<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title></title>
</head>
<body>
<div name="messageBodySection">
<div dir="auto">Dave F wrote:<br />
> The link role of a route relation can be placed on *any* way. It's <br />
> commonly used as an indicator for 'this is the way to go to get <br />
> to the route' and not just connections between different routes.<br />
> <br />
> In the UK National Cycle Routes have signs for this with the ref <br />
> number in brackets.<br />
<br />
Yes - this is indeed the correct historic explanation for why the 'link' role exists.<br />
<br />
Here's an example near me of how this might be signed (apologies for GSV, but I'm too much of a wuss to ride out and take a picture in this weather):<br />
<span style="white-space:pre"> </span><a href="https://goo.gl/maps/gcKAzfXX3CTMEgwr5" target="_blank">https://goo.gl/maps/gcKAzfXX3CTMEgwr5</a><br />
<br />
Personally I tend to tag this as a short discrete network=lcn route, but the 'link' role on the main relation isn't wrong.<br />
<br />
The explanation on the wiki page looks like someone who didn't understand the role attempting to document it. (On the OSM wiki! I'm shocked too! This has never happened before!)<br />
<br />
<br />
Minh Nguyen wrote:<br />
> At the current usage level, data consumers probably would be <br />
> unable to depend on the "link" role for any practical purpose.<br />
<br />
I do use it in cycle.travel's turn-by-turn directions code. If a route is tagged as a "link" to the River Loire cycle route, then by definition it's not the main trunk of the Loire cycle route. Therefore cycle.travel won't say "Follow the Loire cycle route" when turning onto this way.<br />
<br />
Richard</div>
</div>
</body>
</html>