[Imports] [Talk-us] TriMet will begin OSM improvements in the Portland area

[Thomas Emge]

Serge,

In the translation process all nodes are turned into points and no information is lost - if we are losing data then that would be considered a bug and please let us know if that occurs. I am not sure what is implied when you say "nodes with other data" - in this context what is "other data"? 

In the current implementation the not "worthy" tags that would classify a node as a supporting point (and not a point feature in its own merit) are "created_by", "source", "attribution", and "note". Again all nodes are converted into points and based on the tags above they are getting assigned the attribute "supportingElement" of "yes" or "no".  The points with an attribute value of "supportingElement=yes" can be filtered out for viewing purposes but they continued to exist in the data base to maintain the integrity of the original OSM data.

The assumption then would be that the user doesn't necessarily have to be concerned about the supporting nodes as they are a descriptive elements of a way or a relation - or in the GIS world a vertex of a line or polygon feature. In the process of editing a feature (point, line, or polygon) the user might want to get back down to the vertex level but doesn't have to do.

Our concept of "layers" is then a user defined abstraction we are using to define symbology and type (pharmacy icon for tag amenity=pharmacy) and thematically group features. As of now we have a generic OSM abstraction but different communities are looking into creating their own like HOT, etc..

I would definitely like to have a chat with you about the challenges of importing data into OSM - are you at the SOTM in Denver?

Thanks,

- Thomas

-----Original Message-----
From: Serge Wroclawski [mailto:emacsen at gmail.com] 
Sent: Friday, March 04, 2011 4:12 AM
To: PJ Houser
Cc: imports at openstreetmap.org
Subject: Re: [Imports] [Talk-us] TriMet will begin OSM improvements in the Portland area

Obviously this mail was incomplete and fat fingered.... Let me finish now.

My main concern here is that the Arc tools are abstracting away details about OSM that are unimportant when the workflow fits the Arc model, but not the OSM model and workflow.

Specifically in this context, I'm concerned that if nodes are just considered points, that nodes with other data might be inadvertently lost, or else made wrong in a transfer.

This goes hand in hand with working in OSM vs one's own dataset, where one can (and should) feel free to make large modifications in one's one dataset, but working in a shared playground is much harder.

This concern isn't Tom specific, nor your project specific, but I'm wondering how it is you're addressing this. I asked this question of the ESRI folks who developed the tools and their response was that it was the responsibility of the user. Compounding the challenge, if I recall, the "layer" analogy extends to the OSM tools, so you might not see objects' other tagging.

How are you overcoming these technical and cultural challenges as you work to bring your data into OSM?

- Serge

_______________________________________________
Imports mailing list
Imports at openstreetmap.org
http://lists.openstreetmap.org/listinfo/imports