Re: [OSM-talk] bot proposal: shop values cleanup (low use values only, 1 used 250 times, three over 100 times, many used less)

[Andy Townsend]

On 20/04/2023 19:50, Mateusz Konieczny via talk wrote:

For start I want to propose to people to review shop tags in their area
with undocumented shop values or ones documented as problematic.

See http://overpass-turbo.eu/s/1u2o



Reviewing "odd" tags before tagfiddling them away seems to me a very 
sensible approach.  However, running that query locally finds, alongside 
a couple of typos by me, lots that are very much correct, but just not 
on your list - there are some very odd shops out there.


To change "shop=veryrarevalue" where it was correct to 
"shop=lessrarevalue"without preserving the detail somehow loses detail 
from OSM and is therefore by definition a Bad Thing.  Some of the 
entries on your list I'd definitely want to check onsite ("gun" and 
"firearms" is one obvious one such, but there are others).


Also, changing rare shop types into "yes" helps absolutely no-one.  If a 
data consumer wants to handle a catch-all for "shop" they can; they 
don't need them to be set to "yes" first.


That doesn't mean we should keep obvious rubbish in OSM - 
"shop=stationary" is an obvious and amusing typo for "shop=stationery".  
Single-use examples, especially featuring capital letters like 
https://overpass-turbo.eu/s/1u2R ("shop=Chandlery") are also obvious 
candidates for changing to something at least non-capitalised.


The rest, however, I really won't worry about.  I disagree very much 
with your paragraph that starts "For quite long time...", but let's not 
let that get in the way of fixing what ought to be fixed - either 
manually or automatically.


Best Regards,

Andy



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

Re: [OSM-talk] Proposed automated edit of some barrier=kerb kerb=raised nodes (forum crosspost)

[Marc_marc]

Le 19.04.23 à 13:50, Matija Nalis a écrit :

StreetComplete places kerb=raised by itself on a node where a footway meets a 
road


on the shared node between the footway and a road ?
well it's not a SC news, highway=crossing combinaison
have a high number of kerb=*, valid !
https://taginfo.openstreetmap.org/tags/highway=crossing#combinations


The bad news is that there are now about 1,500 nodes tagged barrier=kerb 
kerb=raised that really shouldn’t have the barrier=kerb tag.


I don't understand the mix

kerb=* (raised or not) without a highway=crossing
is a kerb at this location and is thus a barrier=kerb
or a missing highway=crossing [1] [2]

kerb=* on highway=crossing isn't a barrier because
it is a property that describes a nearby element (like
bin=yes on a bus stop) and is therefore 1170 errors
https://overpass-turbo.eu/s/1tZo


How would people feel about editing them all in one go, following the automated 
edit code of conduct?

I propose doing something like this

 use Overpass (query) to get all barrier=kerb kerb=raised nodes on roads


you forget all values other than kerb=raised
you forget the highway=service which have the same problem

I propose a simpler logic:
highway=crossing + barrier=* is an error -> check a series of cases
to see if it is correct to mechanically remove barrier=* (for me : ok)
For the others [3], see case by case, I have only encountered cases 
where highway=crossing is missing, removing barrier=kerb is not wrong 
but it only solves half of the problem



see if it was created_by iD


SC also add it [4]
how will you do this mechanically?

Regards,
Marc

[1] valid kerb=raised on a road 
https://www.openstreetmap.org/node/1936886800

[2] missing highway=crossing https://www.openstreetmap.org/node/6447916337
[3] barrier=kerb on road without highway=crossing 
https://overpass-turbo.eu/s/1tZx

[4] SSC adding barrier=kerb on highway=crossing



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

Re: [OSM-talk] bot proposal: shop values cleanup (low use values only, 1 used 250 times, three over 100 times, many used less)

[Andy Mabbett]

On Thu, 20 Apr 2023 at 19:50, Mateusz Konieczny via talk
 wrote:

> shop = chandler → shop = ship_chandler

In the UK at least, chandlers are more likely to cater for boats than ships.

> shop = vaping → shop = e-cigarette> shop = vape_store → shop = e-cigarette

"vape" seems to be the more widely-used term, in the UK.

> shop = collectibles → shop = collector

I have yet to see a shop that sells collectors; "collectibles" seems fine

> shop = unattended → shop = vacant

"Unattended" is not "vacant"; it can mean a place where one pays in an
"honesty box", or uses vending machines.

