Re: [OSM-dev] Standardized feature/tag database idea & proposal

[Mateusz Konieczny via dev]

Nov 16, 2020, 10:21 by si...@poole.ch:

>
> the iD presets  are the law of the land and anything else is simply 
> irrelevant.
>
>
That is so overstating things that it ceases to be an useful statement.

iD presets and 
"press button here to `upgrade tags`  without explanation what and why and how 
actually happens"
are powerful, but it is only a part of what influences tagging.
___
dev mailing list
dev@openstreetmap.org
https://lists.openstreetmap.org/listinfo/dev

Re: [OSM-dev] Standardized feature/tag database idea & proposal

[Christoph Hormann]

> Simon Poole  hat am 16.11.2020 10:21 geschrieben:
> 
> [...] On the other hand, not just because of the demographics, but also 
> because of the OSMFs involvement, like it or not, the iD presets are the law 
> of the land and anything else is simply irrelevant.

From my experience with OSM-Carto that is not really accurate.  Whenever iD 
presets suggest something different from established practice or community 
consensus the result typically is not the meaning of the tag changing to what 
iD says.  The result tends to be a huge mess of inconsistent uses.

In other words: iD's power over the development tagging and tag meaning in OSM 
is primarily either destructive or affirmative.  iD has the power to wreak 
havoc and destroy a tag or it can affirm and support a trend in already 
predominantly consistent use.  Same is the case with OSM-Carto.

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

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

Re: [OSM-dev] Standardized feature/tag database idea & proposal

[Simon Poole]

Frederik has already pointed out some of the relevant historic points.

On top of that there is https://github.com/osmlab/editor-presets and I 
proposed  something similar a couple of years back as a GSOC project.


I would consider the concept moot. On the one hand, while theoretically 
still possible, the iD preset scheme now has little generalisation and 
has become very UI specific, requiring to capture a lot more editor 
specific data than just a couple of years back if you wanted the output 
to be similar to the original. On the other hand, not just because of 
the demographics, but also because of the OSMFs involvement, like it or 
not, the iD presets are the law of the land and anything else is simply 
irrelevant. That is not a statement about quality (I would maintain that 
my presets continue to be the best curated set of presets), but about mass.


You should have also given my talk at SOTM 2018 a look 
https://2018.stateofthemap.org/2018/T061-An_excursion_in_to_the_world_of_OSM_tagging_presets/


In any case both iD and Vespucci use taginfo and associated information 
to synthesize parts of, or complete presets. Vespucci as part of an 
explicit search process, iD built in to how their presets work. For a 
number of reasons the results are at best mixed, but clearly could be 
improved with some work on taginfo. I would consider this a better way 
forward than starting something new.


Simon

Am 15.11.2020 um 19:14 schrieb Seth Deegan:

_Idea_
Has anyone ever thought of creating an official database that stores 
all of the approved and in-use tags/features in OSM?


_Reasoning_
This could allow the editors (iD and JOSM, StreetComplete, GoMap!) and 
data consumers (osm-carto,. Mapbox etc.), to easily stay up-to-date 
with new features, without requiring their developers to browse the 
Wiki etc.


_Examples_
Both iD and JOSM have their own preset file/repo and are independent 
of each other. Creating a database that could be used as a dependency 
of both that would store these feature presets and their fields means:

 - Both are up-to-date and in-sync
 - Adding new presets could be done automatically by retrieving and 
parsing data from the DB.


Creating applications that use OSM data can be hard and 
time-lengthened by requiring developers to browse the Wiki to find all 
of the features and their keys and values. Having a database that they 
could easily get the keys they want, their values, etc. would 
*significantly* allow greater OSM developer potential.


_Specifics_
The specifics as to how this database would be arranged such as to 
where presets/fields/tags/features go has not been thought of yet. I 
just wanted to ask if this has ever been proposed before. If someone 
would like me to make a DB layout to help them better understand what 
I'm proposing, I'd be happy to do so.


