Re: [nzopengis] OSM for outdoors
104 views
Skip to first unread message
Hamish
Jan 13, 2013, 3:37:26 PM1/13/13
to nzop...@googlegroups.com
Hi Jonathan,
> By way of introduction, I want to understand OSM and its suitability for
> several outdoors datasets I am involved in. On behalf of Whitewater NZ, I
> look after a river guide with data on kayaking sections on NZ rivers,
> including river rapids (names, grades). I also look after the Guide to the
> Whanganui River, and that has names for 239 rapids alone. I think OSM would
> be a good place to master this data and facilitate reuse elsewhere.
glad to hear it. OSM could be a wonderful common canvas/infrastructure
to hold that info, although I get a bit worried when the words
"master" and "canonical" are used, since OSM has no barrier to entry,
minimal quality control, and the last edit wins for better or ill. If
you can deal with the risk that some overly-helpful person on the
other side of the world might decide to change your data at any time
for any reason, then great. I don't mean to discourage you at all--
the vast majority of editors are here to create and make things better
and in practice it's not a big issue; just to avoid nasty surprises
later on since we can't control what other people do.
> http://linz2osm.openstreetmap.org.nz/data_dict/field_stats/mainland/rapid_poly/name/
> shows a handful of rapids, but there are hundreds out there.
Just to note that LINZ's rapid_poly was created with cartography in
mind, not river sports. That doesn't mean more shouldn't be labeled
though, just what you see there is likely to come from what's labeled
on the topo50 sheets.
> Also, on behalf of the New Zealand Alpine Club (NZAC) I've build
> ClimbNZ.org.nz that now has almost 8,000 climbing routes ranging from
> individual boulders to multi-pitch mountain routes. ClimbNZ has an
> extensive multiple-parent place hierarchy (e.g. Aoraki/Mt Cook (
> http://climbnz.org.nz/nz/si/aoraki/mt-cook-range/aoraki-mt-cook ) "is_in"
> Mt Cook Range and "is_near" Hooker Glacier, Grand Plateau.
for my 2c the "is_in" tags are most useful for administrative
distinctions, since the lat/lon of the object means the computer can
figure out what it's physically near to with a spatial query much
better than a human labeling everything can. I believe there are
"is_in:country=" and "is_in:district=" etc. tags commonly used for
islands.
> I'm interested in seeing if OSM can be the canonical repository for the
> spatial elements, with guide descriptions and photos being maintained in
> the ClimbNZ.org.nz website but referencing back to OSM for location.
sounds really great, but as before "canonical repository" wrt OSM
always makes me a bit nervous. NZ is pleasantly civil, but there are
jerks out there in the wider world..
> However, I'm not sure that OSM sufficiently maps the hierarchy of places
> necessary for ClimbNZ and similar places. I'll need to do some research to
> compare the hierarchy of both systems.
would you use OSM's search tools or copy/sync the OSM data locally
into a PostGIS database and do your own queries? If you are using
PostGIS db locally the world's your oyster.
> I'm interested in getting the backcountry hut & track information into OSM
> - perhaps I can assist the linz2osm effort there.
sure thing! the tracks are ready to go. the building_pnt LINZ layer
contains the huts, I don't think it is approved for uploads right now,
but I think that just needs minor review, it's 99.8% ready to go.
> If there's any Christchurch meetups, please let me know. I'd like to learn
> from more experienced people regarding workflows etc. esp. with respect to
> merging between incoming bulk data and existing OSM data.
a couple of simple hints: be respectful of others' time and effort,
and keep in mind the ultimate goal is making the map the best it can
be. sometimes that means the community is more important than the
edit, and other times the edit is more important than pissing someone
off; use your best judgment. I tend to put a lot of weight on if the
original uploader bothered to use a source= tag or not.
If you notice things are misaligned, and labeled with source=Bing
imagery (or can assume that), look for a nearby trig station in the
linz2osm Geodetic dataset and see if it is visible on the Bing
backdrop image in JOSM, if so use the imagery-align tool there to
correct the image alignment by nudging with the arrow keys, then save
the offset as a bookmark.
> A few specific questions to start with:
>
> a. http://linz2osm.openstreetmap.org.nz/data_dict/layer/rapid_poly/preview/
> appears
> to use <tag k="whitewater" v="rapids"/> but
> http://taginfo.openstreetmap.org/keys/whitewater#values suggests rapid
> (singular) is more common.
consider that whatever we decide to use will likely dominate future
taginfo results, and whatever some other bulk uploader decided to do
may have already dominated it. It doesn't necessarily mean a consensus
of OSM users agree with it.
I would guess that the LINZ rapids polygons more show sections of
rivers as opposed to individual runs, so plural would be more
appropriate, but as before, use your best judgment.
The tagwatch stats and osm wiki are just guidelines and should be
taken with grains of salt.
I'd note in our previous discussions about the rapids_poly linz layer
that a lot of the work figuring out what to do with whitewater on the
OSM wiki was done by Harry Wood, who's been around OSM forever and in
general probably thought the thing out better than your average wiki
proposal.
