iD forces mistagging again

classic Classic list List threaded Threaded
19 messages Options
Reply | Threaded
Open this post in threaded view
|

iD forces mistagging again

Tomas Straupis
Hi

  I've noticed that iD started forcing people to retag waterbodies
from original OpenStreetMap scheme like landuse=reservoir to new'er,
less popular and in no way better scheme:
natural=water+water=reservoir (and similar).

  What can be done about it? Is the only way to solve the problem is
by going the way of France and deploy own iD (or different) editor
with such misbehaviour removed?

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

FW: iD forces mistagging again

Phil Wyatt


-----Original Message-----
From: Phil Wyatt [mailto:[hidden email]]
Sent: Saturday, 29 June 2019 5:00 PM
To: 'Tomas Straupis'
Subject: RE: [OSM-talk] iD forces mistagging again

Hi Tomas,

I think you are underestimating the intelligence of OSM users - I see lots of suggestions but only change the ones I have personal knowledge of. Again, you still need to push the upgrade the tags button so they are not forcing you!

Personally, I am happy they are offering suggestions as I am learning more about the (somewhat confusing) tagging system.

Cheers - Phil

-----Original Message-----
From: Tomas Straupis [mailto:[hidden email]]
Sent: Saturday, 29 June 2019 4:48 PM
To: Phil Wyatt
Subject: Re: [OSM-talk] iD forces mistagging again

2019-06-29, št, 09:40 Phil Wyatt rašė:
> They are not forcing you - they are giving you the option to update if you so desire, and if you press the button.

  I think this was already discussed under different "suggestions" of
iD. When editor which looks like "official" makes a suggestion, most
users would think this is the way to go. Very few would go to wiki to
find out that iD suggestion is in no way better.

  Or in other words iD should not suggest change which would give no
benefit to OSM, just benefit to iD developers as this would make tags
easier to code in iD.

--
Tomas


_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Phil Wyatt
In reply to this post by Tomas Straupis
Yep - a messy tagging scheme will always be open to individual, or editor, interpretation.

So any excess energy we have should go into organising a better tagging scheme that can be used by all editors for validation purposes. That would go a long way to solving many of the tagging wars.

Cheers - Phil

-----Original Message-----
From: Tomas Straupis [mailto:[hidden email]]
Sent: Saturday, 29 June 2019 5:30 PM
To: Phil Wyatt
Subject: Re: [OSM-talk] iD forces mistagging again

> ..... because of duplicate schemes...... And therin lies the real problem!

  I've said years ago when this duplicate scheme was introduced that
it will lead to problems for everybody. And here we go.
  If now new scheme is accepted this will make a bad precedent.


_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Mateusz Konieczny-3
In reply to this post by Tomas Straupis
29 Jun 2019, 08:11 by [hidden email]:
Hi

I've noticed that iD started forcing people to retag waterbodies
from original OpenStreetMap scheme like landuse=reservoir to new'er,
less popular and in no way better scheme:
natural=water+water=reservoir (and similar).
(1) Have you (or someone else) tried making issue on iD bugtracker requesting revert
and explaining why it should be done? Even assuming that it is unwanted, there is no
good reason to start from tagging list mail

(2) I see significant benefit of natural=water + water=*
This way one may display water with simply looking for natural=water
rather than multiple separate values. And I would not describe using
landuse=reservoir for reservoir water area as a desirable tagging.
What can be done about it?
see (1)


_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Tomas Straupis
2019-06-29, št, 11:58 Mateusz Konieczny rašė:
> (1) Have you (or someone else) tried making issue on iD bugtracker requesting revert
> and explaining why it should be done?

  https://github.com/openstreetmap/iD/issues/6589

> (2) I see significant benefit of natural=water + water=*

  This is your personal opinion. Opinion of OpenStreetMap community is
expressed by those who map - in the data.

> This way one may display water with simply looking for natural=water
> rather than multiple separate values. And I would not describe using
> landuse=reservoir for reservoir water area as a desirable tagging.

  There are millions of arguments for original OpenStreetMap scheme as
well as for the new one. There is simply no point on drowning this
discussion in such pointless discussion.
  As water tagging is very prominent, my opinion is that most
important thing is STABILITY (it is based on my personal experience in
creating maps, analyses, QA as well as talking to other
people/organisations using OpenStreetMap data). There should be some
very good reasons to change such a prominent tag because it is not
only a change in one editor, it is a change in millions of places:
vector tiles, cartography rules, data analysis, QA, documentation,
training.

  Current situation is that original OpenStreetMap scheme (with
landuse=reservoir) is used more even where JOSM allows both schemes
and iD only allows the new one from the very beginning of existence of
iD (and now pushing people to update tags to scheme preferred by iD
developers. I already had to revert a number of such changes and
explain people to ignore such "advices" or switch to a better editor).