_My pre-DB construction proposal _
Before any type of database is made, one would would *construct a 
dedicated page structure on openstreetmap.org 
 *that would display and be in-sync with all 
of the Features from the DB in a format similar to the Wiki, and then 
*remove all of the features from the Wiki altogether.*

*
*
*Why?* If you see in _Compiling and distributing the DB_ below, a Wiki 
bot is going to be required to get all of the pages for all of the 
already-standard features. Most of the features' pages do have a 
similar structure as to how they are arranged (template that shows 
what elements they use, Keys' values and their descriptions are stored 
in a table, etc.), however these pages would be /impossible/ to parse 
thoroughly with a computer and are going to require team of humans to 
get all of their data.


New features that are proposed and approved after the database would 
be created would require them to be hand-added. Making a standard way 
to propose and approve new features on a new part of the website with 
formatting constraints and database-syncing abilities that MediaWiki 
cannot offer, would mean that the DB would never have to be physically 
touched again and ensure greater long-term DB management efficiency.

*
*
*So why basically delete one of the primary functions of the Wiki and 
create a new system on openstreetmap.org ?*
I recently have got into the OSM community head-on in the past two 
months. I have never really contributed to Wikipedia or any other site 
that uses the MediaWiki website format so learning about how to 
contribute and get around the OSM Wiki took time (learning about the 
proposal process, Templates, talk pages, formatting, the list goes on 
and on...). It also made me realize that this could discourage new 
users from ever looking into the depths of OSM or even finding the 
Wiki at all. The WikiMedia interface is not the prettiest either. It 
can take time for new users to explore and find what they are looking 
for.
(Also, this would mean moving the proposal process over to osm.org 
 too)

Re: [OSM-dev] Standardized feature/tag database idea & proposal

[Frederik Ramm]

Seth,

On 11/15/20 19:14, Seth Deegan wrote:
> Has anyone ever thought of creating an official database that stores all
> of the approved and in-use tags/features in OSM? 

Yes, of course, this idea has been around in a variety of flavours. I
remember a conference talk by David Earl in 2010
(https://wiki.openstreetmap.org/wiki/SotM_2010_session:_Tag_Central:_a_Schema_for_OSM)
though even then it was not revolutionary.

I think that going from a wiki to a plain database doesn't actually
solve many problems. Many things that are difficult for people to deal
with - like tag definitions leaving room for interpretation, different
editors having different presets, different language versions describing
tags differently, a lot of vagueness about the proposal process and its
importance, etc.etc. - would not be solved by your idea. You'd only
replace the backend but that would not automatically fix the messy bits
(and for some of the messy bits, "fixing" them might also mean breaking
a part of OSM that many cherish).

For example, one point you mention is that with a database, things could
be somehow "locked", but you're completely ignoring the question of who
would get to lock stuff, who would decide which changes are allowed, and
how this relates to translations, what the appeals process would be
etc.etc. - and *those* are the difficult questions, not the question of
whether I could technically have a mechanism that would restrict editing
to certain people.

And of course, as has been pointed out, the wiki stuff can already be
accessed in a database-like fashion, not only through the data endpoint
but also through the taginfo-wiki SQLite database from
https://taginfo.openstreetmap.org/download.

Bye
Frederik

-- 
Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

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

Re: [OSM-dev] Standardized feature/tag database idea & proposal

[Mateusz Konieczny via dev]

Nov 15, 2020, 19:14 by jayands...@gmail.com:

> Idea
> Has anyone ever thought of creating an official database that stores all of 
> the approved and in-use tags/features in OSM? 
>
Yes. 

Depending on what one means by "database" OSM Wiki can be considered as one.

Wikidata copy ( https://wiki.openstreetmap.org/wiki/Data_items ) is definitely 
one.
How your proposal differs from data items?

