Re: [Tagging] Handle with care

[Martin Koppenhoefer]

sent from a phone

> Am 27.09.2015 um 12:14 schrieb moltonel 3x Combo :
> 
> It's always
> source:foo, not foo:source.


you can find both but the former is more common: 
http://taginfo.openstreetmap.org/search?q=source


cheers 
Martin 
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Dave Swarthout]

Semantically the former makes more sense and as a plus, groups all sources
under one tag:

source:name  specifies the source of the name tag value
source:ref   specifies the source of the ref tag value
source:foo  specifies the source of the foo tag value

On Sun, Sep 27, 2015 at 7:18 PM, Martin Koppenhoefer  wrote:

>
>
> sent from a phone
>
> > Am 27.09.2015 um 12:14 schrieb moltonel 3x Combo :
> >
> > It's always
> > source:foo, not foo:source.
>
>
> you can find both but the former is more common:
> http://taginfo.openstreetmap.org/search?q=source
>
>
> cheers
> Martin
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>



-- 
Dave Swarthout
Homer, Alaska
Chiang Mai, Thailand
Travel Blog at http://dswarthout.blogspot.com
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Marc Gemis]

My fear is that some overzealous mappers will start adding those tags to
all objects in their neighborhood, just to "protect' their area and scare
away newbies.

Since we suppose that all data is mean to be correct and everybody makes
edits to improve the map, I do not see a good reason for such a tag.

regards

m


On Sat, Sep 26, 2015 at 10:43 PM, moltonel  wrote:

>
>
> On 26 September 2015 19:05:09 GMT+01:00, "André Pirard" <
> a.pirard.pa...@gmail.com> wrote:
> >It is
> >:warning=
> >which acts only when that key is changed.
> >geometry:warning=  to protect the coordinates of the element
> >name:warning=  to protect its name.
> >Those tags do not warn against changing other tags.
>
> May I suggest 'edit_warning', or something else that explicit (because a
> 'warning' key could be used for so many purposes) ? And use proper
> namespace ordering, ie edit_warning:name=blah rather than
> name:edit_warning=blah (because many keys, like name or phone, are also
> namespaces that can be followed by arbitrary sufixes).
>
> That said, you should open bugs on the various editors as soon as possible
> to discuss with them what they think of such a feature. Unless this tag
> gets editor support, it doesn't bring anything that the already popular
> 'note' tag doesn't give. Sadly, inexperienced mapers are the ones most
> likely to miss a plain 'note' tag, but they are also most likely using iD,
> whose developers are pretty warning-averse...
> --
> Vincent Dp
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[moltonel 3x Combo]

On 27/09/2015, André Pirard  wrote:
> But I'm afraid that the correct namespace order is name:edit_warning=*.
> edit_warning is a qualifier of name and not the opposite.
> It is the edit warning of (for) name and not the name of the edit warning.
> It's just like the order of the words in an English phrase.

I don't see the warning tag as a qualifier of the name, but as
metadata describing the name. As for the "English sentence" argument,
it's very easy and natural to phrase it with either 'name' first or
'waring' first.

> Adding a qualifier makes the word more specific, "with warning" rather
> than plain.
> If you speak of the building antenna type, it's building:antenna:4G=yes
> The left side of the road is road:left and not left:road.

Real-world objects have a left and a right, an antenna that is 4G or
not. They do not have a "warning: location is very precise" or
"warning: sat imagery is outdated". The warnings are not giving a more
specific description of the object, they exist outside the object.

> And BTW, source:maxspeed is a mistake, there is no such thing as the
> speed of a source
> The source of the maxspeed should be maxspeed:source.
> And its date should be maxspeed:source:date and not
> source:date:maxspeed or source:maxspeed:date or date:maxspeed:source..

Thank you for providing an example of a tag that is 'meta' just like
'warning' and correctly lives in its own namespace. It's always
source:foo, not foo:source.


So we've looked at the semantic, linguistic, and current practice
arguments. But here comes the technical one, based on a golden rule of
osm tag-crafting, "New tags should not break consumers that do not
know about the tag" :

Without knowing about these "warning" tags, I'll encounter them as I
read namespaces left to right. name:warning ? Huh, not sure what it
means, but the name tag is so complicated that I just add all of them
as alternative names anyway. phone:warning ? Ohhh I know about the
phone:foo scheme, it means that this phone "number" is the one to use
for warning purposes. etc...If you don't put warning in its own
top-level namespace, it's going to show up in a lot of places where it
shouldn't.

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[moltonel 3x Combo]

On 27/09/2015, Marc Gemis  wrote:
> My fear is that some overzealous mappers will start adding those tags to
> all objects in their neighborhood, just to "protect' their area and scare
> away newbies.
>
> Since we suppose that all data is mean to be correct and everybody makes
> edits to improve the map, I do not see a good reason for such a tag.

To be honest I'm not a fan of the idea either, but it's comming back
so regularly that I think we should treat it in some way.

Bryan proved me wrong by saying that he (as an iD developer) would
support some form of warning, and he makes the very interesting point
that values should be standardised (this begs the question of what to
do when an unknown values comes up, but hey).