> b. If I've interpreted it correctly
> https://github.com/opennewzealand/linz2osm/blob/master/contrib/mp2osm_linz_jr_pl.py
> appears to map huts to tourism=alpine_hut
the mp2osm script for mapping garmin codes to OSM tags hasn't been
used in years, only the NZOpenGPS roads are being used with that. (in
the past-tense, Stephen has already processed them and loaded them
into the linz2osm web app db AFAIK)
but yes, we have been tagging huts as tourism=alpine_hut for the last
years, just that it's done as part of the building_pnt and
building_poly layers. (and maybe descriptive_text too)
> but OSM definition is food provided
> https://wiki.openstreetmap.org/wiki/Tag:tourism%3Dalpine_hut
> I think mapping to
> https://wiki.openstreetmap.org/wiki/Tag:tourism%3Dwilderness_hut is more
> appropriate.
if you look closely at the change history of the alpine_hut osm wiki
page, the wilderness_hut wiki page, and the proposal page (+
discussion tabs for each) you'll notice that the new wilderness_hut
tag and distinctions about when to use each is wholly the work of one
guy in the last months.
that's not to say that it is right or wrong, just to keep in mind that
the osm wiki is not gospel, it's whatever the last person to edit the
page wanted it to say. And it's likely the wiki author was thinking
about what was common meaning and practice in the Swiss Alps, not the
Southern ones or any other alpine area.
personally I'd be a lot more comfortable if the distinction had more
to do with it being, you know, *alpine*, and not if it has a hut
warden who served you food. See also long osm wiki (discussion tab)
pages on the definition of a chalet, and bivvys (which apparently mean
a completely different thing to the French). ...
As it is my 2c vote would be to let the new idea mature in the cellar
for a little while before jumping on board. I didn't check, but I
doubt the mapnik rendering gatekeepers have paid any attention to the
wilderness_hut idea yet. I grant you that apline_hut is a bit weird
for things non-alpine, but it has historical momentum behind it.
> 288 huts for NZ have been entered in OSM as tourism=alpine_hut in every
> case to date, but I think these don't qualify per the OSM description.
The wiki is a last-to-edit-the-page-wins set of guidelines, not strict
law. Take it all with a grain of salt. About the only thing in OSM
which isn't benevolent anarchy are the rules around copyright
assignment.
> If people agreed to the change,
(at this point, for my part I don't agree to change it, just to wait
for a while)
> how best to change a set of nodes from one tag to another?
If it were to be done the way to do it is to do a Xapi extract for
LINZ:source_version=* and tourism=alpine_hut in the NZ bounding box,
then manually change ones you know are truly backcountry, but not I
would guess something like the Aspiring "Hut". See hints about that on
the osm wiki LINZ pages.
hope it helps and you are most welcome,
Hamish
> By way of introduction, I want to understand OSM and its suitability for
> several outdoors datasets I am involved in. On behalf of Whitewater NZ, I
> look after a river guide with data on kayaking sections on NZ rivers,
> including river rapids (names, grades). I also look after the Guide to the
> Whanganui River, and that has names for 239 rapids alone. I think OSM would
> be a good place to master this data and facilitate reuse elsewhere.
glad to hear it. OSM could be a wonderful common canvas/infrastructure
to hold that info, although I get a bit worried when the words
"master" and "canonical" are used, since OSM has no barrier to entry,
minimal quality control, and the last edit wins for better or ill. If
you can deal with the risk that some overly-helpful person on the
other side of the world might decide to change your data at any time
for any reason, then great. I don't mean to discourage you at all--
the vast majority of editors are here to create and make things better
and in practice it's not a big issue; just to avoid nasty surprises
later on since we can't control what other people do.
> http://linz2osm.openstreetmap.org.nz/data_dict/field_stats/mainland/rapid_poly/name/
> shows a handful of rapids, but there are hundreds out there.
Just to note that LINZ's rapid_poly was created with cartography in
mind, not river sports. That doesn't mean more shouldn't be labeled
though, just what you see there is likely to come from what's labeled
on the topo50 sheets.
> Also, on behalf of the New Zealand Alpine Club (NZAC) I've build
> ClimbNZ.org.nz that now has almost 8,000 climbing routes ranging from
> individual boulders to multi-pitch mountain routes. ClimbNZ has an
> extensive multiple-parent place hierarchy (e.g. Aoraki/Mt Cook (
> http://climbnz.org.nz/nz/si/aoraki/mt-cook-range/aoraki-mt-cook ) "is_in"
> Mt Cook Range and "is_near" Hooker Glacier, Grand Plateau.
for my 2c the "is_in" tags are most useful for administrative
distinctions, since the lat/lon of the object means the computer can
figure out what it's physically near to with a spatial query much
better than a human labeling everything can. I believe there are
"is_in:country=" and "is_in:district=" etc. tags commonly used for
islands.
