[Talk-us] Trimet, Portland, Oregon updates

[PJ Houser]

Some background in case you just learned of this right now:I work at Trimet
transit agency in Portland, Oregon, and we are updating OSM for our area. We
are switching to an open source trip planner that uses OSM data, so we are
updating OSM with jurisdictional data that has been made available to us
(and it's public domain). We are using Trimet's 6 inch aerials for
verification (as much as we can without groundtruthing - we hope the
community will help us with that part). If a way is far off the centerline,
we will adjust it; if the type of street or trail is wrong, we will correct
it; directionality; name. We will keep OSM tags and add in some
jursidictional (RLIS and CCGIS) tags so everyone knows the data source.

But we'd like some advice! Here are our questions:

1) We are editing incorrect trails and adding missing trails from RLIS
(Metro, Oregon) and CCGIS (Clark County, Washington). We'd like to tag
handicap accessibility of the trails we edit or add in. What tag would OSM
mappers prefer? We were thinking "accessibility" with values of "yes", "no",
and "unknown". But should we be distinguishing between ADA's standard
accessibility and a general idea of accessibility? What do you think?

2) Since we are using two jurisdictional datasets for editing and adding
streets and trails, should we delete the TIGER tags if we change a way based
on our datasets?

For example, if we adjust a street type and its geometry, we would have an
RLIS:localid tag=##### (or CCGIS:localid) and RLIS:reviewed=no (or
CCGIS:reviewed=no) tag. The RLIS:reviewed=no is for OSM community members so
they know what's been checked by the community. Is this a good idea? We are
making these edits one by one, so it's not like a TIGER import, but we are
basing our changes on high-res aerials and the most recent quarterly
jurisdictional datasets.

Anyways, if we change a way based on the jurisdictional data, then it's no
longer from the TIGER data, so should we delete those tags from the way?

3) In TIGER data, there are name parts, i.e. name_direction_prefix,
name_direction_suffix, name_base, name_type. Should we imitate that in our
datasets' tags? So far, we have made the jurisdictional data include the
tags for name parts and a tag for full name. Should we just include the full
name? Or do we need to keep the name parts for some geocoders or some other
reason?

Thanks everyone!
--
PJ Houser
Trimet
GIS intern, 503-962-5711 (office)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20110309/a1522268/attachment.html>