Standard names offer the opportunity to guide which type of warnings
are good to have and not oversteping. And can link to help text that
can try its best to explain rather than scare away. Picking up on
Bryan's examples:
* outdated_imagery should be self-explanatory, I've used the note tag
for that a few times
* border_dispute can link to the many ressources explaining how to
handle these nicely
* current_event is... the same as outdated_imagery ? unless a precise
date is given ?
* authoritative_data can suggest mapers to alert the authority that
their data is crap ;)
* i_know_the_place_better_than_you is notably absent

Bryan also distinguished 'warning' (which just pop up a message but
lets the user continue) from 'restriction' (which don't let the user
edit until he removes the restriction tag). I think that 'restriction'
goes too far, an editor warning should be enough for all cases (as
long as it is visible enough).

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[André Pirard]

On 2015-09-24 17:32, Kotya Karapetyan wrote :
> Hi André, all,
>
> Shall we discuss an "object_warning" tag? To begin with, it will
> simply contain information. Editors can also choose to show it when
> the tagged object is about to be changed.
Gladly (of course), I suppose that all those discussions were meant to
come to a concrete result.
But beware that it is not "object_warning" that seems to protect the
whole element.
It is
:warning=
which acts only when that key is changed.
geometry:warning=  to protect the coordinates of the element
name:warning=  to protect its name.
Those tags do not warn against changing other tags.

But it's a good implicit remark that the whole object could be
protected, including from deletion.
warning=

"about to be changed" is not "about to be uploaded" (needing to identify
several elements" but when the user clicks for changing (normally) one
element.

Cheers

André.





>
> Kind regards,
> Kotya
>
>
>
>
>
> On Fri, Sep 18, 2015 at 12:53 AM, André Pirard
> > wrote:
>
> On 2015-09-17 18:02, Kotya Karapetyan wrote :
>> Hi André,
>>
>> I don't know why your text was removed. 
>>
>> > It would produce a message saying something like:  
>> > "The coordinates you are trying to change are accurate to 25 cm.  
>> > You probably shouldn't change this tag, certainly not with GPS data.  
>> > Are you certain that you will not destroy valuable data and do you 
>> want to continue?".
>> > And if he replies "no", his attempt is canceled.
>>
>> I like this approach. I wonder if it is technically feasible.
> Forget about my bad examples and the eagerness to pick them.
> Here is the original text.
>>> ... Despite a "don't touch" note explaining why not, a good soul
>>> passes, not reading note and makes a "correction".
>>> What is needed here is an "are you sure?" tag named such as 
>>> [keyname:]warning="text" that the map editing softwate  uses any
>>> time a mapper wants to change that keyname's value  to display
>>> the message and ask for a confirmation (by the tag, at the time
>>> he tries to change it, not when he tries to upload a dozen of
>>> such changes).
>>> ="Reasons why you shouldn't change that tag.  Do you
>>> really want to change it?"
>>> Replying "no" cancels the attempt.
>>> Or should it be [keyname:]note:warn="text" and spare another
>>> wiki page?
>>> keyname can be "geometry" as in source:geometry.
>>> Et voilà.  An all-purpose simple guardrail, a small update to
>>> the wiki and passing the word to the editors.
>>
>> My point was that to make it generic may be more difficult than
>> creating a very specific tag/function for survey-based data.
> IMHO it may be simpler that some specific implementations and
> certainly when their numbers reaches 2.
> The answer will be given by JOSM et al.
> It doesn't address "mechanical" updates, but the persons doing
> them are supposed to know what they're doing, aren't they?
>> And I didn't understand the benefit for your other examples. But
>> otherwise I support it.
> Those examples forgotten, other voices are needed, the wiki update
> has almost been written.
>>
>> Cheers,
>> Kotya
> General tip: Kotya, do you know that you can have your
> kotya.li...@gmail.com  account use
> filters to store messages in by-the-list folders and access those
> folders using IMAP with software like Thunderbird and do things
> like answering to ancient mail?
>
> Cheers
>
> André.
>
>
>

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[moltonel]

On 26 September 2015 19:05:09 GMT+01:00, "André Pirard" 
 wrote:
>It is
>:warning=
>which acts only when that key is changed.
>geometry:warning=  to protect the coordinates of the element
>name:warning=  to protect its name.
>Those tags do not warn against changing other tags.

May I suggest 'edit_warning', or something else that explicit (because a 
'warning' key could be used for so many purposes) ? And use proper namespace 
ordering, ie edit_warning:name=blah rather than name:edit_warning=blah (because 
many keys, like name or phone, are also namespaces that can be followed by 
arbitrary sufixes).

That said, you should open bugs on the various editors as soon as possible to 
discuss with them what they think of such a feature. Unless this tag gets 
editor support, it doesn't bring anything that the already popular 'note' tag 
doesn't give. Sadly, inexperienced mapers are the ones most likely to miss a 
plain 'note' tag, but they are also most likely using iD, whose developers are 
pretty warning-averse...
-- 
Vincent Dp

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[André Pirard]