> I'm interested in seeing if OSM can be the canonical repository for the
> spatial elements, with guide descriptions and photos being maintained in
> the ClimbNZ.org.nz website but referencing back to OSM for location.
sounds really great, but as before "canonical repository" wrt OSM
always makes me a bit nervous. NZ is pleasantly civil, but there are
jerks out there in the wider world..
> However, I'm not sure that OSM sufficiently maps the hierarchy of places
> necessary for ClimbNZ and similar places. I'll need to do some research to
> compare the hierarchy of both systems.
would you use OSM's search tools or copy/sync the OSM data locally
into a PostGIS database and do your own queries? If you are using
PostGIS db locally the world's your oyster.
> I'm interested in getting the backcountry hut & track information into OSM
> - perhaps I can assist the linz2osm effort there.
sure thing! the tracks are ready to go. the building_pnt LINZ layer
contains the huts, I don't think it is approved for uploads right now,
but I think that just needs minor review, it's 99.8% ready to go.
> If there's any Christchurch meetups, please let me know. I'd like to learn
> from more experienced people regarding workflows etc. esp. with respect to
> merging between incoming bulk data and existing OSM data.
a couple of simple hints: be respectful of others' time and effort,
and keep in mind the ultimate goal is making the map the best it can
be. sometimes that means the community is more important than the
edit, and other times the edit is more important than pissing someone
off; use your best judgment. I tend to put a lot of weight on if the
original uploader bothered to use a source= tag or not.
If you notice things are misaligned, and labeled with source=Bing
imagery (or can assume that), look for a nearby trig station in the
linz2osm Geodetic dataset and see if it is visible on the Bing
backdrop image in JOSM, if so use the imagery-align tool there to
correct the image alignment by nudging with the arrow keys, then save
the offset as a bookmark.
> A few specific questions to start with:
>
> a. http://linz2osm.openstreetmap.org.nz/data_dict/layer/rapid_poly/preview/
> appears
> to use <tag k="whitewater" v="rapids"/> but
> http://taginfo.openstreetmap.org/keys/whitewater#values suggests rapid
> (singular) is more common.
consider that whatever we decide to use will likely dominate future
taginfo results, and whatever some other bulk uploader decided to do
may have already dominated it. It doesn't necessarily mean a consensus
of OSM users agree with it.
I would guess that the LINZ rapids polygons more show sections of
rivers as opposed to individual runs, so plural would be more
appropriate, but as before, use your best judgment.
The tagwatch stats and osm wiki are just guidelines and should be
taken with grains of salt.
I'd note in our previous discussions about the rapids_poly linz layer
that a lot of the work figuring out what to do with whitewater on the
OSM wiki was done by Harry Wood, who's been around OSM forever and in
general probably thought the thing out better than your average wiki
proposal.
> b. If I've interpreted it correctly
> https://github.com/opennewzealand/linz2osm/blob/master/contrib/mp2osm_linz_jr_pl.py
> appears to map huts to tourism=alpine_hut
the mp2osm script for mapping garmin codes to OSM tags hasn't been
used in years, only the NZOpenGPS roads are being used with that. (in
the past-tense, Stephen has already processed them and loaded them
into the linz2osm web app db AFAIK)
but yes, we have been tagging huts as tourism=alpine_hut for the last
years, just that it's done as part of the building_pnt and
building_poly layers. (and maybe descriptive_text too)
> but OSM definition is food provided
> https://wiki.openstreetmap.org/wiki/Tag:tourism%3Dalpine_hut
> I think mapping to
> https://wiki.openstreetmap.org/wiki/Tag:tourism%3Dwilderness_hut is more
> appropriate.
if you look closely at the change history of the alpine_hut osm wiki
page, the wilderness_hut wiki page, and the proposal page (+
discussion tabs for each) you'll notice that the new wilderness_hut
tag and distinctions about when to use each is wholly the work of one
guy in the last months.
that's not to say that it is right or wrong, just to keep in mind that
the osm wiki is not gospel, it's whatever the last person to edit the
page wanted it to say. And it's likely the wiki author was thinking
about what was common meaning and practice in the Swiss Alps, not the
Southern ones or any other alpine area.
personally I'd be a lot more comfortable if the distinction had more
to do with it being, you know, *alpine*, and not if it has a hut
warden who served you food. See also long osm wiki (discussion tab)
pages on the definition of a chalet, and bivvys (which apparently mean
a completely different thing to the French). ...
As it is my 2c vote would be to let the new idea mature in the cellar
for a little while before jumping on board. I didn't check, but I
doubt the mapnik rendering gatekeepers have paid any attention to the
wilderness_hut idea yet. I grant you that apline_hut is a bit weird
for things non-alpine, but it has historical momentum behind it.
> 288 huts for NZ have been entered in OSM as tourism=alpine_hut in every
> case to date, but I think these don't qualify per the OSM description.
