[HOT] Updating the Humanitarian Data Model?
kate at maploser.com
Tue Feb 28 03:52:17 GMT 2012
I think there is really a need to update the wiki pages about the
Humanitarian Data Model. Other organizations currently ask if there
is a plan to maintain it and how it will develop. In its current
state I think it is just going to be too difficult. There has not been
much discussion openly of the HDM. Most of the original development
occurred initially after the earthquake in Haiti, so a more thorough
look could be taken to match it more closely with existing OSM tags
I think the best move would be to have sets of thematic tags, but to
keep it simple. For example in Indonesia we have building tags for
impact modeling. They are pretty straightforward and use a
combination of tags that already exist in OSM and the couple that we
had to add that did not.
I think having some sort of version of Tag Watch or something similar
to show what is in use will allow more dynamic development as well. I
find there is a lot of disconnect between tag development on the
ground and the wider community. This will likely lead to different
tagging conventions and difficulty in creating a more universal
I think having a better way to record tags and turn them into presets
would be great. Something easily human readable but parsable into
JOSM presets, OSM2PGSQL styles, etc
On Sat, Feb 25, 2012 at 6:36 AM, Chris Blow <cgblow at gmail.com> wrote:
> Yesterday on the chat I was asking about the Humanitarian Data Model:
> "HDM transportation objects are based on the UN Spatial Data Infrastructure for Transport (UNSDIT). The below humanitarian transportation framework is a subset of the UNSDIT tailored to match Haiti operational requirements based on direct field needs assessments and experiences gained in past humanitarian responses."
> It appealed to me initially as an excellent point of collaboration and data integration and discussion, for people who want to contribute icons, etc.
> However the HDM seems to have gone cold since initial application in Haiti, in 2010, without a lot of discussion of revision — perhaps because is so complicated, and it does not seem to be aligned with the needs of recent HOT projects on the ground.
> There is some discussion on the wiki, but it's a bit opaque to me. I think the UNSDIT probably a useful thing to reference but after working with it a bit, I started to feel that it is not a strong starting point for collaborative tag work in OSM. It seems to be an ontology with a lot of detail but low usability.
> It is featured prominently in the HOT Package: wiki.openstreetmap.org/wiki/Humanitarian_OSM_Team/HOT_Package
> I would love to see a more dynamic view of which tags are actually in use in HOT projects, which can then be used as the basis for recommended tag set. My hope is that the HDM, if I understand its role correctly, would be a very dynamic model, not something that has been decreed in detail, then forgotten.
> I greatly admire OSM's dynamic tools for understanding tag taxonomies, especially Jochen Topf's Taginfo: http://taginfo.openstreetmap.org/about
> Taginfo provides a REST interface for examining which tags are actually in use and thus (it seems to me) is a better recipe for a recommended data model. I relied on it (and tagstat and other tools) when I was developing recommended tags for the indonesia maps last year. The existing HDM simply didn't make sense in my context, and I suspect others may be confused by it as well.
> I am curious;
> - am I missing the point of the HDM entirely?
> - has there been much discussion about the HDM that I missed, perhaps on another list? (eg the tag list)
> - is there strong consensus around the posture of the HDM, or is there perhaps some energy for updating it and changing course?
> - Is a more dynamic recommended tag set possible and desirable to others?
> - Do others see value in an "official" recommended set of core tags in the first place?
> PS: is there not a better format for storing such a model than in mediawiki (or JOSM presets)? Something easier to read and update perhaps?
> HOT mailing list
> HOT at openstreetmap.org
More information about the HOT