On 2015-09-26 22:43, moltonel wrote :
>
> On 26 September 2015 19:05:09 GMT+01:00, "André Pirard" 
>  wrote:
>> It is
>> :warning=
>> which acts only when that key is changed.
>> geometry:warning=  to protect the coordinates of the element
>> name:warning=  to protect its name.
>> Those tags do not warn against changing other tags.
> May I suggest 'edit_warning', or something else that explicit (because a 
> 'warning' key could be used for so many purposes) ? And use proper namespace 
> ordering, ie edit_warning:name=blah rather than name:edit_warning=blah 
> (because many keys, like name or phone, are also namespaces that can be 
> followed by arbitrary sufixes).
OK with me for a more explicit term as long as it is not too long.  The
best inventor will win.
But I'm afraid that the correct namespace order is name:edit_warning=*.
edit_warning is a qualifier of name and not the opposite.
It is the edit warning of (for) name and not the name of the edit warning.
It's just like the order of the words in an English phrase.
Adding a qualifier makes the word more specific, "with warning" rather
than plain.
If you speak of the building antenna type, it's building:antenna:4G=yes
The left side of the road is road:left and not left:road.
And BTW, source:maxspeed is a mistake, there is no such thing as the
speed of a source
The source of the maxspeed should be maxspeed:source.
And its date should be maxspeed:source:date and not
source:date:maxspeed or source:maxspeed:date or date:maxspeed:source..
But it's typical of an OSM article about namespaces not to speak of the
order.
Let us stop making more mistake.
> That said, you should open bugs on the various editors as soon as possible to 
> discuss with them what they think of such a feature. Unless this tag gets 
> editor support, it doesn't bring anything that the already popular 'note' tag 
> doesn't give. Sadly, inexperienced mapers are the ones most likely to miss a 
> plain 'note' tag, but they are also most likely using iD, whose developers 
> are pretty warning-averse...
Don't speak like that of mappers and developers ;-)  You're going to get
in trouble with Julien Fastré who will say that you're trying to
convince us, and yourself, that "they are doing bad job" and that you
will frighten newcomers (sic) ;-)
This said, seriously and frankly, I have a serious dilemma when I'm
hippy hopping the JOSM Area Selection Plugin along the way and I meet
polygons completely askew and 4-5 m or more away from their place with
just "building=yes" and that in one single click I can draw the house
with a 20 cm precision and complete tags, the house number being
incremented for the next house !  I feel like a 5 years' work has to be
restarted afresh.  Fortunately, not afresh for fixing road geometry in
Improve Way Accuracy mode of that JOSM again.
Isn't there a convincing tutorial to compare methods and show that (fixing)?
I know two friends who are ¾ convinced but who can't get rid of their
habits and make the first step.
I myself know too little of the other products to be a good comparator
and convincing advisor.
All that I can say is that a good migration method is to use the old and
the new programs simultaneously, more and more of the new one.  It's
what I did when simulating a shift from JOSM to Merkaartor, but it
turned out that there were too many things that Merkaartor couldn't do.

Cheers

André.



___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Bryan Housel]

Hi, this issue does pop up on the mailing lists sometimes.
Here are my thoughts from the last time:  
https://lists.openstreetmap.org/pipermail/tagging/2015-June/025438.html 


To recap, here is what I would and would not support in iD:


- I would not support arbitrary free-text warning messages popping up as users 
try to edit things.  This is too susceptible for abuse.


- I would support a standardized tag scheme that lets us warn a mapper about 
conditions that anybody would understand.  These warnings would have very 
narrow scope and attach to areas.  We would translate them to a user’s native 
language.
e.g. 
`editor:warning=outdated_imagery`
  “Imagery in this area may be outdated. Please do not make edits without local 
knowledge."
`editor:warning=border_dispute`
  “There is an ongoing border dispute in this area. Please do not make edits 
without local knowledge."
`editor:warning=current_event`
  “A current event may have affected ground conditions in this area. Please do 
not make edits without local knowledge."


- I would support a standardized tag scheme that lets us restrict certain 
edits.  
(with the caveat that a determined user could remove the tag and edit the thing 
anyway).
e.g. 
`editor:restriction=authoritative_data`
`editor:restriction:source=nyc.gov`
   “You may not move this item because its position has been surveyed with high 
accuracy."


- I would support API-based restrictions (if the OSM API supported such a 
thing, which it doesn't), e.g. to counter abuse, edit wars, vandalism.  We 
already poll the `/api/capabilities` endpoint every few minutes, so maybe it 
makes sense to stick API-based editing restrictions in there.
e.g.
   “Editing in this location is currently unavailable due to high amounts of 
vandalism."


Anyway feel free to discuss here, or on the iD issue list:   :-)
https://github.com/openstreetmap/iD/issues 



Thanks, Bryan




> On Sep 26, 2015, at 4:43 PM, moltonel  wrote:
> 
> 
> 
> On 26 September 2015 19:05:09 GMT+01:00, "André Pirard" 
>  wrote:
>> It is
>> :warning=
>> which acts only when that key is changed.
>> geometry:warning=  to protect the coordinates of the element
>> name:warning=  to protect its name.
>> Those tags do not warn against changing other tags.
> 
> May I suggest 'edit_warning', or something else that explicit (because a 
> 'warning' key could be used for so many purposes) ? And use proper namespace 
> ordering, ie edit_warning:name=blah rather than name:edit_warning=blah 
> (because many keys, like name or phone, are also namespaces that can be 
> followed by arbitrary sufixes).
> 
> That said, you should open bugs on the various editors as soon as possible to 
> discuss with them what they think of such a feature. Unless this tag gets 
> editor support, it doesn't bring anything that the already popular 'note' tag 
> doesn't give. Sadly, inexperienced mapers are the ones most likely to miss a 
> plain 'note' tag, but they are also most likely using iD, whose developers 
> are pretty warning-averse...
> -- 
> Vincent Dp
> 
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Kotya Karapetyan]

Hi André, all,

Shall we discuss an "object_warning" tag? To begin with, it will simply
contain information. Editors can also choose to show it when the tagged
object is about to be changed.