> shop = for_rent → shop = vacant

Or is this a shop where one rents equipment? Or an apartment?

> shop = beauty33 → shop = beauty

Is "Beauty33" a name or brand?

> shop = herbs → shop = herbalist

Not necessarily the same; a herbalist does not sell cooking herbs.


[No view on other values]


--
Andy Mabbett
@pigsonthewing
http://pigsonthewing.org.uk

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

Re: [OSM-talk] bot proposal: shop values cleanup (low use values only, 1 used 250 times, three over 100 times, many used less)

[Marc_marc]

Le 20.04.23 à 20:50, Mateusz Konieczny via talk a écrit :

bot proposal: shop values cleanup


ok


For start I want to propose to people to review shop tags in their area
with undocumented shop values or ones documented as problematic.
See http://overpass-turbo.eu/s/1u2o 


I find that this point has no place in a mass edition and it makes
your proposal less understandable
making 2 topics, one with mechanical editing of cases that do not 
require individual review and the other with a call for review of 
objects, would be more readable



tags with highest use, among ones that will be retagged
shop = chandler with 113 uses

..

and witch fix for thoses values ?



shop = local_shop → shop = yes (though looking at
https://www.openstreetmap.org/node/6771559662/history 
 and other - maybe

this import should be reverted due to dubious quality?)


I don't see the issue, addr:ward ?
let's talk about it in another thread


shop = retail → shop = yes
shop = Retail → shop = yes
shop = Retails → shop = yes
shop = generic → shop = yes
shop = ??? → shop = yes
shop = retailer → shop = yes
shop = retails → shop = yes


I think some of this may stem from the controversial decision to hide 
shop=yes in osmcarto, a decision I find incomprehensible. seeing that
a shop exists is better than seeing nothing, even if the type of shop is 
not yet known (there is a rendering for highway=road)

not against... but we are regressing as much as we are progressing


add/remove s as needed - can I do this for all shop values where such
modification will change them to a searchable shop value present in iD
presets, also without going through review like this one?


"add/remove s as needed to match more common/preset"
is for me an acceptable rule


with trailing space at the end - would it be fine to do it also
with other known valid values (listed on Wiki or in iD presets as valid)
if they appear, without a separate bot proposal?


removing all trailing space is for me an acceptable rule


would it be fine to do it also with other known valid values
(snip '_shop', ' shop', ' store', '_store', '_products', ' products'


you meaning removing space and _ at both begin and end ?
if so yes


etc at the end of shop value if it produces valid shop type


and also if it's not an already documented/preset value


shop=patisserie


shop=bakery


shop=bijouterie


shop=jewelry

Regards,
Marc



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

[OSM-talk-fr] [édition mécanique] supprimer barrier=kerb sur highway=crossing en France

[Marc_marc]

Bonjour,

sur la liste talk, un utilisateur signale un problème avec barrier=kerb
après analyse, tant StreetComplete que iD ont dans le passé
ajouté erronément barrier=kerb sur les objets ayant l'attribut kerb=*
décrivant les caractéristiques d'un passage piéton highway=crossing
Les bugs sont entre temps corrigé dans les 2 éditeurs.

cela pose problème parce que cela donne décrit
qu'il y a une bordure à franchir sur la voie "véhicule".
c'est d'autant plus problématique avec kerb=raised
que je ne voudrais pas franchir à vélo :)

proposition : supprimer barrier=kerb sur les objets highway=crossing
en France

Avis ?

Cordialement,
Marc



___
Talk-fr mailing list
Talk-fr@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-fr

Re: [OSM-talk-fr] Le forum et la liste

[Marc_marc]

Le 20.04.23 à 17:03, Vincent de Château-Thierry a écrit :

Marc qui a voté sur le forum https://localhost, 
ben quoi faut fragmenter...