The wiki is a last-to-edit-the-page-wins set of guidelines, not strict
law. Take it all with a grain of salt. About the only thing in OSM
which isn't benevolent anarchy are the rules around copyright
assignment.
> If people agreed to the change,
(at this point, for my part I don't agree to change it, just to wait
for a while)
> how best to change a set of nodes from one tag to another?
If it were to be done the way to do it is to do a Xapi extract for
LINZ:source_version=* and tourism=alpine_hut in the NZ bounding box,
then manually change ones you know are truly backcountry, but not I
would guess something like the Aspiring "Hut". See hints about that on
the osm wiki LINZ pages.
hope it helps and you are most welcome,
Hamish
Hamish
Jan 13, 2013, 3:46:16 PM1/13/13
to nzop...@googlegroups.com
I wrote:
> I'd note in our previous discussions about the rapids_poly linz layer
> that a lot of the work figuring out what to do with whitewater on the
> OSM wiki was done by Harry Wood, who's been around OSM forever and in
> general probably thought the thing out better than your average wiki proposal.
To clarify since I tend to leave some words out, "I'd note in our
> I'd note in our previous discussions about the rapids_poly linz layer
> that a lot of the work figuring out what to do with whitewater on the
> OSM wiki was done by Harry Wood, who's been around OSM forever and in
> general probably thought the thing out better than your average wiki proposal.
previous discussions about the rapids_poly linz layer" [that it was
noticed] "that a lot of the work figuring out what to do with
whitewater on the OSM wiki was done by Harry Wood".
(i.e. it sort of sounded like Harry was in on our discussions; he wasn't)
H
Stephen Davis
Jan 13, 2013, 3:59:43 PM1/13/13
to nzop...@googlegroups.com
> b. If I've interpreted it correctlythe mp2osm script for mapping garmin codes to OSM tags hasn't been
> https://github.com/opennewzealand/linz2osm/blob/master/contrib/mp2osm_linz_jr_pl.py
> appears to map huts to tourism=alpine_hut
used in years, only the NZOpenGPS roads are being used with that. (in
the past-tense, Stephen has already processed them and loaded them
into the linz2osm web app db AFAIK)
Pretty much. The script I used was https://github.com/opennewzealand/linz2osm/blob/master/scripts/mp2postgis.py which is based on that one, but it is converted to postgis rather than straight into a .osm file. The tagging can still be changed through linz2osm as no-one's done any of the NZOpenGPS roads yet. Speaking of which, I might as well upload a later version if and when anyone wants to tackle those.
Glen Barnes
Jan 13, 2013, 4:00:08 PM1/13/13
to nzop...@googlegroups.com
So mostly what Hamish said with some of my notes below ;-)
On Sunday, 13 January 2013 09:54:37 UTC+13, Jonathan Hunt wrote:
On Sunday, 13 January 2013 09:54:37 UTC+13, Jonathan Hunt wrote:
By way of introduction, I want to understand OSM and its suitability for several outdoors datasets I am involved in. On behalf of Whitewater NZ, I look after a river guide with data on kayaking sections on NZ rivers, including river rapids (names, grades). I also look after the Guide to the Whanganui River, and that has names for 239 rapids alone. I think OSM would be a good place to master this data and facilitate reuse elsewhere.
http://linz2osm.openstreetmap.org.nz/data_dict/field_stats/mainland/rapid_poly/name/ shows a handful of rapids, but there are hundreds out there.
Also, on behalf of the New Zealand Alpine Club (NZAC) I've build ClimbNZ.org.nz that now has almost 8,000 climbing routes ranging from individual boulders to multi-pitch mountain routes. ClimbNZ has an extensive multiple-parent place hierarchy (e.g. Aoraki/Mt Cook ( http://climbnz.org.nz/nz/si/aoraki/mt-cook-range/aoraki-mt-cook ) "is_in" Mt Cook Range and "is_near" Hooker Glacier, Grand Plateau.
Do you have these spatial mapped at the moment? i.e. do they have lat/lngs? If they do then we can import them into linz2osm and you can get going on merging this data into OSM. This would be a great additional resource for the map.
I'm interested in seeing if OSM can be the canonical repository for the spatial elements, with guide descriptions and photos being maintained in the ClimbNZ.org.nz website but referencing back to OSM for location.
I would say yes with the proviso that you have some form of checking scripts that look for any deleted data and new data so you can make sure things a kept in a tidy format.
However, I'm not sure that OSM sufficiently maps the hierarchy of places necessary for ClimbNZ and similar places. I'll need to do some research to compare the hierarchy of both systems.
That is only because there may not be the higher up regions mapped. There is no reason why we shouldn't add "Mount Cook" as a place (http://wiki.openstreetmap.org/wiki/Key:place) and then you can use spatial queries to work out the nesting. Also near is a spatial query kind of thing rather than a direct link.
A few specific questions to start with:
a. http://linz2osm.openstreetmap.org.nz/data_dict/layer/rapid_poly/preview/ appears to use <tag k="whitewater" v="rapids"/> but http://taginfo.openstreetmap.org/keys/whitewater#values suggests rapid (singular) is more common.
