Hi MCDA, thanks for the query. This was part of some large-scale maxspeed updates over the past few months to improve routing & for any other data applications - I've tried my best to minimise errors, but I accept there may be some.

Regarding street lights however, I researched this as much as possible (there is not much documened online, and the legislation is somewhat open to interpretation). The clearest source I could find was this [2009 discussion on the RAC forums](https://forum.rac.co.uk/threads/9339-When-does-an-unrestricted-road-become-restricted/page2) with information purported to be from the DfT and Thames Valley Police, stating their shared interpretation that a truly "restricted" road (i.e. 30 limit denoted by presence of street lights) can only be enforced if it has a 30 limit sign when you enter the section of road. Thus, when in the zone, no repeater signs are required because the street lights denote the restricted nature of the road. Meanwhile, if you enter a stretch of road with street lights and no indication of entering a restricted (30) limit, this is not considered to be a restricted road and therefore retains the limit last signposted (e.g. NSL).

[Incidentally, I'm not aware of any mention in legislation of a implied 40 limit as per your message, only interpretations of an implied 30 limit.]

In practice, if traffic were to slow down to 30 at any section with street lights (e.g. around junctions on the dual carriageway A4) it would actually be dangerous to them and other road users. The other practical issue I can think of is where drivers use a "black box" speed monitor to keep their insurance premiums low. If they were to drive along a NSL single carriageway at 60 mph, pass a few streetlights at a junction where the road has been marked on OSM as a 30 limit, due to only the presence of lamp posts, they will be unfairly penalised when this is processed as a transgression.

General opinion is that the legislation around speed limits is a bit of a mess, but the DfT and TVP intepretation appears to provide fairly clear and implementable standard. Happy to discuss here or somewhere more public, if this doesn't seem like the right interpretation :-)

Oops indeed! My bad, thanks for spotting 😊

I must have been splitting ways at the roundabout while I was at it, left out the step of adding relations back in. Fixed now.

Thanks for those, I may get to them at some point 😊

Even a short distance NW of this edit there's a long section of the Cavan Road which should just be a single carriageway, with traffic islands mapped as nodes. The approach I take is as documented on the OSM wiki (osm.wiki/Dual_carriageway) - if there is no physical barrier separating the carriageways, it's not a dual carriageway and should be mapped as a single way. As a result the Cavan Road has driveways linked to only one side of the road, but you can guarantee the road design is not meant to be left-in left-out only for those access ways. There are occasional exceptions to what's in the wiki, but roads like these are pretty clear cut (see also examples on the wiki page).

It does take some unpicking of relations though (bus routes and the like), so needs some care. I also check which way is the oldest of the two, to retain as much edit history as possible. Give it a try some time – or otherwise I may get round to it.

*Typo, gat=that

Hi vas111, yes it's technically not required (except in some instances, e.g. where other transport methods do not treat the road as one way), but it doesn't do any harm. I tend to add the forward/backward tags habitually as I'm working on other items nearby, as it means I'm always applying the same standard whether it's on one or two-way roads, and with the more consistent view gat results, I'm able to more easily spot any errors I've made. This changeset wasn't specifically to add directions to traffic lights, but to ensure they were mapped on respective ways rather than on both nodes at the intersection, which results in any routing across the junction on the single carriageway "seeing" two sets of traffic lights instead of one. But I did add the directions as a by-the-by.

Hi vas111, thanks for the query.

For intersections between dual and single carriageway I work off the recommendations on the (now archived it seems) Mapbox intersection modelling page (https://web.archive.org/web/20230602212339/https://labs.mapbox.com/mapping/mapping-for-navigation/modeling-intersections-for-map-navigation/). This dictates that a split in the carriageway should not occur at the point of intersection, as it can confuse routing software and produce unpredictable results. The recommendation in this instance would be to either continue the separate ways of Market Street to the north side of the junction and join them into a single carriageway post- intersection, or alternatively have the intersection occur between a single and a dual carriageway, by combining the Market Street carriageways on the south side. I opted for the latter, as there is a natural space on the south side of the intersection for this to occur, allowing the mapped "theoretical" ways to better represent real life, versus the former.

If a further U-turn relation is required then that's not too big an issue - it is after all just a rule that will block U-turn routing from occuring. But the recommended intersection layout has a big effect on compatibility with different routing software, so I'd argue it's the more important need here. I should add though that the Aberdeen roads are mapped to a high standard, so great work if you've been involved in this 😊

Hi GinaroZ, no not specifically, apologies if I've used the wrong designation. I was going by the fact that it is used by both pedestrians and traffic traversing the same road (i.e. pedestrians are not separated onto walkways) so is a living street at least in usage/spirit, if not in legal designation. Service road seemed unlikely as it was built at the same time as nearby residential roads, with similar usage and to similar proportions - perhaps the best solution is Residential with a surface tag to denote a change from asphalt?

