[OSM-talk] ways, rendering

[Robert T Wyatt]

Jochen Topf wrote:
> Hi!
> 
> I have added the following to the FAQ at
> http://wiki.openstreetmap.org/index.php/Tagging_FAQ:
> 
> -------
> Segments should not have any tags (except "created_by" which is
> automatically added by the editor software). Put all tags on the ways.
> Split up ways as necessary.
> 
> This is a much debated topic. See Tagging Segments for more details and
> reasons.
> -------
> 
> And I have added the new Wiki page
> http://wiki.openstreetmap.org/index.php/Tagging_Segments
> 
> with basically the text from my earlier posting here. Not perfect, but
> feel free to improve the explanations.
> 
> Generally I think we need to add things discussed here to the Wiki more
> often. Otherwise they are "lost" and we have nothing to point to when the
> issue comes up again.
> 
> Jochen

I would like to see on the wiki the parts of this discussion which
delineate the considerations here regarding markup of segments/ways. It
would be nice to have a summary of the pros and cons of the different ways
to markup this situation for instance:

         |
---------+
         |
         +------------
         |
         |

Clearly we already let the user decide how to do it, but I think we can at
least help the user make informed decisions. This discussion has been very
enlightening to me and I just now have finally understood why my Garmin
Quest2 sometimes tells me to "continue on ..." or why I am sometimes
directed to turn at a bend in the road where there is no intersection.

I think in the illustrated example above adopting the convention(-?) used
for roundabouts of using the highest-order roadway would be advisable
assuming(!) this is the vertical road in this example. Here in Austin, the
people have demonstrated a proclivity for giving multiple names to the same
roadway and, worse, the multiple names sometimes overlap and sometimes
don't. I have gleaned from this discussion some valuable guidance for how
to deal with these things.

.02 more, the intrinsic value in maps is the (accuracy of the) information
they contain. More is better, however an unreadable map is a waste of time
and effort. The underlying design and implementation should look to future
(rendering and routing) functionality for guidance rather than be
restricted by current (rendering) technology. Does this mean that the
current renderers may be less readable due to too much embedded
information? Maybe in some cases. The classic solution is to use layers to
selectively display certain data in the present view. The current OSM
renderers already allow for certain zoom levels to display certain types of
data don't they? I suppose that this method still won't resolve
double-markup on divided ways, but I believe renderers could be educated to
understand this. There should not be any prejudice against redundant markup.

Back to my day job,
Robert