Just how do we tackle cavalier mapping ?

Just noticed a number of 3 digit Highways north of BKK have had their status changed from Primary, to Trunk… had it come from a regular Thai contributor with experience to their credit, you bow to their judgement…
but when its down to an anonymous newbie, with less than 200 edits in 4 months… shall we insist on a total reverts by user name ? Maybe its just the lack of respect for the existing Thai group that irritates.
Maybe its the fact its a lazy edit, where not all U-turns get the correct corresponding tag. https://www.openstreetmap.org/edit#map=19/14.41587/99.97279
Anyway, I just wrote the following comment on one of his changesets, https://www.openstreetmap.org/changeset/82286616#map=15/14.4686/100.0630

I await a response.

Isn’t that an important highway for far-distance high-speed traffic beyond changwat boundaries?

I hadn’t looked at imagery of the highway in question before but now that I have I can see that it might fit someone’s definition of a trunk road. There is disagreement about this topic as I discovered in Alaska where someone changed what I consider primary highways to trunk despite the fact that they were only 2-lane, not limited access and had traffic signals, intersections and even driveways connected to them. The definition had more to do with purpose and the fact that they were the sole connection between population centers than the definition in the Wiki, the proponents said.

An aside: I noticed a tag that I haven’t seen before, dual_carriage=yes. It would seem that tag is totally redundant. One can clearly see it’s a dual carriageway. There’s no definition in the Wiki so what does it mean?

Unfortunately, with OSM, this is one of the problems in that once you define a road by its purpose, it become very subjective. And with a situation like this, if one mapper decides a status tag should be primary, then just because someone subsequently changes it without documented reason, does that make it right ? So why should I not change it back ? It just gets into an “edit wars” situation.

With this Hwy, I see that Bernhard had added his dual-carriage tags, but was happy to leave the road status as is… on that assumption, if a well established mapper didn’t see it fit to change the road status, why should we give credence to a newbie who cant be bothered to introduce himself to us?

And this raises one of the fundamental issues in Thailand mapping … if we go by purpose (and even with a description), its becomes very subjective, and even more so if you are armchair mapping. If we instigate guidelines that follow the Government road network classification, it makes the map consistent, and prevents arguments over interpretation of purpose.

I dont know of any other map that is commercially produced, where they classify roads by purpose, and certainly none where sections of the same Highway number, change status within their length.

Myself and Bernhard are totally opposed in these views, and until we reach a consensus in our loosely defined group of Thai mappers, it will never change. Dave … if yourself, Stephan, Mishari, Johnny (and any other old timers I have missed) agree with me, then we should change the Wiki to reflect this.

So here are the relevant pieces from the Wiki that are germane to our discussion:

highway=trunk The most important roads in a country’s system that aren’t motorways. (Need not necessarily be a divided highway.)

highway=primary The next most important roads (after trunk) in a country’s system. (Often link larger towns.)

Use highway=trunk for high performance or high importance roads that don’t meet the requirement for motorway. In different countries, either performance or importance is used as the defining criterion for trunk

The wording in the Thailand Wiki is slightly different but IMO boils down to the same thing

“Trunk classification should be applied to roads that look like motorways and are geared towards fast far-distance traffic, but are toll-free and with unrestricted access; even bicycles and pedestrians may use them.”

By these definitions, the highway mentioned in the original post could be tagged as a trunk highway. Due to the way the table of highways types is laid out in the Thailand Wiki, it’s difficult to determine exactly where the example 3-digit highway ref lies. But based on that table I believe one can make the legitimate assumption that a 3-digit highway can legally be tagged as a trunk highway as long as it meets the other criteria.

So, unless we want to rewrite the Thai Wiki to make it clearer or possibly more restrictive, I’d have to say the way the newbie did it is not incorrect. Interpreting the wording of Wiki entries is always a tough prospect. We have people for whom English is a second language, and speaking frankly, people who don’t know how to write a clear technical description, creating these entries. It’s a recipe for confusion.

Dave

Exactly Dave … which is why I’m looking for support to make the Wiki far clearer/less ambiguous … but where does that come from?
It was sort of “hallowed ground” until Kaart hijacked it… what if I changed it ? Is that Kosher ?

I add them so that I can render such roads like trunk. Please do not remove them, this tag does not interfere with other people’s number classification, and still helps me to draw a useful map. I am fed up with ideological discussions, I look for workable solutions.

Take a look back into long gone history, e.g. https://forum.openstreetmap.org/viewtopic.php?id=22577
If I remember correctly, the posts of RocketMan were about lesser (tertiary/secondary) roads.

Changes come from users like you who find the Wiki topic confusing and/or incorrect. Working with a group helps generate support for the changes and may also help with the wording of those additions or changes to make them clearer and hence easier to use. I don’t think of the Wiki as hallowed ground at all.

