Also show highway=motorway information on junctions/forks

[artischocke]

I tried to look into this a bit and it seems that after #15 was implemented, the tool now displays the way information for highway=motorway_link ways leading from the highway=motorway_junction node. However, there are also cases where the ways leading from the exit node are not links, but motorways themselves. Such a fork is northbound I-75 in South Florida where it turns west towards Naples and the Sawgrass Expressway (also a motorway) beginning at the junction (exit 19). Here, the tool only displays information about the node, but not the branching motorway, which also has destination tags.

Now, of course I could load the other motorway in the editor to check the values, but the problem of course is that the tool is not marking the node with the green ring applied to junctions which do have destination tags, which it should as the branching way does indeed have destination tags.

This leads to the second issue, which is that if I load the relation for the Sawgrass Expressway (Florida SR 869 Toll), the first node going north is marked yellow since it's an exit but it can't know that its first way is itself the exit way. In this case, I'm not sure how much value there is by checking the origin node.

[k1wiosm]

Thank you for your comment @artischocke!

I understand what the issue is and I have seen it myself sometimes when using checkautopista.

However, if there is no way tagged highway=motorway_link leading from the highway=motorway_junction node, there is no way to easily identify which way is the actual exit ramp and which way is just the motorway itself.

I would say that at least the first way leading from a highway=motorway_junction node should be a highway=motorway_link way. Then, the second way can be a highway=motorway.

This also fixes the second problem. The relation for the Sawgrass Expressway (Florida SR 869 Toll) should not start right at the highway=motorway_junction node but a little bit further down the road.

Let me know what you think.

[artischocke]

Hi @k1wiosm, thanks for your response.

I would say that at least the first way leading from a highway=motorway_junction node should be a highway=motorway_link way.

No, that's precisely the problem I'm describing above: there are no links at a motorway fork. The situation in question is where a motorway forks into another motorway since the latter's starting point is precisely the junction node. In addition to the example in my original post, you can also look at the junction of Florida's Turnpike and the Florida's Turnpike Extension. As you travel southbound on the Turnpike (highway=motorway), the exit to the Turnpike Extension to the right is also highway=motorway. Both leading branches contain destination tags since it is signed that way: continue on the Turnpike to the south towards Miami, or take the Turnpike Extension to Homestead. The tool chokes in this scenario and only displays the node information:

I would expect the tool to show me both ways here instead of none. That's the point I was trying to make with the screenshot in my original post, namely that no information is displayed at all for the ways at the junction node.

Since the tool knows the ways in the relation, it should be possible for it to determine which ways leading off from it are exits without requiring that they be highway=motorway_link

The relation for the Sawgrass Expressway (Florida SR 869 Toll) should not start right at the highway=motorway_junction node but a little bit further down the road.

Both relations are correct in starting at the junction node. We shouldn't map something a certain way against what's on the ground and official sources just to satisfy a QA tool 😊

Hope I was able to explain the situation more clearly & thanks for your time and effort! :)

[artischocke]

I would consider the following a bug though: Both ways leading from this node have destination tags (see the I-595 way and the I-75 way), yet it shows a red circle and doesn't display any way information at all.

I understand the complexity, but I'm concerned people will start mangling the map to make the warning go away (as they're prone to do).

Thanks again for looking into this! :)