Rapids was for a section of river that contained multiple individual rapids as far as I am aware which is why we went for the plural. You seem to have far more in-depth data so I suggest we get the LINZ data in as a base and then bulk it out with what you have. There is no rule as to who wins (apart from on the ground confirmation).
288 huts for NZ have been entered in OSM as tourism=alpine_hut in every case to date, but I think these don't qualify per the OSM description. If people agreed to the change, how best to change a set of nodes from one tag to another?
I would stick with what we have for now - It is easy to change later if the general consensus changes over time. The tagging pages on the linz2osm tool are the goto place for how we are tagging things:
- http://linz2osm.openstreetmap.org.nz/data_dict/layer/building_poly/tagging/
- http://linz2osm.openstreetmap.org.nz/data_dict/layer/building_pnt/tagging/
Suggestions on updates to the tagging appreciated. We've tended to start a new thread for a particular layer as we go through the approval process. The building layers are probably the most complex one we have so it does need a good review.
An one note in terms of how to display this in a nice way on your site/guides. Check out TileMill - http://mapbox.com/tilemill/. You can take OSM data plus any of your own data and generate some nice looking maps that expose the climbing/rapid data in a way that works for your users.
Hamish
Jan 13, 2013, 4:15:09 PM1/13/13
to nzop...@googlegroups.com
a couple more quick points--
re. "is_in" and "is_near": This will all become a whole lot more
useful once the linz geographic_name layer gets uploaded!
re. rapids: to avoid double handling, probably the river_cl and
river_poly features for an area should be uploaded before doing too
much hand editing on the rapids. Currently the river_cl layer is in
limbo awaiting discussion on if the default should be waterway=river
or waterway=stream. (not to make my bias too clear but 95+% of the
centre lines are for steams :-)
Glen wrote:
> The building layers are probably the most complex one we have so it does need a
> good review.
fwiw there's already been a proposal/seconded/finalized pass:
http://wiki.openstreetmap.org/wiki/LINZ_attribute_matching#building_pnt
so I'd amend that to be "needs final review" to double check what has
changed since the previous review (which was based on the older V16
version of the data and things have drifted a bit since :).
regards,
Hamish
re. "is_in" and "is_near": This will all become a whole lot more
useful once the linz geographic_name layer gets uploaded!
re. rapids: to avoid double handling, probably the river_cl and
river_poly features for an area should be uploaded before doing too
much hand editing on the rapids. Currently the river_cl layer is in
limbo awaiting discussion on if the default should be waterway=river
or waterway=stream. (not to make my bias too clear but 95+% of the
centre lines are for steams :-)
Glen wrote:
> The building layers are probably the most complex one we have so it does need a
> good review.
http://wiki.openstreetmap.org/wiki/LINZ_attribute_matching#building_pnt
so I'd amend that to be "needs final review" to double check what has
changed since the previous review (which was based on the older V16
version of the data and things have drifted a bit since :).
regards,
Hamish
Robert Coup
Jan 14, 2013, 3:23:11 PM1/14/13
to nzop...@googlegroups.com
On Mon, Jan 14, 2013 at 10:15 AM, Hamish <hamish...@gmail.com> wrote:
Currently the river_cl layer is in
limbo awaiting discussion on if the default should be waterway=river
or waterway=stream. (not to make my bias too clear but 95+% of the
centre lines are for steams :-)
Can we just write a tagging rule that looks for a "River" in the name, then make the rest streams? Hacky, but will probably be correct...
Rob :)
Glen Barnes
Jan 14, 2013, 3:40:35 PM1/14/13
to nzop...@googlegroups.com
Nice idea. The Javascript is not strong in me. Care to add that rule?
Robert Coup
Jan 14, 2013, 3:49:10 PM1/14/13
to nzop...@googlegroups.com
On Tue, Jan 15, 2013 at 9:40 AM, Glen Barnes <barnacl...@gmail.com> wrote:
Nice idea. The Javascript is not strong in me. Care to add that rule?
/me naively assumed that the river_cl layer had names... "This object class does not have attributes." #fail.
Neither does river_poly.
Hamish
Jan 14, 2013, 4:05:08 PM1/14/13
to nzop...@googlegroups.com
Hamish wrote:
>> Currently the river_cl layer is in
>> limbo awaiting discussion on if the default should be waterway=river
>> or waterway=stream. (not to make my bias too clear but 95+% of the
>> centre lines are for steams :-)
Rob:
>> Currently the river_cl layer is in
>> limbo awaiting discussion on if the default should be waterway=river
>> or waterway=stream. (not to make my bias too clear but 95+% of the
>> centre lines are for steams :-)
> Can we just write a tagging rule that looks for a "River" in the name, then
> make the rest streams? Hacky, but will probably be correct...
It's not a bad idea, but apparently neither river_cl or river_poly
> make the rest streams? Hacky, but will probably be correct...
have name attributes.
