<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 29/05/12 18:10, Tom Chance wrote:
<blockquote
cite="mid:CACD80NTh=X-zc=tXWx3WBTVqXorQfFp7h0vNS8o_vi=ko6+OTQ@mail.gmail.com"
type="cite">
<div class="gmail_quote">On 29 May 2012 17:19, Chris Hill <span
dir="ltr"><<a moz-do-not-send="true"
href="mailto:osm@raggedred.net" target="_blank">osm@raggedred.net</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="im">
<blockquote class="gmail_quote" style="margin:0pt 0pt 0pt
0.8ex;border-left:1px solid
rgb(204,204,204);padding-left:1ex">They need to be
manually entered as relations sharing nodes with those
features. <br>
</blockquote>
</div>
I would say that sharing nodes can lead to problems.
Boundaries that get imported or manually traced from OS data
often have no visible reference on the ground. If you share
nodes with something else, when someone aligns that something
else to aerial imagery, or a GPS trace or whatever, the
boundary (which was probably right) gets moved too. Why do
nodes of one object need to be shared when they are quite
different objects?</blockquote>
</div>
<br>
This probably varies according to a number of factors, but where
boundaries are just abstract information coterminous with physical
features it makes sense to me that the objects in OSM share nodes.
Many boundaries in urban areas could just be relations containing
lots of roads.<br>
<br>
</blockquote>
<br>
My question is: how do you know the boundary aligns with an existing
object? If they align, what evidence on the ground do you have for
that? How do you *know* that a road, stream or whatever aligns with
the boundary, other than using other sources such as complete
(copyright) OS maps?<br>
<br>
I would say import or draw the boundary based on the OS Open data as
a separate entity from anything else. <br>
<br>
<blockquote
cite="mid:CACD80NTh=X-zc=tXWx3WBTVqXorQfFp7h0vNS8o_vi=ko6+OTQ@mail.gmail.com"
type="cite">Take the example of someone moving a road but not also
moving the boundary. That introduces an inaccuracy right away. </blockquote>
<br>
How do you know that? What rule is there that says a boundary must
follow a road? Why, if the OS Boundary data has been used, would you
want to move a boundary unless an administrative change has been
made?<br>
<br>
<blockquote
cite="mid:CACD80NTh=X-zc=tXWx3WBTVqXorQfFp7h0vNS8o_vi=ko6+OTQ@mail.gmail.com"
type="cite">They could also move both but not bother to get the
locations exactly right, say by making the gap between them
larger, getting a kink in the wrong place, or not having them
exactly coterminous. Perhaps they aren't all that interested in
boundaries. But then along another person comes to check if a
house is in this or what ward and they're misled. I prefer to have
the boundaries share nodes so that people are forced to move
boundaries with roads/streams/etc. and forced to break them apart
if they really aren't coterminous.<br>
</blockquote>
<br>
I think some areas such as some landuse can be described as
coterminous and might benefit from sharing nodes. Sharing nodes
between disparate object types causes more trouble than it is worth
IMHO. I don't see how a road and a boundary can be described as
coterminous since there is no evidence of it from a source we can
use. <br>
<br>
<blockquote
cite="mid:CACD80NTh=X-zc=tXWx3WBTVqXorQfFp7h0vNS8o_vi=ko6+OTQ@mail.gmail.com"
type="cite">
<br>
To my mind in these cases boundaries should be treated the same as
routes.<br>
</blockquote>
<br>
I don't understand that comparison.<br>
<br>
<blockquote
cite="mid:CACD80NTh=X-zc=tXWx3WBTVqXorQfFp7h0vNS8o_vi=ko6+OTQ@mail.gmail.com"
type="cite"><br>
Of course if you don't know that a stream and a boundary are
supposed to be related and you go and share nodes because they
happen to be in roughly the same place, then moving the stream to
align with a GPS trace obviously shouldn't move the boundary so
you introduce problems.<br>
</blockquote>
<br>
That's my point, how would you know that a boundary is 'supposed' to
be related to anything else? Are boundaries ever anything more than
arbitrary? They don't exist on the ground, even the boards saying
"Welcome to My Town" are not always placed on the boundary because
of traffic sign placement rules, indeed, most parish boundaries are
not marked, the village name sign is no where near the parish
boundary.<br>
<br>
People will do as they think best, and I'm sure some boundaries will
end up with shared nodes - I just don't see the benefit nor the
justification and I do see problems. <br>
<pre class="moz-signature" cols="72">--
Cheers, Chris
user: chillly
</pre>
</body>
</html>