>
> Reasoning
> This could allow the editors (iD and JOSM, StreetComplete, GoMap!) and data 
> consumers (osm-carto,. Mapbox etc.), to easily stay up-to-date with new 
> features, without requiring their developers to browse the Wiki etc.
>
I see no benefit whatsoever of data items (from perspective of someone involved 
in making 
StreetComplete and was active in osm-carto development).

Browsing Wiki, Taginfo, reviewing how tags are actually used and so on would be 
definitely needed.

>
> Examples
> Both iD and JOSM have their own preset file/repo and are independent of each 
> other. Creating a database that could be used as a dependency of both that 
> would store these feature presets and their fields means:
>  - Both are up-to-date and in-sync
>  - Adding new presets could be done automatically by retrieving and parsing 
> data from the DB. 
>
There is a good reason why presets are created manually and not pulled from 
unreviewed
dataset (note that JOSM and iD have separate presets despite that pulling from 
another preset
would be technically possinle)

> Creating applications that use OSM data can be hard and time-lengthened by 
> requiring developers to browse the Wiki to find all of the features and their 
> keys and values. Having a database that they could easily get the keys they 
> want, their values, etc. would > significantly>  allow greater OSM developer 
> potential.
>
I am unsure what would be difference.

> Specifics
> The specifics as to how this database would be arranged such as to where 
> presets/fields/tags/features go has not been thought of yet. I just wanted to 
> ask if this has ever been proposed before. If someone would like me to make a 
> DB layout to help them better understand what I'm proposing, I'd be happy to 
> do so. 
>
> My pre-DB construction proposal 
>
> Before any type of database is made, one would would > construct a dedicated 
> page structure on > openstreetmap.org >  > that 
> would display and be in-sync with all of the Features from the DB in a format 
> similar to the Wiki, and then > remove all of the features from the Wiki 
> altogether.
>
> Why?>  If you see in > Compiling and distributing the DB>  below, a Wiki bot 
> is going to be required to get all of the pages for all of the 
> already-standard features. Most of the features' pages do have a similar 
> structure as to how they are arranged (template that shows what elements they 
> use, Keys' values and their descriptions are stored in a table, etc.), 
> however these pages would be > impossible>  to parse thoroughly with a 
> computer and are going to require team of humans to get all of their data. 
>
> New features that are proposed and approved after the database would be 
> created would require them to be hand-added. Making a standard way to propose 
> and approve new features on a new part of the website with formatting 
> constraints and database-syncing abilities that MediaWiki cannot offer, would 
> mean that the DB would never have to be physically touched again and ensure 
> greater long-term DB management efficiency.
>
> So why basically delete one of the primary functions of the Wiki and create a 
> new system on > openstreetmap.org > ?
> I recently have got into the OSM community head-on in the past two months. I 
> have never really contributed to Wikipedia or any other site that uses the 
> MediaWiki website format so learning about how to contribute and get around 
> the OSM Wiki took time (learning about the proposal process, Templates, talk 
> pages, formatting, the list goes on and on...). It also made me realize that 
> this could discourage new users from ever looking into the depths of OSM or 
> even finding the Wiki at all. The WikiMedia interface is not the prettiest 
> either. It can take time for new users to explore and find what they are 
> looking for.
> (Also, this would mean moving the proposal process over to > osm.org 
> >  too)
>
> Therefore, I think creating a easily accessible, pretty, and 
> easily-contributable interface on > osm.org >  would 
> strengthen the OSM community significantly. For example, a heading called 
> "Features" could be added to the left "GPS traces". It would greet the user 
> with the primary features of OSM (kind of like the "Map features" on the Wiki 
> already does) and Feature pages would have a standard format that is 
> consistent throughout the website (unlike Wiki pages where tables for values 
> can be different, proposals can be different, etc.). 

[OSM-dev] Standardized feature/tag database idea & proposal

[Seth Deegan]

*Idea*
Has anyone ever thought of creating an official database that stores all of
the approved and in-use tags/features in OSM?