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Christoph Hormann-2
On Saturday 29 June 2019, Tomas Straupis wrote:
> > (2) I see significant benefit of natural=water + water=*
>
>   This is your personal opinion.

No, that is a statement of fact.  If this is a good reason for choosing
a certain tagging over another is a matter of opinion.

>   Opinion of OpenStreetMap community
> is expressed by those who map - in the data.

And it is clear from the data that both tagging schemes enjoy widespread
support.

The situation is actually much more balanced for reservoirs than for
waterway=riverbank vs. water=river where pushing for the latter is much
more questionable.  In other words:  You are kind of "barking up the
wrong tree" here.

There are many reasons why preset decisions of iD deserve a critical
look.  Choosing one of two widely accepted and widely used tagging
schemes not exclusively based on absolute use numbers or structural
conservativism (which was there first) is not one of them.

--
Christoph Hormann
http://www.imagico.de/

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Tomas Straupis
2019-06-29, št, 13:38 Christoph Hormann rašė:
>>   This is your personal opinion.
> No, that is a statement of fact.  If this is a good reason for choosing
> a certain tagging over another is a matter of opinion.

  When somebody simply says "I think this is better" - it is a
subjective opinion.
  For it to be a fact (or more correctly expressed "objectively true")
you need some objective criteria.
  "It is cool that everything blue is natural=water" is not an
objective criteria, because say lake and basin/pond are VERY different
and have different usage for both data analysis as well as
topology/cartography.

>>   Opinion of OpenStreetMap community
>> is expressed by those who map - in the data.
> And it is clear from the data that both tagging schemes enjoy widespread
> support.

  ? landuse=reservoir is used two times more. And if iD would not be
forcing people to use new scheme (by not giving a choice) and now
retaging to the new scheme, this ratio could be different.

> The situation is actually much more balanced for reservoirs than for
> waterway=riverbank vs. water=river where pushing for the latter is much
> more questionable.  In other words:  You are kind of "barking up the
> wrong tree" here.

  I'm against changing ANY of original OpenStreetMap water tagging. I
was given no good reason or benefit why I or any other data consumer
should rewrite everything what is already done just because someone
says "I think it is better". For somebody it is simply writing some
comments, but for others it is writing harder stuff, testing,
educating etc.
  Ad extremum example: anyone for changing "highway" key to something
more appropriate? <<< NOT TO BE TAKEN LITERALLY :-)

> There are many reasons why preset decisions of iD deserve a critical
> look.  Choosing one of two widely accepted and widely used tagging
> schemes not exclusively based on absolute use numbers or structural
> conservativism (which was there first) is not one of them.

  I'm currently only asking to stop iD from breaking existing stuff.

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Mateusz Konieczny-3
In reply to this post by Tomas Straupis



29 Jun 2019, 11:59 by [hidden email]:
2019-06-29, št, 11:58 Mateusz Konieczny rašė:
(2) I see significant benefit of natural=water + water=*

This is your personal opinion. Opinion of OpenStreetMap community is
expressed by those who map - in the data.

All "I prefer tag Y over X" are "rule Z is good way to deciding which tag is better" is
a personal opinion, so I am not sure why you are pointing this out.

I even started from "I see (...)".

29 Jun 2019, 11:59 by [hidden email]:
Opinion of OpenStreetMap community is expressed by those who map - in the data.

(1) most of lead of landuse=reservoir for areas is a result of a bot edit, maybe an import
(2) since 2016 water=reservoir is growing much faster than landuse=reservoir

This quick check is a bit better than just checking total usage count, but obviously
it is not worth much. It may be useful to investigate actual usage among mappers
who are actually selecting tagging scheme used.
As water tagging is very prominent
This concern tagging details of body of water what is actually not prominent.
I already had to revert a number of such changes and
explain people to ignore such "advices" or switch to a better editor).
"had to revert" is untrue.

Are you manually checking every single part of such edits? Otherwise you
are making automated edits in violation of automatic edits code of conduct.

It is also not too useful, most what you achieve is to confuse other mappers and
likely scare away some of them.

Please stop doing that.

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Mateusz Konieczny-3
In reply to this post by Tomas Straupis
29 Jun 2019, 13:02 by [hidden email]:
2019-06-29, št, 13:38 Christoph Hormann rašė:
And it is clear from the data that both tagging schemes enjoy widespread
support.

? landuse=reservoir is used two times more.
"two times more" is not a significant difference, especially in case of total usage count.

You checked only total usage count, without even checking that large part of landuse=reservoir
is on nodes or history of use.

And you use it as a sufficient to loudly complain about developers and harass other mappers.