Kind regards,
Kotya





On Fri, Sep 18, 2015 at 12:53 AM, André Pirard 
wrote:

> On 2015-09-17 18:02, Kotya Karapetyan wrote :
>
> Hi André,
>
> I don't know why your text was removed.
>
> > It would produce a message saying something like:
> > "The coordinates you are trying to change are accurate to 25 cm.
> > You probably shouldn't change this tag, certainly not with GPS data.
> > Are you certain that you will not destroy valuable data and do you want
> to continue?".
> > And if he replies "no", his attempt is canceled.
>
> I like this approach. I wonder if it is technically feasible.
>
> Forget about my bad examples and the eagerness to pick them.
> Here is the original text.
>
> ... Despite a "don't touch" note explaining why not, a good soul passes,
> not reading note and makes a "correction".
> What is needed here is an "are you sure?" tag named such as
> [keyname:]warning="text" that the map editing softwate  uses any time a
> mapper wants to change that keyname's value  to display the message and ask
> for a confirmation (by the tag, at the time he tries to change it, not when
> he tries to upload a dozen of such changes).
> ="Reasons why you shouldn't change that tag.  Do you really want to
> change it?"
> Replying "no" cancels the attempt.
> Or should it be [keyname:]note:warn="text" and spare another wiki page?
> keyname can be "geometry" as in source:geometry.
> Et voilà.  An all-purpose simple guardrail, a small update to the wiki and
> passing the word to the editors.
>
>
> My point was that to make it generic may be more difficult than creating a
> very specific tag/function for survey-based data.
>
> IMHO it may be simpler that some specific implementations and certainly
> when their numbers reaches 2.
> The answer will be given by JOSM et al.
> It doesn't address "mechanical" updates, but the persons doing them are
> supposed to know what they're doing, aren't they?
>
> And I didn't understand the benefit for your other examples. But otherwise
> I support it.
>
> Those examples forgotten, other voices are needed, the wiki update has
> almost been written.
>
>
> Cheers,
> Kotya
>
> General tip: Kotya, do you know that you can have your
> kotya.li...@gmail.com account use filters to store messages in
> by-the-list folders and access those folders using IMAP with software like
> Thunderbird and do things like answering to ancient mail?
>
> Cheers
>
> André.
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[Kotya Karapetyan]

Hi André,

I don't know why your text was removed.

> It would produce a message saying something like:
> "The coordinates you are trying to change are accurate to 25 cm.
> You probably shouldn't change this tag, certainly not with GPS data.
> Are you certain that you will not destroy valuable data and do you want
to continue?".
> And if he replies "no", his attempt is canceled.

I like this approach. I wonder if it is technically feasible.

My point was that to make it generic may be more difficult than creating a
very specific tag/function for survey-based data.
And I didn't understand the benefit for your other examples. But otherwise
I support it.

Cheers,
Kotya




On Fri, Sep 11, 2015 at 7:16 PM, André Pirard 
wrote:

> On 2015-09-09 23:39, moltonel wrote :
>
> On 9 September 2015 21:46:54 GMT+01:00, "André Pirard" 
>   wrote:
>
> There are various reasons for warning other mappers to be careful about
> their updates.
> I once temporarily overlaid two walking routes to show the effect of
> displaying two sorts of icons.
> Or I left in for a while drawing errors of a plugin as the best way to
> show the author what I talk about.
> Despite a don't touch note explaining why, a good soul passes, not
> reading note and makes a "correction".
>
> Please run experiments like this on a test db, not on the main one. It's easy 
> to point your editor to dev.openstreetmap.org for example (quoting from 
> memory, not 100% sure). You never know when a data consumer will stumble upon 
> your experiment, live or in a downloaded snapshot. Nobody expects osm data to 
> be perfect all the time, but there's no point in knowingly making it worse.
>
> You are off topic, as well as the following messages.
> While I admit that my examples are suboptimal, the matter is extending
> very simply to other tags the idea of preventing to replace precisely
> triangulated coordinates by loose GPS ones.
> Let us, for a better example, say that someone tagged a strange looking
> name and that he knows for sure that the spelling is correct.  After the
> third time the name was changed to a apparently better but wrong spelling,
> he will want to enforce reading the note that nobody reads. That's all
> there is to the suggestion you removed from this message.
>
> Now responding to your accusations.
> What big sin is that to discover errors and leave them a few more days on
> the map for the developer of the tool that produced them to have a look at
> them?  Is there a prescribed time limit?
> Like you, I have always advocated a sandbox, especially for helping
> novices. I've never heard of one and it's the first time I do. But JOSM
> says "The server responds with the return code 404 instead of 200. " when
> trying to validate http://dev.openstreetmap.org/ as well as
> http://api06.dev.openstreetmap.org/
> Thanks, but please give correct information.
> But a sandbox wouldn't help with the first bad example because it's to be
> looked at on Waymarked trails and that program does not display sandbox
> data.  And as we're told that those URL if they worked wouldn't have a
> renderer, they wouldn't be very convenient to use.
> Please make practical suggestions !!!
>
> On 2015-09-10 18:41, Kotya Karapetyan wrote :
>
> But otherwise I think there is a difference between a general warning or
> message from one mapper to another (which in its own is an interesting idea
> but can lead to dialogues and discussions) and a specific technical feature
> that would prevent moving an accurately positioned tag.
>
> Imagine there is a real-world marker at 50.000° N:
> http://www.dieweltenbummler.de/geografisches/geografische-besonderheiten/50-breitengrad/
> Someone draws it in OSM at 50° N. Then I come there with a smartphone,
> measure the location, find it at 49.9° and edit the OSM accordingly. It is
> wrong by definition (providing that the real-world marker location is known
> precisely), but there is no mechanism to prevent such editing.
>
> I think it's a very specific and relevant gap, and would love seeing it
> solved elegantly.
>
> That is what my suggestion does and I wonder why the heck it has been
> removed from this message !!!
> It would produce a message saying something like:  "The coordinates you
> are trying to change are accurate to 25 cm.  You probably shouldn't change
> this tag, certainly not with GPS data.  Are you certain that you will not
> destroy valuable data and do you want to continue?".
> And if he replies "no", his attempt is canceled.
> This kind of message would be possible for any tag and I don't understand
> why you want it to be specific.
> What elegance does it lack? You should explain. It allows to make updates
> to better that 25 cm.
> We know that it's typical of OSM to be crowded with stupid vandalism ("can
> I erase what I don't understand?"), even the DWG changing names to spelling
> mistakes i could explain, but someone defying such 

