On 28/07/2011 15:13, Nathan Edgars II wrote:
Colin Smale wrote:
I assume all the variations would be easily understood anyway so it
probably doesn't matter that much. Not worth a protracted debate in any
case... In this case it might make sense to opt for the variant with the
least possiblity of misunderstanding, so "Parkway". If only the three
abbreviations were used, then "Pkwy". How about always picking the
"longest form present on a sign"?
Then the process for finding the name would need to include traveling the
entire length of the road and looking at each sign. Because of sign
inconsistency, one would end up with parallel streets, some abbreviated and
some not.
It's still the same road, even in the face of variations in the
spelling. I can't for the life of me think of a reason for worrying
about whether it's Pkwy or Parkway, as both are easily understood to be
equivalent. There will only be one official spelling (cf. your plat
example) which, if it differs from the evidence on the ground, should go
in "official_name". The key criterion (IMHO) is whether the name shown
on renderings based on OSM data is useful. Anyone following a satnav's
directions onto "Orlando Parkway" is not going to decide to ignore a
sign saying "Orlando Pkwy" on the grounds that it is probably a
different street and carry on scouring the area for a sign that says
"Orlando Parkway".
In the case where a street has different spellings on the sign at the
two ends, where's the boundary? You can't tell, because there isn't one.
It's the same street. Pick a spelling, and stick to it. Many renderers
(e.g. mkgmap+Garmin nav) will be relying on the spelling being
consistent from end to end. If the street name changes in the course of
the street, even by one letter, it can be considered a "different
street" leading to extraneous directions. To me, consistency is more
important than 100% accuracy as judged against soft criteria.
Colin
_______________________________________________
talk mailing list
talk@openstreetmap.org
http://lists.openstreetmap.org/listinfo/talk