That said, one problem potential Wiki editors will come up against is the difficulty of editing those entries using the built-in Wiki Editor, especially when tables are involved. It’s such a beast and its format so arcane that I hesitate to involve myself with such projects. I’m suspicious that many people wishing to clarify or add to the Wiki are enough put off by it to just go ahead and implement their changes without updating the Wiki. I used to be a computer programmer and have even hand-built HTML tables in code but I find the Wiki editor to be such a PITA that I fall into that category myself. :frowning:

By the way, I looked at the thread from 2013 that Bernard refers to in his reply and it is valuable background for this discussion.

While primary, secondary, tertiary is catagorized mainly by importance.
The trunk road, which stated at https://wiki.openstreetmap.org/wiki/Tag:highway=trunk as “high performance or high importance roads that don’t meet the requirement for motorway”, should rely more on appearance, like motorway.
What is the criteria for trunk? I think https://wiki.openstreetmap.org/wiki/WikiProject_Thailand is stated clearly. Also, I would like to add some words to make things more obvious like: wide shoulder or frontage road available, has an island or ditch with a few u-turn point and not like a painted island.
For these kind of criteria, the edited section of Highway 33 is clearly not trunk road. While Highway 321 is still debatable.

Yes, what did happen to RocketMan ??

Anyway, we can discuss a lot, but the underlying argument I have, is that the definition in our Thai Wiki of Trunk and Primary, is so close and blurred, its down to “interpretation of importance” !! So when I revert Hwy 321 back to Primary, and a few weeks later its changed back, and then I’m mad so change it again … how do we stop these childish antics ?
And then if you make it too specific, we result in a map which looks like a “patchwork quilt of colours”, which I know goes against OSM principles. After all, I dont think its right to change the main East-West Hwy 12, to Tertiary for the 80 or so km, it goes through a National Park, but yet its only a 2-lane single carriageway road here. It certainly does not meet many of our Trunk road criteria, but yet it is Trunk … and it would be crazy to change this section. Fortunately either common sense prevails, or our adherence to the digit principle works here

Now if you examine the problem that started this at Suphanburi -https://www.openstreetmap.org/#map=12/14.5208/100.0460
The Hwy 321 that comes in from South has been made Trunk, so is the Hwy 340 that comes in from the North. But unless your destination was the city, and you were heading North-South (or vice versa), you hope your map/routing would use the Hwy 357 bypass, to save going through the longer and considerably more Traffic Lighted (!!.. made up word) 321-340 trunks.
But yet the 357 bypass is still only primary, I guess because its a two lane d/c and not three. Maybe Mr Oum forgot, didn’t understand, didn’t know … but its cavalier and makes no sense. In fairness it does not cross Changwat boundaries, so maybe that’s the reason.

We need a system that takes away this personal interpretation of status, and I for one, will “interpret” the roads to match the digits (1&2=trunk, 3=primary,4=Secondary, & RR provincial roads=tertiary). Its served the North of Thailand well for years, and I don’t see that I should just bend over and allow a mapper who wont even communicate with us, the right to second guess me.

Nuff said !

I think that maybe the problem is come from the poor highway category system from Thailand’s government. For example, the 1-digit number is not the highest performance, but just a “first” road that run from Bangkok to the province, since 70-80 years ago. Some part of them is worse than the lower ranking road. The recent constructed road by The Department of Highway is mostly 3 or 4 digit, with many of them is better choice for traveling than the older road. And for the new road by The Department of Rural Roads, all of them can be only 4 digit with province prefix, but some of them have very high performance. For some extremely example, the Rural Road นบ.3021 (Ratchaphruek Road) is a better road than the nearby Highway 9. นบ.3021 is running from Bangkok through Nonthaburi to Pathum Thani with no traffic light at all, no u-turn point at the island at all, and frontage road is available for more than half of the length. Another example is, for Highway 1, the section from Lop Buri to Ta Khli might should not be tagged as trunk, as most part of it is only just 2 lanes or 4 lanes with painted island. Moreover, it is not a best way people choosing for travel from Lop Buri to Ta Khli.

If the main purpose of categorizing the highway is for route selection for navigation, then I think we should give weight to “importance for traveling” more than “importance for administration”. The only problem is “importance for travelling” might be hard to find. Maybe we have to develop a more specific criteria for such a thing. Since the importance of the road is not an absolute value but relative, so we maybe have to compared the road we’re working on with other roads nearby or the similar roads run to the nearby area, then ordered for route selection.

I know that by this way maybe more conflict will happen, but at least it fits more for using the map for navigation.

Regards

I’m adding a new twist to this thread because what I have to say falls under the topic of cavalier mapping IMO.