(ISTR river_cl used to have them in earlier version of the linz dataset)
geographic_name has a STRM description code, but none for rivers, and
that will just be for named streams not the many smaller unnamed ones.
(?)
http://wiki.openstreetmap.org/wiki/LINZ_geo_name_matching#S-Z
So is our only source of river names from the descriptive_text point
layer?! (and then just hope that the nearest river_cl feature happens
to be the right one and not a nearby tributary feeding in)
Hamish
Jonathan Hunt
Jan 14, 2013, 5:53:04 PM1/14/13
to nzop...@googlegroups.com
On Monday, 14 January 2013 09:37:26 UTC+13, Hamish wrote:
Hi Jonathan,
> By way of introduction, I want to understand OSM and its suitability for
> several outdoors datasets I am involved in. On behalf of Whitewater NZ, I
> look after a river guide with data on kayaking sections on NZ rivers,
> including river rapids (names, grades). I also look after the Guide to the
> Whanganui River, and that has names for 239 rapids alone. I think OSM would
> be a good place to master this data and facilitate reuse elsewhere.
glad to hear it. OSM could be a wonderful common canvas/infrastructure
to hold that info, although I get a bit worried when the words
"master" and "canonical" are used, since OSM has no barrier to entry,
minimal quality control, and the last edit wins for better or ill. If
you can deal with the risk that some overly-helpful person on the
other side of the world might decide to change your data at any time
for any reason, then great. I don't mean to discourage you at all--
the vast majority of editors are here to create and make things better
and in practice it's not a big issue; just to avoid nasty surprises
later on since we can't control what other people do.
I guess my approach to that would be to mirror the relevant elements locally and take a regular update from OSM with some sanity checking like thresholds for edits and deletes. It looks like matching on OSM element id is necessary; if the ids change then local changes will be required.
I need to understand how to easily reverse a malicious edit, escalation procedures to block bad actors etc.
> http://linz2osm.openstreetmap.org.nz/data_dict/field_stats/mainland/rapid_poly/name/
> shows a handful of rapids, but there are hundreds out there.
Just to note that LINZ's rapid_poly was created with cartography in
mind, not river sports. That doesn't mean more shouldn't be labeled
though, just what you see there is likely to come from what's labeled
on the topo50 sheets.
I expect the bulk of rapids to be modelled as nodes, but some are long enough to be modelled as a polygon (e.g. Nevis Bluff on the Kawarau or Roaring Lion on the Karamea).
Given that there are only a few rapids in LINZ's data, is there any harm in adding individual rapids to OSM now and they will be merged when the LINZ upload occurs?
> Also, on behalf of the New Zealand Alpine Club (NZAC) I've build
> ClimbNZ.org.nz that now has almost 8,000 climbing routes ranging from
> individual boulders to multi-pitch mountain routes. ClimbNZ has an
> extensive multiple-parent place hierarchy (e.g. Aoraki/Mt Cook (
> http://climbnz.org.nz/nz/si/aoraki/mt-cook-range/aoraki-mt-cook ) "is_in"
> Mt Cook Range and "is_near" Hooker Glacier, Grand Plateau.
for my 2c the "is_in" tags are most useful for administrative
distinctions, since the lat/lon of the object means the computer can
figure out what it's physically near to with a spatial query much
better than a human labeling everything can. I believe there are
"is_in:country=" and "is_in:district=" etc. tags commonly used for
islands.
I need to wrap my head around spatial queries, e.g.
http://wiki.openstreetmap.org/wiki/Overpass_API/Language_Guide#Select_Region_by_Polygon
> I'm interested in seeing if OSM can be the canonical repository for the
> spatial elements, with guide descriptions and photos being maintained in
> the ClimbNZ.org.nz website but referencing back to OSM for location.
sounds really great, but as before "canonical repository" wrt OSM
always makes me a bit nervous. NZ is pleasantly civil, but there are
jerks out there in the wider world..
> However, I'm not sure that OSM sufficiently maps the hierarchy of places
> necessary for ClimbNZ and similar places. I'll need to do some research to
> compare the hierarchy of both systems.
would you use OSM's search tools or copy/sync the OSM data locally
into a PostGIS database and do your own queries? If you are using
PostGIS db locally the world's your oyster.
I think I'll be mirroring OSM content locally until I understand the dynamics of OSM better. If syncing is no big deal, then maintaining a local subset would provide a buffer to catch bad data (hopefully). It would be more elegant if we could rely purely on OSM (e.g. via Overpass API).
I'll follow up on huts in a separate thread.
> how best to change a set of nodes from one tag to another?
If it were to be done the way to do it is to do a Xapi extract for
LINZ:source_version=* and tourism=alpine_hut in the NZ bounding box,
then manually change ones you know are truly backcountry, but not I
would guess something like the Aspiring "Hut". See hints about that on
the osm wiki LINZ pages.
Can you specify the page you were thinking of?
hope it helps and you are most welcome,
Thanks for your assistance.
