<!doctype html public "-//W3C//DTD W3 HTML//EN">
<html><head><style type="text/css"><!--
blockquote, dl, ul, ol, li { padding-top: 0 ; padding-bottom: 0 }
--></style><title>Re: [Talk-us] Freeway
directions</title></head><body>
<blockquote type="cite" cite>On Thu, Oct 17, 2013 at 3:39 PM, Brad
Neuhauser <<a
href="mailto:brad.neuhauser@gmail.com">brad.neuhauser@gmail.com</a>>
wrote:<br>
<blockquote>From <a
href=
"http://en.wikipedia.org/wiki/Interstate_Highway_System#Primary_.28one-_and_two-digit.29_routes_.28contiguous_U.S..29"><span
></span
>http://en.wikipedia.org/wiki/Interstate_Highway_System#Primary_.28on<span
></span>e-_and_two-digit.29_routes_.28contiguous_U.S..29</a> :<br>
</blockquote>
<blockquote>"In the numbering scheme, east-west highways are
assigned even numbers and north-south highways are assigned odd
numbers. Odd route numbers increase from west to east, and
even-numbered routes increase from south to north (to avoid confusion
with the U.S. Highways, which increase from east to west and north to
south), though there are exceptions to both principles in several
locations."<br>
</blockquote>
</blockquote>
<blockquote type="cite" cite><br></blockquote>
<blockquote type="cite" cite>Field signage is sometimes inconsistent
with the official rules; for example, US 68 is mostly (entirely?)
signed north-south, and I-69 becomes east-west between Lansing and
Port Huron. States may have their own rules; some states (MS,
FL) follow the even-odd rules that the national routes do, some use an
opposite pattern (even N-S, odd E-W), and some have no pattern at all
(GA, TN).</blockquote>
<blockquote type="cite" cite><br></blockquote>
<blockquote type="cite" cite>There are also cases of signage by loop
nesting ("inner" clockwise/"outer"
counterclockwise for RHD) - I-495 around Washington and I-440 around
Raleigh NC are examples, along with GA 10 Loop around Athens GA. And
in Canada the QEW is "directionally" signed by destination
(Toronto on the clockwise carriageway, Niagara and then Fort Erie on
the counterclockwise one). There may be a few more oddballs I've
missed.<br>
</blockquote>
<blockquote type="cite" cite><br></blockquote>
<blockquote type="cite" cite>IMO preferred practice should be a
relation for each continuous cardinal direction, to keep validation
simple; undivided roads should use forward/backward roles to
distinguish which relation applies to the underlying way's
forward/backward traversal. It shouldn't be too terribly hard to come
up with an algorithm to fixup the existing single-relation cases,
particularly for the ones where the routes are entirely dual
carriageway, although occasionally the heuristics will be wrong and
need a manual edit.</blockquote>
<div><br></div>
<div>Well said, Chris. +1, lands right in the middle of where it
might, imho.</div>
<div><br></div>
<div>Manual edits on exceptional cases means that somebody, somewhere
is paying attention. In a big set, with careful management,
that's to be expected. Good show.</div>
<div><br></div>
<div>Encore, author. This is a serious project which aims to
accurately map "what is." Nice job so far, everybody.
I very much like this crowd-sourced map. It only keeps getting
better and better.</div>
<div><br></div>
<div>SteveA</div>
<div>California</div>
</body>
</html>