https://it.wikipedia.org/wiki/Ciclovia_Alpe_Adria : "Da qui con il treno navetta in 11 minuti si raggiunge Mallnitz (1.191 m), e quindi di nuovo in bicicletta si attraversa la Carinzia toccando Spittal a. d. Drau, Villach e Arnoldstein, al confine italo-austriaco"

https://www.alpe-adria-radweg.com/en/bahn-rad/unterwegs-am-alpe-adria-radweg-von-salzburg-nach-grado/ (the official site): "At the end of Gastein Valley, you will have to use the 'Tauernschleuse' motorail service (Böckstein – Mallnitz) in order to reach the south side of the main Alpine divide"

https://italy-cycling-guide.info/international-cycle-routes/ciclovia-alpe-adria-radweg/ : "it heads through the Gasteinertal towards Bad Gastein and Böckstein where you pick up the Tauerntunnel motorrail shuttle that takes you to Mallnitz"

There is literally no other way to get from Böckstein to Mallnitz other than a 160km detour involving significant backtracking. If you don't think the route does follow the rail shuttle, which way do you think it does go?

Bike routes in OSM have included ferries since time immemorial. Trains are less common but no different in practice. I put code specifically in cycle.travel to cope with routes like this.

See https://www.alpe-adria-radweg.com/en/ - it's very definitely an official part of the route.

Hi,

Could you explain why you've removed the railway from the bike route relation?

The bike route requires going on the train. It's an integral part of the route so should be included in the relation.

Removing the railway has broken routing on cycle.travel for this route. :(

Richard

thank **** for that

Wow - just spotted this incredibly detailed bit of mapping. Hats off!

Nice work! Just spotted this on the Rails-to-Trails group on Facebook and wondered if it had been mapped… delighted to see it has :)

Thanks \o/

I suggest maybe using a proper editor like iD instead of JOSM which is well-known for breaking data like this

<runs away very very fast>

Joshua - you need to fix your router. Blacklisting barrier=* will fail unexpectedly in lots of other places: for example, the edges of central London had anti-terrorist barriers (actually just little checkpoint booths) which were tagged as barrier=checkpoint nodes for many months.

The only way to parse barrier= nodes is to blacklist a defined list of values and assume everything else is ok. Believe me, I've been through this with getting cycle.travel's routing working. :)

Hi Lewis,

Great to see the edits you've been doing!

In this one you (accidentally I guess) deleted a cycleway alongside the road. This means that it's harder for bike maps and routing sites to show the best way. In particular, it's broken the continuous route of National Cycle Network route 8 (the cross-Wales route, Lon Las Cymru).

Generally you should be reluctant to delete stuff from OSM - it's usually been put there for a reason even if that's not obvious! You can leave changeset comments like this if you want to ask an editor what they were doing.

I've fixed the NCN 8 cycleway route for now.

Best wishes
Richard

I'm not aware of any convention in the UK that "request stops" are always tagged as railway=halt. The UK railway system doesn't really have a clear distinction between station and halt as some other countries do.

(FWIW a station can't really be a "request stop" because stopping by request is an attribute of a train service rather than of a station. Usually if one service stops by request then all do, but it's not always the case!)

Hi - you've added a lot of place=town nodes here, but they don't really seem appropriate for the features in question and they duplicate existing information. For example, there's already a place=village name for Cranberry - you don't need to add a duplicate place=town. Because these are small places they're better mapped in OSM as villages.

Could I suggest you revert the changes?

Best wishes -- Richard

Hi Roland,

The statement is at https://www.eurovelo.at/en/ev7.html .

Although lifecycle prefixes are reasonably common for amenities etc., they aren't in general use on route relations - only seven occurrences worldwide according to https://taginfo.openstreetmap.org/tags/type=proposed%3Aroute . As such they aren't understood by data consumers that parse route relations.

Ah good - I didn't spot that bit from the train :)

Reinstated type=route.

Hi Roland,

In this you've changed the type tag from "route" to "proposed:route". This isn't standard tagging for route relations - it should be type=route, state=proposed.

Is it really the case that all of the route in Austria is only proposed? According to the RadLobby EV7 website, "The Austrian section of EuroVelo 7 is signposted as 'EuroVelo7' everywhere apart from the province of Upper Austria." So I'd think it should be predominantly type=route.

Let me know.

Richard

OSM is rarely 100% consistent :)

From my point of view (I run cycle.travel) it's not actually OpenCycleMap's rendering that's the issue - OCM shows route=mtb anyway. The problem is more that most _routers_, including cycle.travel, use route=bicycle as a flag that "this is a good way to route an ordinary bicycle along". That isn't really the case for some of the gnarlier bits of the Great North Trail!

I've changed it to route=mtb already so no need to do any more work on that.

I would suggest you do need to get clearance from CUK for the route to be entered into OSM though - they may assert copyright over the route (this isn't that unusual for route creators) and we can't have it in OSM if the creators don't give permission.

I've changed it to route=mtb, unsigned=yes.

This shouldn't really be in OSM I'm afraid. If we were to tag unsigned routes invented by some well-meaning sort then we'd have the complete collection of Cicerone guidebook routes and so on. Long-standing precedent is that we do not map these in the UK.

The route is also potentially copyrighted by Cycling UK ("freely available" does not mean free of copyright).

At the very very least it absolutely shouldn't be route=bicycle, network=ncn, which in the UK is used for National Cycle Network routes suitable for most bicycles. route=mtb is the correct tagging for a (waymarked) route like this.

I think that the shared-use path in Wilton, along Salisbury Road, postdates the original Wiltshire Cycleway (perhaps a Sustrans Connect2 project?) - and it's that which has enabled the route to be rerouted away from the A3094 and onto South Street. So it's quite plausible that OSM and the Wiltshire PDFs are just reflecting the previous route.