions based on our feelings or inclinations.

You can quote me on this: “There will never be a Wiki definiton that is perfect or even close to perfect. The judgement call will always be a part of our mapping scenario. Human decison making will never be perfect either and there will always be aruguments about important topics.”

Dave

Well, no problem for me…
Since I render the maps for my use myself, I can use some tricks. The South is my main area for holidays, that means far less roads than North or Isan, so I can handle an inofficial extra tagging there (just in case some one else among you stumbles upon the “hwpi” tags - feel free to change the “highway” tags however you wish, but only don’t touch “hwpi”).

I tend to agree that for some roads the classification simply by digits does not work well.
It does work probably well in telling the “classification”, but in practical life we want the highways rendered prominently on a map based on importance. For the similar reason we want them to be preferred when routing on OSM data.

Right now, for a lot of data consumers (mainly map rendering or routing), the road classification gives them some reasonable defaults.
You can probably assume that a highway=secondary is paved. Or that a highway=primary has better driving conditions than a tertiary (like number of lanes, overpass instead of traffic lights or such).

We have to maintain a robust set of critera on how we can tag and make this tagging verifiable by other mappers. It is like in science. If you claim something there must be a possibility to prove you right or wrong.

So there must be criteria that if I claim a road is a primary, then others on the ground must be able to verify that. And in case road-work happened maybe a road can be re-classified based on this criteria.

The classification by digits served well for an initial classification.

And I think we discussed it before that there might be defined criteria which allow “upgrades” of roads in a higher category. A four-lane road with shoulder and a middle-divider might be upgraded from tertiary into a secondary.

Maybe we can also come up with some criteria which allows to downgrade roads. Or hard criteria which will prevent a specific classification.

I for example would not want to see any unpaved road in Thailand with a classification higher than tertiary.
This is a criteria which can be easily verified by others on the ground.

Another criteria which is verifiable might be the intensity of usage of a road.

Taking the example from Russ: So technically (based of DRR database) a section of a road might be tertiary. But based on the intensity of usage, it could still be used as a primary road. Number of vehicles per hour might be a criteria.
By counting vehicles on the ground for half an hour you should be able to get a good idea how “intense” a road is used.
So in this case it would “protect” a road section from a downgrade to create a more uniform grid of the road network.

I remember also a tertiary road (according to DRR database) in greater Bangkok which was just an agriculture track. Maybe planned and earmarked already for future expansion, but not yet built.
For such cases I would like to have some objective criteria which allows to not tag it as tertiary.

The classification should give a hint, but as government databases can be wrong as well, the authority should be the ground-truth by mappers.

But not based on “opinion, impression” or other subjective criteria. I saw edit-wars where people wanted to do poor-man’s traffic management and reduce via-traffic through their quarter by messing around with the database in the hope that routing engines would prefer other ways.

So what kind of “minimum” requirements would we have for each specific road type (unclassified, tertiary, secondary, primary, trunk)? surface, width, number of lanes, other factors? Are there “de-railers” which prevent a category? What could allow up/downgrades? What could allow a road to keep its classification a bit longer?

@Bernhard: I am no fan of “private” tags which are neither documented nor verifiable by others and OSM is just used as a cheap way to store private data. Same reason we don not want to have other “internal” identifiers stored in OSM data.

In you case: I believe you tried to mark road categories on a different classification, outside of the OSM scheme, right? So could you document it? No real need to keep it private. Why not trying to run it in parallel with the existing tagging. If we see later a benefit, we can switch over. Otherwise it is easy to spot and remove.

I use some tags for overriding the highway tag during the map creation process.
Long ago, it was about dual carriage highways, when I added a “dual_carriage=yes” to those highways (I do not add it anymore).
When the discussion about reference-number-based mapping started again in april, I decided that I need to protect the tagging which I did during the last years, and invented “hwpi” (think of “highway performance or importance”). Thus we can avoid edit wars.

The rendering rules are:

  • “facebook”: import=yes and highway!=track => highway=road
  • “dual carriage”: dual_carriage=yes => highway=trunk
  • “hwpi” depending on value
    0 => trunk
    1 => primary
    2 => secondary
    3 => tertiary
    4 => unclassified
    5 => residential
    6 => service
    7 => track

When I add those tags, I actually follow common standards for highway classification - even “our” Thai wiki with the exception of the reference numbers.
Preferentially, it is based on ground truth. But unfortunately, most of the roads have to be classified from imagery. Experience with the ground truth both from Thailand and from well-mapped countries elsewhere helps alot. I expect some misclassifications - but far less than now.

You have put me in a bit of a quandary with this information, Bernhard. I don’t think this sort of tagging is legitimate. Sure, one can create any tag one wants on OSM but it’s a two-edged sword. If everybody felt free to create just any tag they want based on his or her personal needs, the database would soon fill up with millions of undecipherable tags, tags like"hwpi". Just imagine if all of the top 10 mappers in Thailand were to add just two or three of their pet tags. That would imediately add 20 or 30 new “mystery tags” to OSM. Clearly, that is not a workable situation.

These special tags should be removed IMO. I’m sorry to have to say that to such a dedicated Thailand mapper but I’m afraid that’s how I feel about it. If the current tagging scheme isn’t working well, then let’s work together to improve it.

Respectfully,

Dave

I prefer to think of that tagging as something temporary, nothing permanent.
You can view it as an experiment: we can later compare maps based on the different styles, and decide then which one is better.