Je rebondis sur cette remarque (pique ?) de Marc au sujet du forum et de la 
fragmentation


ce n'est pas une pique, la vie est trop courte et le temps trop précieu 
que pour passer du temps à cela.

c'est une désolation de voir qu'on se tire une balle dans le pied
trop souvent.

quand je parle de fragmentation, je ne joue pas à qui à la plus grande 
(ce qui est évidement facile quand on spam à tout va, j'ai moi-même

un compte sur le forum qui me spam de "résumé des sujets" alors
que je ne m'y suis jamais inscrit, comportement digne des spammeurs pro)
ou du fait que le moi passé une entreprise me signalait qu'on leur
avait signé que cette liste oü j'écris en ce moment avait été fermé (!)

non je parle plutot du fait :
osmf a une liste de diffusion... osm-fr a la sienne
osmf avait un forum... osm-fr avait le sien
osmf a décidé de basculer sur discourse... osm-fr installe vite le sien
osmf a un canal de discussion irc bridgé sur Element... un groupuscule 
fait un 2ieme canal exactement identique

chacun dans son coin, ne surtout pas se regrouper.

pendant ce temps à l'avant dernier CA, l'asso se demande comment
faire pour sortir de la saturation des sysadmins. un exemple parmi 
d'autres :

le rendu humanitaire a ses maj cassé depuis... 6 mois !
non c'est pas une erreur de frappe, c'est en mois, pas en jours !
c'est un sysadmin osmf qui est venu le signaler parce qu'en interne
on est sous-eau.
on parle de faire appel à une entreprise externe pour gérer l'infra.

mais si on arretait ce concours de "il nous faut une copie
de ce que le monde a sans plusvalue ?"
est-ce que osm-be a un serveur de liste de diffusion ?
non ils utilisent celui d'osmf.
et un discourse ? non ils utilisent celui d'osmf
est-ce qu'ils ont 2 canaux Element ? non un
est-ce que bzh a une serveur de diffusion ? non c'est chez osmf
est-ce que la liste de transport internationale a son serveur ?
non c'est sur celui d'osmf
Si on se concentrait sur ajouter de la plusvalue au lieu
de dupliquer pour en avoir "un à soi" ?

en plus discourse est prévu pour fusionner l'interface web-email
pour un même contenu, bien plus malin que d'avoir des discussions
fragmentée entre 2 lieux sur 2 interfaces incompatibles.
mais pour cela il fallait laisser un peu de temps au temps,
rien ne pressait d'ajouter une interface web (que j'appelle
mickey-mouse non pas péjorativement mais parce que je trouve
que cela met la mise en forme devant le contenu) à ce qui existait.
du moins pas si en même temps on est sous-eau avec des
choses qui me semble, à mon avis, prioritaires.

il y a quelques jours je découvre une nouvelle appli...
qui demande elle aussi de traduire à nouveau les mêmes
clefs que plein d'autre applis... chaque appli se doit
d'avoir son propre transiflex, on va quand même pas
faire confiance à la traduction de kerb=* faite ailleurs,
faut tout recommencer encore et encore depuis 0.

demain je suis interviewé par des étudiants.
l'une des question : Comment cette situation peut être améliorée / 
renforcée dans deux ans. Avec quels moyens et pratiques ? (en lien
avec la question précédente Comment le citoyen (individu, entité privée 
ou publique) participe et contribue en 2023 à cette initiative ?),


cela me fait pennser à l'exemple que j'avais déjà posté du temps
de "cro" :
on a plein d'éditeurs dans tous les sens (ce qui est aussi une richesse)
et aucun n'est finit pour que l'entrepreneur référence son commerce
de manière éfficace, iD n'a pas d'éditeur d'horaire, osmmybiz
n'a pas accès à l'opendata, et plein d'autre sont relégé aux 
connaisseurs, invisible depuis osm.org qui ne renseigne que id/josm

même les règles de validation sont à dupliquer dans tous
les éditeurs, seul osmose partager une partie des règles avec josm.

je leur dit quoi ?
j'espère sincement que la commuanuté puisse un jour réalisé
qu'elle est elle-même son propre fossoyeur, trop oqp à
savoir si highway=track est une piste agricole ou autre chose,
trop oqp à courrir après avoir "la plus grosse" dans n'importe
quel domaine...

et pendant ce temps Ouverture parle d'une même voix
à ses utilisateurs de données...
ils ont même réussit, pour une entreprise américaine,
à normaliser height=* en mètre, c'est quand même plus facilement
utilisable que de laisser chaque utilisateur faire les conversions...
faite une utilisation de la hauteur des batiments pour rigoler...


inscrits (pour un total de 3000


tu compares des pommes et des poires
1) je ne suis pas inscrit mais je compte dans tes 3000
combien d'autres ont été inscrit contre leur volontée ?
2) si 1000 personnes s'inscrive sur le forum et meurent,
cela te fait +1000 dans tes stats.
sur la liste, ils vont se faire désincrire automatiquement
et cela fait +0
Je parle aussi des messages style "je ne sais pas mais je l'écris"
du forum ? ou "salut moi c'est Jean"
bref, la course au nombre... ce qui ne 

