Fwd: Re: Roadmap for deprecation of name tags in OSM

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|

Fwd: Re: Roadmap for deprecation of name tags in OSM

pangose



-------- Originalmeddelande --------
Från: pangoSE
Skickat: 9 augusti 2020 15:40:41 CEST
Till: [hidden email]
Ämne: Re: [OSM-talk] Roadmap for deprecation of name tags in OSM

This is another good reason to abandon this suggestion in favor of our own wikibase instance.

Philip Barnes <[hidden email]> skrev: (9 augusti 2020 15:12:21 CEST)

>On Sun, 2020-08-09 at 09:04 -0400, James wrote:
>> Not to mention if someone wants to add a name for a new item/object,
>> does the user need to create a wikidata item on top of it? Will the
>> editor do it automatically? How does it pick the right one? Does it
>> offer a list to the user? This is going to make osm a massive turn
>> off to new contributors on the steep learning curve(which is already
>> pretty high) to contributing to osm.
>> This whole idea is really terrible and could just be offered as a
>> post-processed data set: wikidata? use that instead of name tag.
>>
>
>This leads me to a fairly fundamental question, what if a mapper does
>not want to be associated with wikidata and refuses to sign up?
>Phil (trigpoint)

--
Skickat från min Android-enhet med k9.

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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

General Discussion mailing list
It still fails to provide even a single benefit over the current situation.

Aug 9, 2020, 20:11 by [hidden email]:



-------- Originalmeddelande --------
Från: pangoSE
Skickat: 9 augusti 2020 15:40:41 CEST
Ämne: Re: [OSM-talk] Roadmap for deprecation of name tags in OSM

This is another good reason to abandon this suggestion in favor of our own wikibase instance.

Philip Barnes <[hidden email]> skrev: (9 augusti 2020 15:12:21 CEST)
>On Sun, 2020-08-09 at 09:04 -0400, James wrote:
Not to mention if someone wants to add a name for a new item/object,
does the user need to create a wikidata item on top of it? Will the
editor do it automatically? How does it pick the right one? Does it
offer a list to the user? This is going to make osm a massive turn
off to new contributors on the steep learning curve(which is already
pretty high) to contributing to osm.
This whole idea is really terrible and could just be offered as a
post-processed data set: wikidata? use that instead of name tag.
>This leads me to a fairly fundamental question, what if a mapper does
>not want to be associated with wikidata and refuses to sign up?
>Phil (trigpoint)

--
Skickat från min Android-enhet med k9.

_______________________________________________
talk mailing list
https://lists.openstreetmap.org/listinfo/talk


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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

pangose
Could you reply with your arguments in favor of the current one2one tag model system in the other thread where I listed the benefits as I see them?

E.g. permanent unique ids, talk pages if we want that for every osmid, SPARQL support, standardization benefits "riding the current ride in open data", scripting support for botmakers, and above all support for references and linking interactively to other data sources.
Bots are very useful e.g. to notify an editor of a possible tagging mistake or checking urls of references. Martin could also reference an image in Wikimedia commons directly on the statement it relates to.

Unique Qids for every osm object that is decoupled from the underlying osmid gives some possibility we don't have today regarding integration with e.g. wikidata.

Cheers

Mateusz Konieczny via talk <[hidden email]> skrev: (9 augusti 2020 20:14:03 CEST)
It still fails to provide even a single benefit over the current situation.

Aug 9, 2020, 20:11 by [hidden email]:



-------- Originalmeddelande --------
Från: pangoSE
Skickat: 9 augusti 2020 15:40:41 CEST
Ämne: Re: [OSM-talk] Roadmap for deprecation of name tags in OSM

This is another good reason to abandon this suggestion in favor of our own wikibase instance.

Philip Barnes <[hidden email]> skrev: (9 augusti 2020 15:12:21 CEST)
>On Sun, 2020-08-09 at 09:04 -0400, James wrote:
Not to mention if someone wants to add a name for a new item/object,
does the user need to create a wikidata item on top of it? Will the
editor do it automatically? How does it pick the right one? Does it
offer a list to the user? This is going to make osm a massive turn
off to new contributors on the steep learning curve(which is already
pretty high) to contributing to osm.
This whole idea is really terrible and could just be offered as a
post-processed data set: wikidata? use that instead of name tag.
>This leads me to a fairly fundamental question, what if a mapper does
>not want to be associated with wikidata and refuses to sign up?
>Phil (trigpoint)

--
Skickat från min Android-enhet med k9.

_______________________________________________
talk mailing list
https://lists.openstreetmap.org/listinfo/talk


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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

john whelan-2
I honestly can't see any benefit.  Splitting the data into two places adds the danger of it getting out of sync.

Standard naming conventions would be nice but defining the standard name is practically impossible.  Compare taginfo to the map features wiki page.  One problem with map features is what is written is often one person's idea of what the standard name should be.

And different features really are called difference things in different countries.  It can be difficult to stretch a "standard" name to cover many things.

For example in Canada many highways have wide shoulders to dump snow on in winter.  In summer they are often used by cyclists but they aren't cycle lanes by any stretch of imagination even though some are paved.

Cheerio John

On Sun, Aug 9, 2020, 15:04 pangoSE <[hidden email]> wrote:
Could you reply with your arguments in favor of the current one2one tag model system in the other thread where I listed the benefits as I see them?

E.g. permanent unique ids, talk pages if we want that for every osmid, SPARQL support, standardization benefits "riding the current ride in open data", scripting support for botmakers, and above all support for references and linking interactively to other data sources.
Bots are very useful e.g. to notify an editor of a possible tagging mistake or checking urls of references. Martin could also reference an image in Wikimedia commons directly on the statement it relates to.