Re: [Tagging] Handle with care

[André Pirard]

On 2015-09-17 18:02, Kotya Karapetyan wrote :
> Hi André,
>
> I don't know why your text was removed. 
>
> > It would produce a message saying something like:  
> > "The coordinates you are trying to change are accurate to 25 cm.  
> > You probably shouldn't change this tag, certainly not with GPS data.  
> > Are you certain that you will not destroy valuable data and do you
> want to continue?".
> > And if he replies "no", his attempt is canceled.
>
> I like this approach. I wonder if it is technically feasible.
Forget about my bad examples and the eagerness to pick them.
Here is the original text.
>> ... Despite a "don't touch" note explaining why not, a good soul
>> passes, not reading note and makes a "correction".
>> What is needed here is an "are you sure?" tag named such as 
>> [keyname:]warning="text" that the map editing softwate  uses any time
>> a mapper wants to change that keyname's value  to display the message
>> and ask for a confirmation (by the tag, at the time he tries to
>> change it, not when he tries to upload a dozen of such changes).
>> ="Reasons why you shouldn't change that tag.  Do you really
>> want to change it?"
>> Replying "no" cancels the attempt.
>> Or should it be [keyname:]note:warn="text" and spare another wiki page?
>> keyname can be "geometry" as in source:geometry.
>> Et voilà.  An all-purpose simple guardrail, a small update to the
>> wiki and passing the word to the editors.
>
> My point was that to make it generic may be more difficult than
> creating a very specific tag/function for survey-based data.
IMHO it may be simpler that some specific implementations and certainly
when their numbers reaches 2.
The answer will be given by JOSM et al.
It doesn't address "mechanical" updates, but the persons doing them are
supposed to know what they're doing, aren't they?
> And I didn't understand the benefit for your other examples. But
> otherwise I support it.
Those examples forgotten, other voices are needed, the wiki update has
almost been written.
>
> Cheers,
> Kotya
General tip: Kotya, do you know that you can have your
kotya.li...@gmail.com account use filters to store messages in
by-the-list folders and access those folders using IMAP with software
like Thunderbird and do things like answering to ancient mail?

Cheers

André.


___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[Ruben Maes]

Friday 11 September 2015 19:16:09, André Pirard:
> But JOSM says "The server responds with the return code 404 instead of 200. " 
> when trying to validate http://dev.openstreetmap.org/ as well as 
> http://api06.dev.openstreetmap.org/
> Thanks, but please give correct information.

Hi André

The exact URL you need to enter is:
http://api06.dev.openstreetmap.org/api

To edit the map you'll have to create another account at 
http://api06.dev.openstreetmap.org/.

-- 
The field "from" of an email is about as reliable as the address written on the 
back of an envelope. That's why this message is OpenPGP signed.

signature.asc
Description: This is a digitally signed message part.
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[André Pirard]

  
  
On 2015-09-09 23:39, moltonel wrote :



  On 9 September 2015 21:46:54 GMT+01:00, "André Pirard"  wrote:

  
There are various reasons for warning other mappers to be careful about
their updates.
I once temporarily overlaid two walking routes to show the effect of
displaying two sorts of icons.
Or I left in for a while drawing errors of a plugin as the best way to
show the author what I talk about.
Despite a don't touch note explaining why, a good soul passes, not
reading note and makes a "correction".

  
  
Please run experiments like this on a test db, not on the main one. It's easy to point your editor to dev.openstreetmap.org for example (quoting from memory, not 100% sure). You never know when a data consumer will stumble upon your experiment, live or in a downloaded snapshot. Nobody expects osm data to be perfect all the time, but there's no point in knowingly making it worse.


You are off topic, as well as the following messages.
While I admit that my examples are suboptimal, the matter is
extending very simply to other tags the idea of preventing to
replace precisely triangulated coordinates by loose GPS ones.
Let us, for a better example, say that someone tagged a strange
looking name and that he knows for sure that the spelling is
correct.  After the third time the name was changed to a apparently
better but wrong spelling, he will want to enforce reading the note
that nobody reads. That's all there is to the suggestion you removed
from this message.

