[Tagging] Two side-of-road parking questions

[Matthew Woehlke]

On 23/07/2020 17.26, Paul Allen wrote:
> On Thu, 23 Jul 2020 at 21:00, Matthew Woehlke wrote:
>> Interesting. By that criteria, I would think that
>> https://www.openstreetmap.org/way/826561593 has on-street parking,
> 
> Tough call.  In isolation it looks like a parking lane, but it has markings
> for car parking.  On-street parking (at least where I am) doesn't usually
> have that.  However, when I switch to Virginia imagery (much clearer)
> I see the parking extends around the corner.

The Virginia Ortho isn't bad, and I sometimes use it as a secondary 
reference, but MapBox is even better. (That doesn't seem to be the case 
everywhere; in some places, MapBox is apparently quite bad, but for much 
of the US east coast it seems to be the highest-resolution imagery 
available. NAIP is the *newest*, but it's lower resolution than many 
other sources.)

> From the geometry, I'd say that was a parking lot.

Currently, I have the non-parallel spots marked as a lot. To my mind, 
parallel parking and on-street parking are nearly synonymous. That said, 
I can see your argument, especially if I look at JOSM's lane rendering 
as a gauge for whether normal traffic flow would be occluded by parked 
vehicles. In that sense, there are a lot of folks parking on e.g. 3rd 
and 4th Avenues (where there are additionally not marked spaces) that I 
would definitely call "on street parking". (I'm also not at all inclined 
to tag that.)

>> but would you argue that all of Potomac Avenue
>> (https://www.openstreetmap.org/way/686513005
> 
> From the fact that parking spaces are marked, it's not a parking
> lane, in my opinion.

Well it's certainly not a parking *lane*; you clearly are not meant to 
possibly drive through it. I was thinking that the fact the parking 
spaces are arranged so as to not occlude traffic was what was inclining 
me to model it as a "lot".

Really, it's the notion of a parking lot for which the aisle is also a 
main road that's throwing me... Maybe I should look at what's been done 
somewhere like Manhattan...

>> Actually, I did mean the six spaces on the west side of Broadway. I
>> recommend looking at e.g.
>> https://www.openstreetmap.org/edit#map=22/38.52057/-77.29185 for all of
>> those, it will be easier to tell what's centered.
> 
> That's a parking lot rather than on-street parking.  At least, that's how
> I'd map it.  If no cars were parked you couldn't drive along it without
> some turns.  If you mapped the highway itself as an area, that
> parking would be a pregnant bulge.

Yeah, that line of thinking is similar in effect to asking if it 
occludes normal traffic flow. Different questions, but likely to have 
the same answer.

> This is how I handled a similar one:
> https://www.openstreetmap.org/?mlat=52.08562&mlon=-4.65829#map=19/52.08562/-4.65829
> 
> Somebody objected that whilst that looked right when rendered, when
> you examined it in the editor it misleadingly implied that you could
> park with one end of your car blocking half of the street.
Well, there's an easy solution to that; map the spaces, also ;-). That 
said, I find that attitude slightly asinine; it's normal for a parking 
lot area to include at least parts of the aisles.

> I did one car park which attempted to deal with that complaint:
> https://www.openstreetmap.org/?mlat=52.10572&mlon=-4.37367#map=19/52.10572/-4.37367
> but it looks so ugly that I doubt I'll do that again.

Agreed (on the 'looking ugly'). I also tend to line up edges of lots 
along ways (example: https://www.openstreetmap.org/way/826298876). 
However, part of that was from an effort to make the lot only cover the 
area where you can actually park, without disconnecting it from the ways 
that serve it (the aforementioned lot is a good example of my 'older' 
technique). However, now that I am drawing the spaces also (thank you 
gridify!), I'm generally mapping the entire surface, up to some 
semi-arbitrary cutoff representing the entrances. I suppose if the roads 
were also mapped as areas, I would probably connect those rather than 
mapping them as separate driveways, although I'm on the fence there 
also. For example, with https://www.openstreetmap.org/way/826239218 I've 
connected the parking aisle directly to the main road, but in other 
places (e.g. https://www.openstreetmap.org/way/829371430) I've modeled a 
separate driveway first.

Anyway, your example would be much more sane if the entire road had a 
mapped area, rather than just the little piece by the parking lot.

> In the end, do what works for you and hope nobody else who comes across it
> thinks it so egregiously wrong that they change it.  At least not until
> you've finished your project.

As long as they don't *remove* information, I don't mind people 
fiddling. That said, I'm not going to hold my breath, there doesn't seem 
to have been much activity on the base until I started working on it.

...and to be honest, another argument for modeling as lots is that the 
parking_lane tagging is rather more obtuse...

Anyway, thanks very much for sharing your thoughts! It's useful and 
definitely appreciated.

-- 
Matthew