Hi midnight2024, I see you have altered a large section of the A92 in Arbroath to highway=construction, I assume in relation to note [4183658](note/4183658). Are you sure there is absolutely no access from one side of the town to the other, and the Guthrie Port Roundabout is entirely closed to traffic? As the changes you have made will currently stop GPS software routing road users along any of these roads. I have a hunch this needs to be mapped differently to take into account what access routes are still open on the ground, to ensure GPS road users are not affected.

Ah! Well spotted, thank you 😊 Yes I must have missed this in changeset/123379950. That road was definitely opened, I can confirm from survey (although not a recent survey). Please do remove the access restriction.

It looks like there are some minor name capitalisation issues in the adjacent ways which could be fixed at the same time, if you don't mind doing so.

Hi borovac,

That was quite a large changeset at the time, and I can't see any ways that are still mapped as closed. Which particular way/road are you referring to?

Yeah, you're right - I think I was erroneously swayed by 3D rendering, which defaults to displaying as roof:levels=1 unless otherwise specified, so was tagging relative to that. I'll change to 0.5 and add a roof:height tag.

Oh, wow! That's a failure of pretty epic proportions, especially considering the timescales. Thanks for the background info!

I've updated based on those tweeted images (and another which shows updated routing of the crossing to the NW). I've also added notes to explain why the map doesn't match what's on the ground, and that the way is "due" to be relined...

Thanks for bringing it to my attention!

(It seems strange I know - but it reflects what is on the ground.)

Hi, thanks for checking. Yes, it is meant to be one way - there is painted text on the cycle path at this point that says "no entry" in eastbound direction, and eastbound cyclists are forced to take the crossing. Painted lines and arrows reinforce this. Routing should direct cyclists onto the opposite carriageway.

Ugh, forgive poor typing on a phone - "the previous routing"

I should add, tge previous routine of the way was incorrect, and made it look more like the cycleway would have continued from Lampton Park straight through to the Great West Road - but as you'll see on updated layout, it's not the case and there is a dogleg junction instead. I think it most likely that the designated cycle route continues west to Willow Gardens and does not dogleg up this footpath, while it may have looked more likely with the old mapped layout.

I haven't, and from your thread you've done plenty of digging in other council areas so well done! Street-level imagery shows what may be byelaw signage at the point this path meets the Great West Road, which might shed some light - but I no longer live in London so can't check. I suppose I was taking a different approach of mapping only affirmative signage, as the cycleway to the south and another on the east side of Lampton School are signed as shared cycleways, while this is not, and the omission leads me to mark this as foot designated only. Perhaps the most pragmatic option is to add a fixme tag asking someone to survey the path and check smallprint signage at North end to confirm whether bicycles are allowed?

Apols for typos in last sentence, I mean bicycle=yes and bicycle=permissive of course.

Thanks for the link - an interesting read (and irritating for mappers and the public, naturally! Councils' ambiguous application of signage, standards and byelaws certainly makes a mess of the situation...)

My main concern on that particular way is that it was set for cycles only with no provision for foot traffic, when the exact opposite appears to be the case on the ground. Historically it was always a footpath (hence changing to footway), and on a recent visit I saw no signage designating it for cycles. It also has a rather dangerous blind junction at the south end (incidentally with a path which *is* explicilty marked as a shared cycleway). At the north end the footpath joins with a segregated foot pavement (physically separated from the cycleway), and the path itself isn't in a park, although Lampton Park is in the vicinity.

So I think this one is a bit more clear cut to be honest? Although separately to the south, where the adjacent shared cycleway joins footways in Lampton Park, which I updated with cycle=yes (at least all those that are wide enough and have good sightlines), your method of cycle=permissive could be more appropriate.

There are issues with changing ways from under construction, to constructed with access=no and no highway classification, to fully open with full classification.

1. Some ways will go between these states in quite quick succession - OSM practice with roads (e.g. when an existing road is closed for work) is to only change closure status when that status will remain for around 3 months or more - reason being that it reduces the chances of GPS navigation software updating their maps during a period that the road is briefly in flux, and then retaining that status potentially for the following year or more.
2. There is no established threshold as to whether the road is constructed or not. Is it when the first tar goes down? When the wearing course is added? When lines are painted? How do we know there's nothing still to be constructed like barriers, or if tar needs to be lifted and relaid? IMO the only point that a road is no longer under construction is when opened to traffic. Construction traffic doesn't count, because if it did, we'd be mapping all the site traffic tracks, which would change from day to day anyway - there is no benefit to mapping what contruction traffic can use, as opposed to considering the road still under construction.
3. When other editors come to open the road, they don't necessarily know to remove or convert the access=no, proposed:ref proposed:bicycle etc into normal access/ref tags, so there is a mess of legacy tags that has to be tidied up (I've been doing some recently around N22 Macroom). But if it was left as under construction, the way can contain all of the relevant tags ready to go, so that all that remains is to convert construction tags to highway tags when the road opens.

I'd strongly advocate VictorIE's suggested way of working with road construction.