Now responding to your accusations.
What big sin is that to discover errors and leave them a few more
days on the map for the developer of the tool that produced them to
have a look at them?  Is there a prescribed time limit?
Like you, I have always advocated a sandbox, especially for helping
novices. I've never heard of one and it's the first time I do. But
JOSM says "The server responds with the return code 404 instead of
200. " when trying to validate http://dev.openstreetmap.org/
as well as http://api06.dev.openstreetmap.org/
Thanks, but please give correct information.
But a sandbox wouldn't help with the first bad example because it's
to be looked at on Waymarked trails and that program does not
display sandbox data.  And as we're told that those URL if they
worked wouldn't have a renderer, they wouldn't be very convenient to
use.
Please make practical suggestions !!!

On 2015-09-10 18:41, Kotya Karapetyan wrote :

  But
  otherwise I think there is a difference between a general
  warning or message from one mapper to another (which in its
  own is an interesting idea but can lead to dialogues and
  discussions) and a specific technical feature that would
  prevent moving an accurately positioned tag. 
  

  Imagine
  there is a real-world marker at 50.000° N: http://www.dieweltenbummler.de/geografisches/geografische-besonderheiten/50-breitengrad/
  Someone draws it in OSM at 50°
  N. Then I come there with a smartphone, measure the location,
  find it at 49.9° and edit the OSM accordingly. It is wrong by
  definition (providing that the real-world marker location is
  known precisely), but there is no mechanism to prevent such
  editing.
  

  I think it's a very specific
  and relevant gap, and would love seeing it solved elegantly.

That is what my suggestion does and I wonder why the heck it has
been removed from this message !!!
It would produce a message saying something like:  "The coordinates
you are trying to change are accurate to 25 cm.  You probably
shouldn't change this tag, certainly not with GPS data.  Are you
certain that you will not destroy valuable data and do you want to
continue?".
And if he replies "no", his attempt is canceled.
This kind of message would be possible for any tag and I don't
understand why you want it to be specific.
What elegance does it lack? You should explain. It allows to make
updates to better that 25 cm.
We know that it's typical of OSM to be crowded with stupid vandalism
("can I erase what I don't understand?"), even the DWG changing
names to spelling mistakes i could explain, but someone defying such
a warning should be banned for life.
Why was my text removed?

Cheers



  

  André.

  


  


___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Colin Smale]

 

I meant my question seriously, not hypothetically I assume all these
boundary posts are tagged with something like "source:location=MA State
Data Set 2015-01-19"? If not, how is a mapper to compare his
"correctness" to the existing "correctness"? 

Without this provenance information in the database, the accuracy is
unknown to other mappers, and may in fact be worse (for example an
armchair mapper digitising from small scale map). 

--colin 

On 2015-09-11 20:02, Greg Troxel wrote: 

> Colin Smale  writes:
> 
>> How does the chap with the GPS on his smartphone know that the old
>> coordinate of 50.000 is "more correct" than his own measurement?
> 
> This is generally a hard question, but if you're using a phone, and the
> value you see is within 20m (or maybe 10m) of what's in the db, then you
> shouldn't change it unless you've thought about why your new value is
> better.  And arguably data with higher-than-normal accuracy deserves a
> note.
> 
> Around me, town boundaries are defined by the actual positions of
> granite markers.  Massaschusetts publishes a dataset with coordinates
> determined from a surveying long ago, reckoned forward to a modern
> coordinate system.  Those coordinates are in OSM.  I've visited a number
> of these, and of course my handheld GPS receiver gets different numbers.
> But they are all close, and I have no reason to believe my measurements
> are better.  In fact, I think my measurements are worse.  So I'm
> certainly not changing the values in the database.
> 
> So really, people just have to understand the expected errors of their
> measurements.
 ___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Colin Smale]

 

And let's not forget that what is correct to one person, may be
inaccurate according to another. 50.N may be exactly accurate
according to a particular frame of reference at a certain point in time,
but if a surveyor with sub-centimeter accuracy equipment says it's
50.0001 then he is also right, in his own way. Similarly with road
numbers or the spelling of street names etc - as it is written on the
sign is correct in a certain frame of reference, but the spelling in the
government's administration or the spelling according to a dictionary
may be different, but also just as correct, just in a different way. 

So when claiming a piece of data is "correct" it needs to be accompanied
by a statement about "according to what standard". 

How does the chap with the GPS on his smartphone know that the old
coordinate of 50.000 is "more correct" than his own measurement? 

--colin 

On 2015-09-11 19:16, André Pirard wrote: 

> On 2015-09-09 23:39, moltonel wrote : 
> 
> On 9 September 2015 21:46:54 GMT+01:00, "André Pirard" 
>  wrote:
> 
> There are various reasons for warning other mappers to be careful about
> their updates.
> I once temporarily overlaid two walking routes to show the effect of
> displaying two sorts of icons.
> Or I left in for a while drawing errors of a plugin as the best way to
> show the author what I talk about.
> Despite a don't touch note explaining why, a good soul passes, not
> reading note and makes a "correction".
> 
> Please run experiments like this on a test db, not on the main one. It's easy 
> to point your editor to dev.openstreetmap.org for example (quoting from 
> memory, not 100% sure). You never know when a data consumer will stumble upon 
> your experiment, live or in a downloaded snapshot. Nobody expects osm data to 
> be perfect all the time, but there's no point in knowingly making it worse.
 You are off topic, as well as the following messages.
While I admit that my examples are suboptimal, the matter is extending
very simply to other tags the idea of preventing to replace precisely
triangulated coordinates by loose GPS ones.
Let us, for a better example, say that someone tagged a strange looking
name and that he knows for sure that the spelling is correct.  After the
third time the name was changed to a apparently better but wrong
spelling, he will want to enforce reading the note that nobody reads.
That's all there is to the suggestion you removed from this message.