I’ve run across many instances lately of improper mapping in northern Thailand by new user:mCloud00. I’ve made some changeset comments but have not received anything back from him or her yet. You can see some of them here:

http://resultmaps.neis-one.org/osm-discussions?c=Thailand&d=90#5/12.805/101.846

This person isn’t doing things because he’s nasty but because he’s inexperienced and isn’t reading the Wiki or can’t read it. My first experience with mCloud00 is when I saw a bunch of “fords” in my local area where a stream or other waterway crossed a highway but the original mapper for whatever reason hadn’t also added a bridge or culvert. I’ve removed those in places where I positively know it’s wrong.

Then yesterday while mapping wood multipolygons I noticed many areas tagged natural=wood that mCloud00 had added as outers to nearby multipolygons even though they shared no ways with them - they weren’t connected on the ground. I suppose that’s not absolutely incorrect but it’s not logical and adds unnecessary complications. mCloud00’s other work is also problematic. Traces are drawn haphazardly, etc., and I’ve cleaned up several long untagged ways he left behind. One of them was part of a complex boundary - waterway relation that was wrecked because he broke the tagging for the Ping River riverbank that was part of it. It is quite obvious he is in over his head but he tried to fiddle with a boundary relation despite his obvious inexperience. Relations are complicated but he seems hellbent to work with them despite that fact.

So, what to do? I call your attention to this just to put you on alert but beyond that I don’t have an answer. It’s too bad this person doesn’t try to learn by following examples and reading the Wiki. But then again, perhaps this mapper cannot read the Wiki.

Dave

I would unequivocally ask the DWG to block him with immediate effect, until he communicates with us, showing he understands the mistakes. We are a community. If you cant understand the “instructions”, then there is no place for you.
Im afraid I had cause to send him a message a few weeks ago…

He did acknowledge me with a brief “ok, but should I add residental area back even if I mapped all of the build in that area

My response …

And there it ended.
So, Dave, I suggest you send him a personal message, as at least I got a basic reply using that method. But if nothing, I would be quick to ask for a block, with my full support.
Russ.

Hi Russ, I did send a PM through the OSM system on May 24th:

I really don’t want to have him blocked unless as a last resort. He’s new and inexperienced and maybe doesn’t know how to ask for help. But I agree with you that his tecnique is sloppy. Yet, he has energy and apparently time on his hands because he’s added a lot of stuff to OSM during the time he’s been involved.

OTOH, it’s been almost a week since I sent that message and so far I’ve heard nothing from him.

Anybody else have a suggestion or approach to use with this fellow?

mCloud00 appears to still be actively editing; has their work improved, or are there still issues? I don’t think they’re going to respond without more direct action.

Regarding highway classification, it’s been seven years since the last revision; I think we can re-visit and rethink the scheme if it’s proven to be inadequate. Maybe a separate thread dedicated to the issue would be more conducive to discussing this?

Yes, he is actively editing and when I say active I mean he’s been very busy. His work quality is borderline. It’s not very accurate and not following guidelines as laid out in the Wiki or by locl convention. He resolved one of my Notes a week or so ago, which was fine, but he has not replied to my changeset comments or my PM.

If someone could just sit down with him and offer some guidance, maybe that would work. But he doesn’t respond to those standard methods so perhaps he needs to be blocked in order to get his attention.

Take a look at this area, in particular the “gardens” he’s added. Check them out with satellite imagery. He has included buildings, orchards and other stuff in those areas. Very casual. And then of course, none of the areas he’s mapped meet the OSM definition of a garden.

https://www.openstreetmap.org/way/786189402/history#map=18/18.68164/98.99092

Some sort of direct action is required IMO.

I’ve emailed the DWG asking that someone look at this thread.

Thanks, Paul. I just checked the stats and saw that mcloud00 is now the #3 mapper in Thailand. He’s been very busy indeed. Best to get his attention ASAP before he makes it impossible to clean up his work.

After I posted this reply, I left this comment on the changeset I referenced in my earlier post:

I’ve received a reply as follows:

A quick note before shutting down for the night:
mCloud00 has responded in brief to all my changeset comments and has indicated his intention to correct the errors I pointed out in those comments. He’s made some edits already but it’s too late to check on his work tonight. I’ll try to take a look tomorrow and report my findings here. I guess the block got his attention.

Dave

Update: This morning, I checked several of those changesets but he hasn’t changed anything as yet. I started deleting “gardens” and removed a couple of incorrect “ford=yes” tags but then decided instead to add advice to those comments and to give him more time to respond. Because he said in one comment reply that he lives in the area and speaks English, I offered to meet with him to discuss OSM mapping. I believe he’s excited to be mapping and he obviously has time on his hands, therefore IMO, he could be a good asset if he understood more about how things should be done.

I’ll let you know.