Unique Qids for every osm object that is decoupled from the underlying osmid gives some possibility we don't have today regarding integration with e.g. wikidata.

Cheers

Mateusz Konieczny via talk <[hidden email]> skrev: (9 augusti 2020 20:14:03 CEST)
It still fails to provide even a single benefit over the current situation.

Aug 9, 2020, 20:11 by [hidden email]:



-------- Originalmeddelande --------
Från: pangoSE
Skickat: 9 augusti 2020 15:40:41 CEST
Ämne: Re: [OSM-talk] Roadmap for deprecation of name tags in OSM

This is another good reason to abandon this suggestion in favor of our own wikibase instance.

Philip Barnes <[hidden email]> skrev: (9 augusti 2020 15:12:21 CEST)
>On Sun, 2020-08-09 at 09:04 -0400, James wrote:
Not to mention if someone wants to add a name for a new item/object,
does the user need to create a wikidata item on top of it? Will the
editor do it automatically? How does it pick the right one? Does it
offer a list to the user? This is going to make osm a massive turn
off to new contributors on the steep learning curve(which is already
pretty high) to contributing to osm.
This whole idea is really terrible and could just be offered as a
post-processed data set: wikidata? use that instead of name tag.
>This leads me to a fairly fundamental question, what if a mapper does
>not want to be associated with wikidata and refuses to sign up?
>Phil (trigpoint)

--
Skickat från min Android-enhet med k9.

_______________________________________________
talk mailing list

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

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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

General Discussion mailing list
In reply to this post by pangose
"riding the current ride in open data" - I am confused what is the meaning of that

"scripting support for botmakers" - as a bot operator and a bot author I am confused what is
supposed to be missing

"support for references and linking interactively to other data sources" - we have that,
see wikidata tag for example

"talk pages if we want that for every osmid" - changeset comments and notes
seem sufficient

"SPARQL support" - we have already some wikidata/OSM query engine and overpass turbo
And frankly Overpass is much nicer that SPARQL (this may be just me).

"Bots are very useful" - maybe, but it is 100% possible to have bots already so I am
confused why it is mentioned

"possibility we don't have today regarding integration with e.g. wikidata." - given
what kind of ideas appear thanks to current wikidata integration (like deprecating name tag)
I prefer reducing wikidata integration rather than deepening it (or, the best solution - keep it as is)

Aug 9, 2020, 21:01 by [hidden email]:
Could you reply with your arguments in favor of the current one2one tag model system in the other thread where I listed the benefits as I see them?

E.g. permanent unique ids, talk pages if we want that for every osmid, SPARQL support, standardization benefits "riding the current ride in open data", scripting support for botmakers, and above all support for references and linking interactively to other data sources.
Bots are very useful e.g. to notify an editor of a possible tagging mistake or checking urls of references. Martin could also reference an image in Wikimedia commons directly on the statement it relates to.

Unique Qids for every osm object that is decoupled from the underlying osmid gives some possibility we don't have today regarding integration with e.g. wikidata.

Cheers

Mateusz Konieczny via talk <[hidden email]> skrev: (9 augusti 2020 20:14:03 CEST)
It still fails to provide even a single benefit over the current situation.

Aug 9, 2020, 20:11 by [hidden email]:



-------- Originalmeddelande --------
Från: pangoSE
Skickat: 9 augusti 2020 15:40:41 CEST
Ämne: Re: [OSM-talk] Roadmap for deprecation of name tags in OSM

This is another good reason to abandon this suggestion in favor of our own wikibase instance.

Philip Barnes <[hidden email]> skrev: (9 augusti 2020 15:12:21 CEST)
>On Sun, 2020-08-09 at 09:04 -0400, James wrote:
Not to mention if someone wants to add a name for a new item/object,
does the user need to create a wikidata item on top of it? Will the
editor do it automatically? How does it pick the right one? Does it
offer a list to the user? This is going to make osm a massive turn
off to new contributors on the steep learning curve(which is already
pretty high) to contributing to osm.
This whole idea is really terrible and could just be offered as a
post-processed data set: wikidata? use that instead of name tag.
>This leads me to a fairly fundamental question, what if a mapper does
>not want to be associated with wikidata and refuses to sign up?
>Phil (trigpoint)

--
Skickat från min Android-enhet med k9.

_______________________________________________
talk mailing list
https://lists.openstreetmap.org/listinfo/talk



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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

dieterdreist
In reply to this post by pangose


sent from a phone

> On 9. Aug 2020, at 21:04, pangoSE <[hidden email]> wrote:
>
> E.g. permanent unique ids, talk pages if we want that for every osmid, SPARQL support, standardization benefits "riding the current ride in open data"


somehow you can have this already through the integration of wikidata: just add a wikidata reference to an object and you could decide to ignore all OpenStreetMap tags and rely only on wikidata (although it would not be advisable IMHO, looking at where wikidata is currently, you would loose lots of relevant information).

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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

dieterdreist
In reply to this post by john whelan-2


sent from a phone

> On 9. Aug 2020, at 21:16, john whelan <[hidden email]> wrote:
>
> And different features really are called difference things in different countries.


+1, moreover, the „same“ features are different in different countries and cultures, and it is part of our work to define when we consider them still the same and where they start to be so different that another tag is required.

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

Re: Fwd: Re: Roadmap for deprecation of name tags in OSM

Clifford Snow
In reply to this post by pangose
As I read your proposal, it sounds like you have a solution but haven't defined the problem. If you could focus on the problem and describe exactly what is wrong with the current arrangement, and what will happen if we do nothing, that might help. Otherwise I can not see the merits of your proposal. From what I've read for the feedback you've received, you haven't convinced anyone that there is a problem.

Best,
Clifford


--
@osm_washington
OpenStreetMap: Maps with a human touch

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