Now responding to your accusations.
What big sin is that to discover errors and leave them a few more days
on the map for the developer of the tool that produced them to have a
look at them?  Is there a prescribed time limit?
Like you, I have always advocated a sandbox, especially for helping
novices. I've never heard of one and it's the first time I do. But JOSM
says "The server responds with the return code 404 instead of 200. "
when trying to validate http://dev.openstreetmap.org/ as well as
http://api06.dev.openstreetmap.org/
Thanks, but please give correct information.
But a sandbox wouldn't help with the first bad example because it's to
be looked at on Waymarked trails and that program does not display
sandbox data.  And as we're told that those URL if they worked wouldn't
have a renderer, they wouldn't be very convenient to use.
Please make practical suggestions !!!

On 2015-09-10 18:41, Kotya Karapetyan wrote :

> But otherwise I think there is a difference between a general warning or 
> message from one mapper to another (which in its own is an interesting idea 
> but can lead to dialogues and discussions) and a specific technical feature 
> that would prevent moving an accurately positioned tag.  
> 
> Imagine there is a real-world marker at 50.000° N: 
> http://www.dieweltenbummler.de/geografisches/geografische-besonderheiten/50-breitengrad/
>  
> Someone draws it in OSM at 50° N. Then I come there with a smartphone, 
> measure the location, find it at 49.9° and edit the OSM accordingly. It is 
> wrong by definition (providing that the real-world marker location is known 
> precisely), but there is no mechanism to prevent such editing. 
> 
> I think it's a very specific and relevant gap, and would love seeing it 
> solved elegantly.
 That is what my suggestion does and I wonder why the heck it has been
removed from this message !!!
It would produce a message saying something like:  "The coordinates you
are trying to change are accurate to 25 cm.  You probably shouldn't
change this tag, certainly not with GPS data.  Are you certain that you
will not destroy valuable data and do you want to continue?".
And if he replies "no", his attempt is canceled.
This kind of message would be possible for any tag and I don't
understand why you want it to be specific.
What elegance does it lack? You should explain. It allows to make
updates to better that 25 cm.
We know that it's typical of OSM to be crowded with stupid vandalism
("can I erase what I don't understand?"), 

Re: [Tagging] Handle with care

[Greg Troxel]

Colin Smale  writes:

> I meant my question seriously, not hypothetically I assume all these
> boundary posts are tagged with something like "source:location=MA State
> Data Set 2015-01-19"? If not, how is a mapper to compare his
> "correctness" to the existing "correctness"? 

There are changeset tags, yes.
And looking at

  https://wiki.openstreetmap.org/wiki/Massachusetts

one is directed to MassGIS imports

  https://wiki.openstreetmap.org/wiki/MassGIS

which as

  
https://wiki.openstreetmap.org/wiki/MassGIS#Town_Boundaries_from_Survey_Points_Layer

which links to the state page about this.

> Without this provenance information in the database, the accuracy is
> unknown to other mappers, and may in fact be worse (for example an
> armchair mapper digitising from small scale map). 

When it isn't clear that what you're doing is an improvement, you should
message the other people involved.  In the Mass case, most of the active
mappers know each other.  Basically my point is that absent having a
good basis to know that what you're doing is an improvement, you should
be careful.  And a phone/hiking GPSr that shows within 10-20m is not
presumption that the new data is better.   If the phone GPS is showing
200m off, then yes, the existing data is bad.

I have actually written to people when I have tracks for hiking trails
in nearby towns that are off about 20m, to ask how they did it.  And
usually they say - "just my phone, feel free to fix".

And if you're not local, and definitely if you're not actually on the
ground, you should be extra careful.


pgpyDrI8nLwNN.pgp
Description: PGP signature
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Greg Troxel]

Colin Smale  writes:

> How does the chap with the GPS on his smartphone know that the old
> coordinate of 50.000 is "more correct" than his own measurement? 

This is generally a hard question, but if you're using a phone, and the
value you see is within 20m (or maybe 10m) of what's in the db, then you
shouldn't change it unless you've thought about why your new value is
better.  And arguably data with higher-than-normal accuracy deserves a
note.

Around me, town boundaries are defined by the actual positions of
granite markers.  Massaschusetts publishes a dataset with coordinates
determined from a surveying long ago, reckoned forward to a modern
coordinate system.  Those coordinates are in OSM.  I've visited a number
of these, and of course my handheld GPS receiver gets different numbers.
But they are all close, and I have no reason to believe my measurements
are better.  In fact, I think my measurements are worse.  So I'm
certainly not changing the values in the database.

So really, people just have to understand the expected errors of their
measurements.


pgpNIqoT5MhMt.pgp
Description: PGP signature
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[Kotya Karapetyan]

Hi André,

I agree with moltonel.

But otherwise I think there is a difference between a general warning or
message from one mapper to another (which in its own is an interesting idea
but can lead to dialogues and discussions) and a specific technical feature
that would prevent moving an accurately positioned tag.