Regards
Jonathan
Jonathan Hunt
Jan 14, 2013, 6:01:46 PM1/14/13
to nzop...@googlegroups.com
On Monday, 14 January 2013 09:37:26 UTC+13, Hamish wrote:
Hi Jonathan,
Ok, I caught up with
http://wiki.openstreetmap.org/wiki/Proposed_features/wilderness_mountain_buildings
There's been a vote, so I assume that is as authoritative as OSM gets?
This defines hut types: Alpine_hut and Wilderness_hut approved, while tourism=Basic_hut rejected in favour of amenity=shelter, shelter_type=basic_hut.
Only a small fraction of NZ huts would qualify as Alpine (food provided). Most would be wilderness_hut and some (sans stove) as amenity=shelter, shelter_type=basic_hut.
It seems to me that http://linz2osm.openstreetmap.org.nz/data_dict/layer/building_pnt/tagging/ should be altered to map hut to "wilderness_hut".
Regards
Jonathan
Glen Barnes
Jan 14, 2013, 6:19:36 PM1/14/13
to nzop...@googlegroups.com
Are we asking the right questions....
Should we be looking for a better source of river data? Does NIWA have anything? DoC?
Hamish
Jan 14, 2013, 6:45:48 PM1/14/13
to nzop...@googlegroups.com
> Should we be looking for a better source of river data? Does NIWA have anything?
We should proceed with river_poly. As for river_cl for my 2c both LINZ
and OSM's targets are basically cartography, and the LINZ river_cl
lines up exactly with their lake outlets, cliff edges, etc, so I'm
mostly inclined to stick with it.
NIWA maintains its REC. I'm not sure about licence compatibility.
Hamish
> You received this message because you are subscribed to the Google Groups
> "nzopengis" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/nzopengis/-/s0hiVV2JRC4J.
> To post to this group, send email to nzop...@googlegroups.com.
> To unsubscribe from this group, send email to
> nzopengis+...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/nzopengis?hl=en.
>
>
Jonathan Hunt
Jan 14, 2013, 6:49:07 PM1/14/13
to nzop...@googlegroups.com
On Tuesday, 15 January 2013 12:19:36 UTC+13, Glen Barnes wrote:
Are we asking the right questions....
NIWA has the River Environment Classification (REC)
It's a segmented waterway model that could possibly be mapped to OSM. It would be a good basis since it's already a network (see the Tracer function) but I'm not sure it has names per LINZ.
Regards
Jonathan
Hamish
Jan 14, 2013, 7:16:03 PM1/14/13
to nzop...@googlegroups.com
While I think it is important to distinguish catered lodges (such as
found along the Great Walks) from back country huts, and in general I
fully agree the DOC huts are in general more "wilderness" than
"alpine", as a matter of process I'm still inclined to go slow and
ignore the issue for a while and stick with the existing tagging. It
is simple enough to retag them later as best we see fit. (note with
the Xapi-esque queries you can search for modification version equal
or not equal to 1, so quickly pull out the unaltered ones which you
can bulk modify without worry)
I would have much preferred if the proposal had not reused the already
widely deployed alpine_hut, and instead created new alpine_lodge and
wilderness_hut, and a suggestion that existing alpine_huts (of the
non-hut or non-alpine flavour) be migrated to either of the new names.
All of a sudden changing the definition of an already widely used name
just makes a big confusing mess IMO.
2c,
Hamish
found along the Great Walks) from back country huts, and in general I
fully agree the DOC huts are in general more "wilderness" than
"alpine", as a matter of process I'm still inclined to go slow and
ignore the issue for a while and stick with the existing tagging. It
is simple enough to retag them later as best we see fit. (note with
the Xapi-esque queries you can search for modification version equal
or not equal to 1, so quickly pull out the unaltered ones which you
can bulk modify without worry)
I would have much preferred if the proposal had not reused the already
widely deployed alpine_hut, and instead created new alpine_lodge and
wilderness_hut, and a suggestion that existing alpine_huts (of the
non-hut or non-alpine flavour) be migrated to either of the new names.
All of a sudden changing the definition of an already widely used name
just makes a big confusing mess IMO.
2c,
Hamish
> --
> You received this message because you are subscribed to the Google Groups
> "nzopengis" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/nzopengis/-/EITHaen-CWAJ.
> You received this message because you are subscribed to the Google Groups
> "nzopengis" group.
> To view this discussion on the web visit
Hamish
Jan 14, 2013, 7:29:50 PM1/14/13
to nzop...@googlegroups.com
(aka if you change it now all the huts will seem to disappear from the
osm.org map as the maintainers of the osm.org mapnik rendering rules
need to be convinced about it as well, and so far they haven't done
anything about supporting wilderness_hut. the votes on the wiki are
largely symbolic, and the wiki itself just a common framework to
coordinate guidelines)
regards,
Hamish
osm.org map as the maintainers of the osm.org mapnik rendering rules
need to be convinced about it as well, and so far they haven't done
anything about supporting wilderness_hut. the votes on the wiki are
largely symbolic, and the wiki itself just a common framework to
coordinate guidelines)
regards,
Hamish
Jonathan Hunt
Jan 14, 2013, 7:49:22 PM1/14/13
to nzop...@googlegroups.com
On Tuesday, 15 January 2013 12:45:48 UTC+13, Hamish wrote:
> Should we be looking for a better source of river data? Does NIWA have anything?