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Blake Girardot
In reply to this post by Mateusz Konieczny-3
If I understand things correctly, the original poster is reverting
perfectly fine changes to an equivalent, accepted, current tagging
scheme, because they do not want to update their own local code that
uses the data.

I can only imagine if HOT or another group started doing that.

"Our scripts were written to use the original tagging scheme from 10
years ago, and this new tagging scheme means I would have to change my
scripts, so I am just going to change all the tags back to what they
were."

lol. Someone would write another directed editing policy just to make
sure that didn't happen.

Cheers,
Blake



On Sat, Jun 29, 2019 at 7:23 AM Mateusz Konieczny
<[hidden email]> wrote:

>
>
>
>
> 29 Jun 2019, 11:59 by [hidden email]:
>
> 2019-06-29, št, 11:58 Mateusz Konieczny rašė:
>
> (2) I see significant benefit of natural=water + water=*
>
>
> This is your personal opinion. Opinion of OpenStreetMap community is
> expressed by those who map - in the data.
>
>
> All "I prefer tag Y over X" are "rule Z is good way to deciding which tag is better" is
> a personal opinion, so I am not sure why you are pointing this out.
>
> I even started from "I see (...)".
>
> 29 Jun 2019, 11:59 by [hidden email]:
>
> Opinion of OpenStreetMap community is expressed by those who map - in the data.
>
> See https://taghistory.raifer.tech/
> https://user-images.githubusercontent.com/899988/60383077-c001ce00-9a6c-11e9-9aa8-ed43c7851a36.png for a quick check.
>
> (1) most of lead of landuse=reservoir for areas is a result of a bot edit, maybe an import
> (2) since 2016 water=reservoir is growing much faster than landuse=reservoir
>
> This quick check is a bit better than just checking total usage count, but obviously
> it is not worth much. It may be useful to investigate actual usage among mappers
> who are actually selecting tagging scheme used.
>
> As water tagging is very prominent
>
> This concern tagging details of body of water what is actually not prominent.
>
> I already had to revert a number of such changes and
> explain people to ignore such "advices" or switch to a better editor).
>
> "had to revert" is untrue.
>
> Are you manually checking every single part of such edits? Otherwise you
> are making automated edits in violation of automatic edits code of conduct.
>
> It is also not too useful, most what you achieve is to confuse other mappers and
> likely scare away some of them.
>
> Please stop doing that.
> _______________________________________________
> talk mailing list
> [hidden email]
> https://lists.openstreetmap.org/listinfo/talk



--
----------------------------------------------------
Blake Girardot
OSM Wiki - https://wiki.openstreetmap.org/wiki/User:Bgirardot
HOTOSM Member - https://hotosm.org/users/blake_girardot
skype: jblakegirardot

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Tomas Straupis
In reply to this post by Mateusz Konieczny-3
2019-06-29, št, 14:23 Mateusz Konieczny rašė:
> All "I prefer tag Y over X" are "rule Z is good way to deciding which tag is better" is
> a personal opinion, so I am not sure why you are pointing this out.
> I even started from "I see (...)".

  I pointed that:
  1. Original OpenStreetMap water tagging scheme is used more often - Fact.
  2. Original OpenStreetMap water tagging scheme is used much longer
(from more or less the beginning of the project) - Fact.

  Here I would note that 2nd point is enough to keep original water
scheme and depreciate the new one. Because of data consumers.

> https://user-images.githubusercontent.com/899988/60383077-c001ce00-9a6c-11e9-9aa8-ed43c7851a36.png for a quick check.

  Yes, it says landuse=reservoir is used twice as often.

> (1) most of lead of landuse=reservoir for areas is a result of a bot edit, maybe an import

  How do you know that?

> (2) since 2016 water=reservoir is growing much faster than landuse=reservoir

  Remove "iD effect" and what will we get?

> I already had to revert a number of such changes and
> explain people to ignore such "advices" or switch to a better editor).
> "had to revert" is untrue.

  It is true. There is an old agreement in Lithuania on how we tag
water (predating new water scheme). All code is written according to
that agreement. Fastest way to not brake existing stuff is revert iD
"advice" and give mappers yet one more reason not to use iD.

> Are you manually checking every single part of such edits? Otherwise you
> are making automated edits in violation of automatic edits code of conduct.

 I'm doing QA in Lithuania for more than ten years and then doing
appropriate action: fixing, informing, educating, updating.

> Please stop doing that.

  You cannot simply tell me what to do without giving reasons with at
least as much practical examples as I have.

  And once again, please step out of "commenter" role and try getting
into data consumer role. Commenting takes minutes and you can change
your opinion every five minutes, consuming takes months and years of
hard work and it is much harder to adopt to changes, it is especially
demotivating to be forced to adopt to the changes which give no
benefit.