*Reasoning*
This could allow the editors (iD and JOSM, StreetComplete, GoMap!) and data
consumers (osm-carto,. Mapbox etc.), to easily stay up-to-date with new
features, without requiring their developers to browse the Wiki etc.

*Examples*
Both iD and JOSM have their own preset file/repo and are independent of
each other. Creating a database that could be used as a dependency of both
that would store these feature presets and their fields means:
 - Both are up-to-date and in-sync
 - Adding new presets could be done automatically by retrieving and parsing
data from the DB.

Creating applications that use OSM data can be hard and time-lengthened by
requiring developers to browse the Wiki to find all of the features and
their keys and values. Having a database that they could easily get the
keys they want, their values, etc. would *significantly* allow greater OSM
developer potential.

*Specifics*
The specifics as to how this database would be arranged such as to where
presets/fields/tags/features go has not been thought of yet. I just wanted
to ask if this has ever been proposed before. If someone would like me to
make a DB layout to help them better understand what I'm proposing, I'd be
happy to do so.

*My pre-DB construction proposal *
Before any type of database is made, one would would *construct a dedicated
page structure on openstreetmap.org  *that would
display and be in-sync with all of the Features from the DB in a format
similar to the Wiki, and then *remove all of the features from the Wiki
altogether.*

*Why?* If you see in *Compiling and distributing the DB* below, a Wiki bot
is going to be required to get all of the pages for all of the
already-standard features. Most of the features' pages do have a similar
structure as to how they are arranged (template that shows what elements
they use, Keys' values and their descriptions are stored in a table, etc.),
however these pages would be *impossible* to parse thoroughly with a
computer and are going to require team of humans to get all of their data.

New features that are proposed and approved after the database would be
created would require them to be hand-added. Making a standard way to
propose and approve new features on a new part of the website with
formatting constraints and database-syncing abilities that MediaWiki cannot
offer, would mean that the DB would never have to be physically touched
again and ensure greater long-term DB management efficiency.

*So why basically delete one of the primary functions of the Wiki and
create a new system on openstreetmap.org ?*
I recently have got into the OSM community head-on in the past two months.
I have never really contributed to Wikipedia or any other site that uses
the MediaWiki website format so learning about how to contribute and get
around the OSM Wiki took time (learning about the proposal process,
Templates, talk pages, formatting, the list goes on and on...). It also
made me realize that this could discourage new users from ever looking into
the depths of OSM or even finding the Wiki at all. The WikiMedia interface
is not the prettiest either. It can take time for new users to explore and
find what they are looking for.
(Also, this would mean moving the proposal process over to osm.org too)

Therefore, I think creating a easily accessible, pretty, and
easily-contributable interface on osm.org would strengthen the OSM
community significantly. For example, a heading called "Features" could be
added to the left "GPS traces". It would greet the user with the primary
features of OSM (kind of like the "Map features" on the Wiki already does)
and Feature pages would have a standard format that is consistent
throughout the website (unlike Wiki pages where tables for values can be
different, proposals can be different, etc.). Pages could also be "locked",
or require a proposal before ever changing any of the contents of their
page/feature. This would ensure the DB is secure and uniform with the
community's agreement on Features (the database is directly synced and
edited through changes on the website) and no "random edits' by users like
on the current Wiki would have to tracked (since anyone can directly edit).
There are other possible benefits that you could probably think of.

Also, other pages on the Wiki would not be deleted. There are plenty of
great pages on it that have nothing to do with the DB and work well in the
open environment the Wiki has to offer.

*Compiling and distributing the DB*

   1. One would probably use a Wiki bot like
   https://en.wikipedia.org/wiki/Wikipedia:AutoWikiBrowser to get all of
   the features with Categories "approved" and "in-use" (and "depreciated" as
   well just to let possible future editors know what to get rid of) and add
   their Keys and Values, descriptions, what