On Mon, 24 Aug 2009 17:00:14 +1000, Roy Wallace wrote:

> On Mon, Aug 24, 2009 at 3:39 PM, David Paleino<d.pale...@gmail.com> wrote:
> > On Mon, 24 Aug 2009 08:53:53 +1000, Roy Wallace wrote:
> >
> >> On Sun, Aug 23, 2009 at 8:48 PM, David Paleino<d.pale...@gmail.com> wrote:
> >> > Hello,
> >> > I'd like to start discussion on the deprecation of the Tag:highway=stop
> >> > in favour of using stop=yes/both/-1.
> >>
> >> First impression: the value of the tag is extremely ambiguous, and in
> >> no way self-explanatory. I don't like it at all.
> >
> > It has the same values as oneway=*. If you use Key:oneway, you know how to
> > use Key:stop.
> 
> 1) no, it doesn't (yes/both/-1 vs yes/no/-1)

Fair enough.

> 2) the meaning of "yes" and "-1" is different for oneway! ("yes" means
> "forward" as opposed to "on the last node"; "-1" means "in the reverse
> direction" as opposed to "on the first node")

Think broader: "yes" might mean "apply this in the same direction of the way",
while "-1" → "opposite direction". However, I don't want to nitpick about
this ;-)

> Seriously, stop=-1 is not self-explanatory! Even if the values of
> oneway matched up (which they don't), it still wouldn't make stop=-1
> self-explanatory.

Ok, fine.

> > Aren't we tagging what we see in the real world? I'm of the opposite
> > opinion, we tag stop *signs* (horizontal or vertical signs), and we're
> > trying to relate those signs to the junction they have effect on.
> 
> If you want to put a stop *sign* on the map, use a separate node with
> traffic_sign=*.
> 
> If you want to describe an attribute of the intersection of ways, it's
> quite alright to assign this attribute to the way/intersection itself,
> because it is indeed an attribute of the way/intersection.

Both things are related -- you shouldn't use Key:stop if there's no stop sign
in the real world.

> >> How about stop=at_last_node, stop=at_first_node and
> >> stop=at_first_and_last_node? More verbose, but a lot clearer than
> >> yes/-1/both.
> >
> > That can be done too. More concise:
> >
> >  stop=first (-1)
> >  stop=last  (yes)
> >  stop=both  (both)
> 
> Hrmm that is more concise, but I think less self-explanatory (remember
> that not everyone reads the wiki before editing).

Well, they must IMHO. The wiki explains the ontology of the tags we're using,
and the wiki is the main regulamentation for tags. Otherwise we go wild, and
everyone uses what she likes best.

> E.g. stop=both could be misunderstood to mean "both directions", or "both
> intersecting ways", etc.

stop=both_sides? Propose something :-)

> Also, need to clarify something...:
> 
> Let's say way A is drawn from West to East, then at some point becomes
> (intersects with) way B, which continues to the East.
> And let's say East-bound travelers have to *stop* at the junction (for
> some reason), but West-bound travelers don't.

If I understood "East-bound" and "West-bound" correctly, you mean:

  http://imagebin.ca/view/bJWJB6.html

?

> This would be tagged as A being stop=at_last_node. Right?

No, it would be stop=at_first_node (or whatever we decide it to be) assigned to
the second segment of B. [1]

> For West-bound travelers, at the instant they cross from B to A,

From the drawing I made above, B has no West-bound travelers. Maybe I
misunderstood your description?

> this would imply that they should stop, because they're at the last_node of
> A. Which is not the case. In other words, it would seem to me that the
> proposal needs clarification in the form of something like:
> 
> "The stop=* tag is applied to a way to specify the node at which the
> stop sign applies. However, the stop sign only applies when the node
> is approached from the way that is tagged."

[1] the use of first_node, first, -1, whatever, means that the stop applies to
    people coming from the opposite direction the way is drawn. This is
    basically what you're suggesting here, I suppose?

Probably some drawing by you would be best! ;)

Kindly,
David

-- 
 . ''`.  Debian maintainer | http://wiki.debian.org/DavidPaleino
 : :'  : Linuxer #334216 --|-- http://www.hanskalabs.net/
 `. `'`  GPG: 1392B174 ----|---- http://snipr.com/qa_page
   `-   2BAB C625 4E66 E7B8 450A C3E1 E6AA 9017 1392 B174

Attachment: signature.asc
Description: PGP signature

_______________________________________________
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk

Reply via email to