  Don't change/brake stuff that works.

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Mike N.
On 6/29/2019 8:08 AM, Tomas Straupis wrote:
>    Here I would note that 2nd point is enough to keep original water
> scheme and depreciate the new one. Because of data consumers.

  I don't remember why but I arrived at the new scheme several years ago
and have been using it ever since.   So apparently data consumers will
be ignoring my tagging?

   (Using JOSM by the way).

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: iD forces mistagging again

Tomas Straupis
2019-06-29, št, 15:38 Mike N rašė:
>   I don't remember why but I arrived at the new scheme several years ago
> and have been using it ever since.   So apparently data consumers will
> be ignoring my tagging?

  There are so many of them (data consumers) that it is possible to
say that some will be ignoring, some will be getting incorrect
information (say interpreting ponds as lakes) etc. But some would be
prepared for both schemas.
  CHANGING schema which works is a problem.

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Mateusz Konieczny-3
In reply to this post by Tomas Straupis
29 Jun 2019, 14:08 by [hidden email]:
(1) most of lead of landuse=reservoir for areas is a result of a bot edit, maybe an import

How do you know that?
See spike in the orange line on the left.
(2) since 2016 water=reservoir is growing much faster than landuse=reservoir

Remove "iD effect" and what will we get?
Not sure, I mentioned that proper analysis would be valuable.
Are you manually checking every single part of such edits? Otherwise you
are making automated edits in violation of automatic edits code of conduct.

I'm doing QA in Lithuania for more than ten years and then doing
appropriate action: fixing, informing, educating, updating.
I am not sure what is your method used for mass reverts that you mentioned but
automatic edits code of conduct still applies to you.

Don't change/brake stuff that works.
Software supporting only landuse=reservoir or only water=reservoir for
detecting reservoirs is already broken and was broken for a long time.

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Tomas Straupis
2019-06-29, št, 15:53 Mateusz Konieczny rašė:
>> Don't change/brake stuff that works.
> Software supporting only landuse=reservoir or only water=reservoir for
> detecting reservoirs is already broken and was broken for a long time.

  How many maps, analysis, QA routines, presentations, documentations
have you created to come up with such a claim?

  This discussion should have been done looong time ago when new water
scheme was introduced, then damage was not this large :-( We're
pointlessly arguing about names of tags when other GIS use simple
codes like gk11 gk12 gk13 for classification with an explanatory table
attached and have no problem with that, because values are more
STABLE.

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Yves-2
In reply to this post by Mateusz Konieczny-3
Don't worry about data consumers, they have a long habit of using concurrent tagging schemes and make extensive use of 'OR'.
Yves

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Mateusz Konieczny-3
In reply to this post by Tomas Straupis
29 Jun 2019, 15:42 by [hidden email]:
2019-06-29, št, 15:53 Mateusz Konieczny rašė:
Don't change/brake stuff that works.
Software supporting only landuse=reservoir or only water=reservoir for
detecting reservoirs is already broken and was broken for a long time.

How many maps, analysis, QA routines, presentations, documentations
have you created to come up with such a claim?
To make such claim it is enough to look at tag statistics, history of usage
and note that water=reservoir is generally not tagged with landuse=reservoir

I am not sure why you are even asking, you are aware about this:
This discussion should have been done looong time ago when new water
scheme was introduced, then damage was not this large
(though I would not describe it as damage)

29 Jun 2019, 15:42 by [hidden email]:
How many maps, analysis, QA routines, presentations, documentations
have you created to come up with such a claim?
Feel free to check JOSM bugtracker, Github and OSM Wiki to see sample of
public contributions. I use usernames "matkoniecz" and "Mateusz Konieczny".

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Tomas Straupis
2019-06-29, št, 18:30 Mateusz Konieczny rašė:
>> How many maps, analysis, QA routines, presentations, documentations
>> have you created to come up with such a claim?
>
> To make such claim it is enough to look at tag statistics, history of usage
> and note that water=reservoir is generally not tagged with landuse=reservoir

  If you would have created the things I've listed above (which
consume OSM data), you would understand why I'm so opposed to random
shuffling of tags without a good reason.

  Tagging in OSM is something like an "API" to data and usually
serious organisations/companies are very strict about changes in API,
and breaking changes are only done if really really really necessary.

--
Tomas

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk
Reply | Threaded
Open this post in threaded view
|

Re: landuse=reservoir vs water=reservoir

Andy Townsend
Usually I'm the one wringing my hands and plaintively saying "won't someone think of the data consumers?".  In this case though anyone who's not looking for "natural=water; water=blah" (as well as any other options) really is doing it wrong - people have been using that scheme for years.

Best Regards,

Andy

_______________________________________________
talk mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/talk