[Tagging] What classification for a connecting link?

[Nathan Edgars II]

On Fri, Jun 18, 2010 at 5:24 AM, Richard Mann
<richard.mann.westoxford at googlemail.com> wrote:
> On Fri, Jun 18, 2010 at 3:37 AM, Nathan Edgars II <neroute2 at gmail.com> wrote:
>> On the other hand, http://www.openstreetmap.org/browse/way/52557869
>> and http://www.openstreetmap.org/browse/way/52557872, among other
>> links south of Elizabeth and West Lawn, should be at most
>> primary_link, since the "independent sections" from NJ 4 can only make
>> it that far before hitting residential streets.
> the second primary (because it's two-way), the third primary_link,

Two-way traffic shouldn't prevent something from being a link.
Otherwise this is not a motorway_link:
http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=gloucester,+nj&sll=40.867777,-74.10303&sspn=0.0034,0.0103&ie=UTF8&hq=&hnear=Gloucester,+Camden,+New+Jersey&ll=39.807893,-75.290852&spn=0.006907,0.020599&t=k&z=17&layer=c&cbll=39.807842,-75.290767&panoid=F9YbCTAmZtoxZm9tdRS9nw&cbp=12,311.49,,0,2.89

>> But what about http://www.openstreetmap.org/browse/way/11660654, a
>> typical U-turn jughandle that has two (mapped) driveway intersections?
>> Do the driveways really prevent it from being a link the whole way?
>
> the fourth could be just about
> anything from trunk to service. Mapnik makes a mess if a link
> intersects a service, but that's cos Mapnik renders a trunk_link under
> a service, which is wrong. The simplest is probably to call the fourth
> a trunk with a note that there's a case for it being a trunk_link, but
> that trunk is more renderer-proof.

That seems incorrect, and hence tagging (incorrectly) for the
renderer. Whether something is a link should not depend on how links
are rendered.