We should proceed with river_poly. As for river_cl for my 2c both LINZ
and OSM's targets are basically cartography, and the LINZ river_cl
lines up exactly with their lake outlets, cliff edges, etc, so I'm
mostly inclined to stick with it.
NIWA maintains its REC. I'm not sure about licence compatibility.
The REC is available at
under a CC-BY license.
The dataset includes an NZ Rivernames.csv file that maps names (e.g. Heaphy River) to reach ids.
Is there a way to match the river centerlines from LINZ to reaches in REC? If so, a ref could be added to OMS with the reach id, and the name could be pulled from Rivernames.csv
Jonathan Hunt
Jan 14, 2013, 7:58:32 PM1/14/13
to nzop...@googlegroups.com
The huts wiki page includes sample renderings for mapnik. If the maintainers for osm.org mapnik don't update based on votes, what does cause them to act/
I'm curious about the render on osm.org in any case. I looked at Aoraki Mt Cook and it looks poor, like islands in a wavy blue sea (Standard layer).
Only a few peaks show up, and no huts or tracks.
Does this mean that a site should always customise its rendering to suit its purposes?
Regards
Jonathan
Glen Barnes
Jan 14, 2013, 8:19:11 PM1/14/13
to nzop...@googlegroups.com
On Tuesday, 15 January 2013 13:58:32 UTC+13, Jonathan Hunt wrote:
Does this mean that a site should always customise its rendering to suit its purposes?
The tools for rendering are becoming a lot more usable and I'm really excited about what can be done with making really nice looking maps. So yes I would strongly consider making your own renderer (or looking for an existing alternative) if it suits especially for what you want. The default renderer isn't the nicest cartography in the world.
Hamish
Jan 14, 2013, 9:33:46 PM1/14/13
to nzop...@googlegroups.com
for custom rendering, tilemill is a nice place to start,
http://mapbox.com/tilemill/
http://adhoc.osgeo.osuosl.org/livedvd/docs/en/overview/tilemill_overview.html
http://adhoc.osgeo.osuosl.org/livedvd/docs/en/quickstart/tilemill_quickstart.html
the wavy lines you see on mt. cook are the glaciers. (tick the Browse
map data box on the osm.org page top right menu) draping over a shaded
relief map would make that a lot more obvious (e.g. in QGIS with the
LINZ LDS WMS shaded relief layer loaded and the OSM plugin rendering
the OSM data)
by all means please help with the tracks and mountain peaks situation :)
probably a proven track record of people using the thing in a rational
way. they have a million and one people constantly yelling at them
wanting their pet thing rendered this way or that, over time I think
they've developed a sort of immunity to it. Currently there are just
225 wilderness_huts tagged in osm; still in the noise level. Give it
time..
I sure do wish they'd fix the show-all-airstrips no-matter-how-small
bug at zoom level 10 though.
From an end user's perspective It's a pity that the live Osmarender
rendering went away, it was much less conservative and you could see a
bunch of different objects that don't turn up in the Mapnik one.
best,
Hamish
http://mapbox.com/tilemill/
http://adhoc.osgeo.osuosl.org/livedvd/docs/en/overview/tilemill_overview.html
http://adhoc.osgeo.osuosl.org/livedvd/docs/en/quickstart/tilemill_quickstart.html
the wavy lines you see on mt. cook are the glaciers. (tick the Browse
map data box on the osm.org page top right menu) draping over a shaded
relief map would make that a lot more obvious (e.g. in QGIS with the
LINZ LDS WMS shaded relief layer loaded and the OSM plugin rendering
the OSM data)
by all means please help with the tracks and mountain peaks situation :)
probably a proven track record of people using the thing in a rational
way. they have a million and one people constantly yelling at them
wanting their pet thing rendered this way or that, over time I think
they've developed a sort of immunity to it. Currently there are just
225 wilderness_huts tagged in osm; still in the noise level. Give it
time..
I sure do wish they'd fix the show-all-airstrips no-matter-how-small
bug at zoom level 10 though.
From an end user's perspective It's a pity that the live Osmarender
rendering went away, it was much less conservative and you could see a
bunch of different objects that don't turn up in the Mapnik one.
best,
Hamish
> --
> You received this message because you are subscribed to the Google Groups
> "nzopengis" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/nzopengis/-/ltkeziWUizMJ.
> You received this message because you are subscribed to the Google Groups
> "nzopengis" group.
> To view this discussion on the web visit
Reply all
Reply to author
Forward
0 new messages