[HOT] Database, OSM & HOT (Was: Request for information about common set of tags for HOT)

john whelan jwhelan0112 at gmail.com
Wed May 20 19:01:39 UTC 2015


OSM has a page of recommended tags,
http://wiki.openstreetmap.org/wiki/Map_Features.  Sometimes these are used
sometimes not.  http://taginfo.openstreetmap.org/ has information on how
common a tag is and is some times used to determine which tag should be
used.  This is a bottom up approach rather than the top down approach more
usual in the business world.

For example many of the mappers locally are enthusiastic cyclists but the
recommended tags for cycle paths have been formulated in Europe, in Canada
we have some cycle paths / lanes that are used for cyclists in summer and
we dump snow on them in winter.  There is a local convention for how these
are tagged.

With a conventional database you usually have a client in mind and they
have specific requirements.  OSM doesn't.  If someone wants to map
hairdressers that's fine as far as OSM is concerned they have contributed
to the map.  Locally many tags were in OSM that wouldn't render on
conventional rendering systems, no one else used the tags and the renderers
just ignored them.  Many mappers have their own personal views about how
something should be tagged and have no interest in following any other
suggestions at all.  It is an issue.

HOT is much more structured, we actually have clients with requirements in
mind so we map to those as best we can.   We have recommended tags and to a
much larger extent people follow them.  We do have a lot of new mappers who
may not even know about the map features page or find that reading through
more than two lines of instructions boring.  Having a two step process with
validation helps as well.  However we still rely on locals on the ground
mappers for more detail and as far as I'm concerned if they want to map
video games, hairdresser, gymnastics, karate, volleyball or football fields
that's fine, they might map something else of use whilst they are mapping
or introduce someone else who might map something more useful to us.

What the agencies like is that we can map places very quickly which is
better than no maps.  Also we are very cost effective, I was going to say
cheap but that has quality implications.  They can add their own specific
tags without having to go through a formal standards committee. Typically
it takes five years to get something through the ISO standards process.
Currently in the background I believe HOT going through a "standard's
process" as we progress.  Nepal has had a big impact on HOT, Ebola came
earlier in the spring and is still around, but the scale of mapping in
Nepal has caused a rethink about how we do things including training,
things are becoming more formalised but having said that I don't think it
will ever be totally rigid.

Cheerio John

On 20 May 2015 at 11:32, althio <althio.forum at gmail.com> wrote:

> Springfield,
>
> Sorry for the partial answer and I don't mean to be harsh because I
> know things around here are not easy to find and understand. We all
> need pointers and FAQ or homepages and portals...
>
> My point is... I do think that you are somehow confused between
> OpenStreetMap and HOT:
>
> ***
> OSM aka OpenStreetMap, the project, its database, its goals, its community
>    - [http://www.openstreetmap.org/welcome] OpenStreetMap, the free
> and editable map of the world
>    - [http://www.openstreetmap.org/about] OpenStreetMap is built by a
> community of mappers that contribute and maintain data about roads,
> trails, caf├ęs, railway stations, and much more (note: also video
> games, hairdresser, gymnastics, karate and volleyball...) (note2: also
> boundaries, hospitals, schools), all over the world.
>
> see also OSM Foundation, the entity to support the project
>    - http://wiki.osmfoundation.org/wiki/Main_Page
>
> ***
> HOT aka Humanitarian OSM Team is using a subset of OSM database and
> building on it, its own goals (some overlap with OSM), its own
> community (some overlap with OSM)
>    - [http://hotosm.org/] The Humanitarian OpenStreetMap Team [HOT]
> applies the principles of open source and open data sharing for
> humanitarian response and economic development.
>    - [http://hotosm.org/about]
>    - [https://wiki.openstreetmap.org/wiki/Humanitarian_OSM_Team]
>
> ***
>
> Back to OSM and "tag soup" database. This is a rather hard and
> technical topic, and not really related to HOT and this list. You will
> not find the answer here, nor the most interested or skilled people.
> HOT uses and contributes to the database, HOT does not control it.
>
> A few more words anyway? ("I am not a lawyer" and "I am not a database
> expert").
> OSM database is open, free, public, iterative and rather rich.
> OSM database is not fixed, not complete, not comprehensive, not
> homogeneous (spatially at least).
> The philosophy and structure have their own advantages and
> disadvantages compared to existing datasets.
> Please appreciate the uniqueness, value and potential of OSM database
> before you try to make it a clone of something existing.
>
>
> All the best,
>
>  - althio
>
>
> On 19 May 2015 at 21:38, Springfield Harrison <stellargps at gmail.com>
> wrote:
> > Hello Stefan & Blake,
> >
> > I concur with the comments about the "tag soup" mess.  As I have
> mentioned
> > before, I am new to this OSM environment but have some years experience
> with
> > GPS and GIS mapping and database design.
> >
> > To be honest, I was appalled when I discovered that the OSM database
> design
> > looked like a glorified scratchpad.  I just downloaded and inspected
> 366,017
> > OSM database records.  There were 18 Key Terms and scores of values.  I
> > extracted the unique combinations of keys/values and ended up with 388
> > records of those.
> >
> > It is difficult to describe the results in detail as patterns are very
> hard
> > to see with this system.  Suffice it to say, there is an abundance of
> > overlap, redundancy, ambiguity and a confusing intermingling of features
> and
> > attributes.  Using traditional methods of querying a database, it would
> be
> > impossible to definitively extract a meaningful subset of any of the
> 366,000
> > records.  Generally speaking, the problem is that one feature may be
> > described in many different ways that are not consistent.
> >
> > Having said all that, since I frequently hear how well all this mapping
> > information is received in the field, I must conclude that this mishmash
> of
> > tagging somehow creates a usable end product.  It may well be that I am
> not
> > aware of magic techniques that bring order to all this chaotic tagging.
> > However, if it works, it is good.  However I do believe that it will work
> > better with a more robust database.
> >
> > Sorry to offer this harsh critique, but in decades of looking at database
> > structures for both geographical and administrative applications, I have
> > never seen such a jumble of terminology.
> >
> > Anyway, I have put together what I believe is a more appropriate Data
> > Dictionary that generally parallels the best practices in database
> design.
> > I have found this approach to be very useful, and also useful in the
> field,
> > since being introduced to it by Trimble Navigation in the early 90s.
> >
> > I am impressed with the enthusiasm that permeates the crowd GIS
> initiative
> > but concerned that the geographical and database underpinnings may be
> less
> > than ideal.  My observation from creating a few software applications, is
> > that the lesser trained are the users, the much greater investment there
> > needs to be in the user interface and training.  GIS and GPS data
> collection
> > is not particularly intuitive.
> >
> > My approach in projects of this kind is always to start at the far end
> with
> > the users - what information are they wanting for whatever it is that
> they
> > do?  Then I look at the reporting requirements and finally design the
> data
> > collection process to feed into that.
> >
> > In the case of this emergency relief operation, I'm hard-pressed to see
> the
> > value in mapping video games, hairdresser, gymnastics, karate and
> > volleyball.  To be fair, many of the other attributes could have value in
> > providing relief services but in the record set that I downloaded, there
> > seems to be little information related to the emergency relief effort.
> In
> > over 366,000 records there are only 19 marked as aeroway = helipad.
> >
> > I'm not sure just how thorough you intend to be with the "updating,
> > streamlining and regularizing" but I would be happy to help where
> possible.
> >
> > It would probably not be overly difficult to substitute a new
> > feature/attribute catalogue into the OSM database.  Translating the
> existing
> > mass of keys and values to their new equivalent might be more
> challenging.
> > Databases succeed because they conform to standard pattern sets.
> >
> > Again, sorry to be less than enthusiastic but perhaps things can be
> > improved.
> >
> >         Thanks for your patience, Cheers . . . . . . . . Spring Harrison
>
> _______________________________________________
> HOT mailing list
> HOT at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/hot
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20150520/5a22dc85/attachment.html>


More information about the HOT mailing list