Imagine there is a real-world marker at 50.000° N:
http://www.dieweltenbummler.de/geografisches/geografische-besonderheiten/50-breitengrad/
Someone draws it in OSM at 50° N. Then I come there with a smartphone,
measure the location, find it at 49.9° and edit the OSM accordingly. It is
wrong by definition (providing that the real-world marker location is known
precisely), but there is no mechanism to prevent such editing.

I think it's a very specific and relevant gap, and would love seeing it
solved elegantly.

Kind regards,
Kotya


On Wed, Sep 9, 2015 at 11:39 PM, moltonel  wrote:

>
>
> On 9 September 2015 21:46:54 GMT+01:00, "André Pirard" <
> a.pirard.pa...@gmail.com> wrote:
> >There are various reasons for warning other mappers to be careful about
> >their updates.
> >I once temporarily overlaid two walking routes to show the effect of
> >displaying two sorts of icons.
> >Or I left in for a while drawing errors of a plugin as the best way to
> >show the author what I talk about.
> >Despite a don't touch note explaining why, a good soul passes, not
> >reading note and makes a "correction".
>
> Please run experiments like this on a test db, not on the main one. It's
> easy to point your editor to dev.openstreetmap.org for example (quoting
> from memory, not 100% sure). You never know when a data consumer will
> stumble upon your experiment, live or in a downloaded snapshot. Nobody
> expects osm data to be perfect all the time, but there's no point in
> knowingly making it worse.
> --
> Vincent Dp
>
> ___
> Tagging mailing list
> Tagging@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/tagging
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[Volker Schmidt]

I am not expert enough to know if it's tecnically feasable:
Could we not put such markers in a separate db and make that available as
TMS/WMS service?

Volker 

>
>
___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care

[Andy Townsend]

On 09/09/2015 22:39, moltonel wrote:
Please run experiments like this on a test db, not on the main one. 
It's easy to point your editor to dev.openstreetmap.org for example 
(quoting from memory, not 100% sure).


While that's a good idea (the test URL I use for such things is 
http://api06.dev.openstreetmap.org/ btw) it's perhaps worth mentioning 
that I'm not aware of anything rendering "standard" map tiles from the 
data in there.  I'm also not aware of anyone making planet files or 
extracts from there either*.


However, for a very small area (but still big enough to show the sort of 
effects you're looking for) I suspect you'd be OK doing an API MAP call 
to get the data and throwing that into a rendering database which you 
can access in the usual "rendering server" way, or by something like 
tilemill.


I had been thinking about writing some notes about how to do this for a 
while, just haven't got the necessary "round tuit" yet :)


Cheers,

Andy

* but I'd be delighted to be proved wrong on both counts of course.



___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[André Pirard]

  
  
On 2014-12-29 15:27, Kotya Karapetyan
  wrote :


  
Happy holidays and 2015 everyone!


> what is
needed here is some tag, saying "don't touch these
> coordinates,
  they've been surveyed with high(est) accuracy".


I second this idea.


Just recently I discovered that something in this direction
  already exists: https://wiki.openstreetmap.org/wiki/WikiProject_France/Rep%C3%A8res_G%C3%A9od%C3%A9siques#Permanence_des_rep.C3.A8res
Example: https://www.openstreetmap.org/edit#map=23/43.42272/6.76665
However it seems to be France-specific. I don't know if a
  similar thing exists e.g. for Germany.


Since such reference points are quite common, I would
  support the idea of creating a special tag for them, requiring
  that they are not moved. However we need a clear consensus on
  how we define the "sufficient" accuracy and how the data for
  such points will be updated.

  

These are very good ideas but restricted to a very particular case.
There are various reasons for warning other mappers to be careful
about their updates.
I once temporarily overlaid two walking routes to show the effect of
displaying two sorts of icons.
Or I left in for a while drawing errors of a plugin as the best way
to show the author what I talk about.
Despite a don't touch note explaining why, a good soul passes, not
reading note and makes a "correction".
What is needed here is an "are you sure" tag named
[keyname:]warning=* or [keyname:caution]=* that the editor uses any
time a mapper wants to change that key's value (not uploads a dozen
updates)  to display the message and ask for a confirmation.
Or should it be [keyname:]note:warn=* and spare another wiki page?
keyname can be "geometry" as in source:geometry.
Et voilà.  An all-purpose simple guardrail, a small update to the
wiki and passing the word to the editors.

What do you think?

Cheers



  

  André.

  




  


___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging

Re: [Tagging] Handle with care (was: Accuracy of survey)

[moltonel]

On 9 September 2015 21:46:54 GMT+01:00, "André Pirard" 
 wrote:
>There are various reasons for warning other mappers to be careful about
>their updates.
>I once temporarily overlaid two walking routes to show the effect of
>displaying two sorts of icons.
>Or I left in for a while drawing errors of a plugin as the best way to
>show the author what I talk about.
>Despite a don't touch note explaining why, a good soul passes, not
>reading note and makes a "correction".

Please run experiments like this on a test db, not on the main one. It's easy 
to point your editor to dev.openstreetmap.org for example (quoting from memory, 
not 100% sure). You never know when a data consumer will stumble upon your 
experiment, live or in a downloaded snapshot. Nobody expects osm data to be 
perfect all the time, but there's no point in knowingly making it worse.
-- 
Vincent Dp

___
Tagging mailing list
Tagging@openstreetmap.org
https://lists.openstreetmap.org/listinfo/tagging