[HOT] paths, tracks and unclassified in West Africa

Thomas Gertin tgertin at gmail.com
Thu Jul 16 21:19:03 UTC 2015


That’s a good point. The current definitions take a variety of attributes into account to define what type of highway it is (physical attributes of the highway, size of urban areas connected by highway, the type of use the highway is used for, and whether cars can travel on the highway). Looking at it from a data dictionary perspective it is better to have less highway types, and put some of the information used to define highway types into the highway attributes instead. Changing the tagging scheme and updating the existing features would be a really big undertaking though.   -Tom G


> On Jul 16, 2015, at 10:17 AM, Springfield Harrison <stellargps at gmail.com> wrote:
> 
> Which road attribute are you attempting to record?
> 
> Surface type
> Width
> Number of lanes
> Type of vehicle
> Access control (toll, etc.)
> Type of user (farmer, commuter)
> Type of destination (farm, village, city, woodlot)
> Owner (state, logging company, village)
> Seasonality (all weather, 4wd, dry season)
> Steepness
> Straightness
> Other
> 
> It seems that some of the confusion stems from trying to choose one term to encompass all possible attributes combinations.
> 
> You might need to apply more than one attribute per road (the list above).  Or define common  attribute sets to cover typical situations (primary, secondary, etc. Or interstate, regional, local, personal).  The difficulty with the latter is getting a common understanding of the attribute set for each umbrella term.
> 
> Every feature should be tagged as Validated=yes/no.
> 
> A good data dictionary will clearly distinguish between a Feature Type and that feature's attributes.  It is difficult to ad hoc a DD once the project is underway.
> 
> Good luck! . . . . .
> 
> Cheers . . . . .   Spring Harrison, Canada
> Samsung Tab 4
> 
> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20150716/63904334/attachment.html>


More information about the HOT mailing list