[OSM-talk] bot proposal: shop values cleanup (low use values only, 1 used 250 times, three over 100 times, many used less)

[Mateusz Konieczny via talk]

bot proposal: shop values cleanup (low use values only, 1 used 250 times, three 
over 100 times, many used less)

For quite long time I am trying to use OSM-based products as Google
Maps replacement. One of major issues are POIs (in many apects). Small
part of that are POIs marked but in way that makes them unusuable
anyway. This is also problems for mappers, especially newbies, confused
for example why nice icon is not appearing on some (and problem is for
example shop=hair_dresser vs shop=hairdresser).


For start I want to propose to people to review shop tags in their area
with undocumented shop values or ones documented as problematic.

See http://overpass-turbo.eu/s/1u2o

For each case either shop should be either


(1) retagged and shop=* changed
(2) such shop value should have its value documented at OSM Wiki (I
documented some, see for example
https://wiki.openstreetmap.org/wiki/Tag:shop%3Dcatalogue ) (3)
sometimes new value should be invented, documented and shop=* retagged
to it


https://community.openstreetmap.org/c/general/tagging/70 may be useful
for discussing new shop=* values (local discussion channel may be also
useful, but I strongly recommend asking wider community about new
values to avoid avoidable confusion). Some people go through
https://wiki.openstreetmap.org/wiki/Proposal_process - but
discussion/review step is the most useful one and you can use just this.


Tagging mailing list also exists and can be used for discussing new
tags.


https://wiki.openstreetmap.org/wiki/Creating_a_page_describing_key_or_value
may be also useful.


But some of shop values can be safely automatically replaced by another
shop value. For example shop=shoe can be safely migrated to shop=shoes
without human review.


-


Getting to the bot edit itself (and I want to note that I am more
excited about finding missing shop values and documenting them and
adding them to presets/documentation than I am about retagging):


So I am proposing to extend
https://wiki.openstreetmap.org/wiki/Mechanical_Edits/Mateusz_Konieczny_-_bot_account/fixing_malformed_shop_tags
by adding more tag replacements.


Please let me know if any of replacements here are dubious and values
require human review/survey to be replaced or are actualy valid. I know
that list is long, so if someone wants to review but needs more than 2
weeks - please write and I can wait for longer.


Also, let me know if anyone would want to get list of affected objects
for review or manual retagging or listing of edits that added this tags
and so on.


tags with highest use, among ones that will be retagged 
shop = chandler with 113 uses
shop = stationary with 116 uses
shop = hardware_store with 60 uses
shop = lamps with 250 uses
shop = knife with 60 uses
shop = unattended with 87 uses (see 
https://www.openstreetmap.org/changeset/130756523 - this mapper added
all* of them and is fine with such change 
*including one as a typo, that is why another mapper may
be credited with it)
shop = local_shop with 53 uses
shop = retail with 145 uses

shop = chandler → shop = ship_chandler
shop = chandlery → shop = ship_chandler
shop = chandlers → shop = ship_chandler
shop = stationary → shop = stationery
shop = hardware_store → shop = hardware (Note: there are weird clusters of
shop=hardware in some places, but that is a bit different story -
I suspect some systematic mistake or bad mapping, unless there are African
towns where 1/4 of all shops are really shop=hardware - though either way
local on the ground survey seems needed)
shop = vaping → shop = e-cigarette
shop = vape_store → shop = e-cigarette
shop = vape → shop = e-cigarette
shop = Vape_Store → shop = e-cigarette
shop = lamps → shop = lighting
shop = lamp → shop = lighting
shop = Lighting_Shop → shop = lighting
shop = knife → shop = knives
shop = collectibles → shop = collector
shop = unattended → shop = vacant
shop = for_rent → shop = vacant
shop = unused → shop = vacant
shop = vacancy → shop = vacant
shop = local_shop → shop = yes (though looking at
https://www.openstreetmap.org/node/6771559662/history and other - maybe
this import should be reverted due to dubious quality?)
shop = retail → shop = yes
shop = Retail → shop = yes
shop = Retails → shop = yes
shop = generic → shop = yes
shop = ??? → shop = yes
shop = retailer → shop = yes
shop = retails → shop = yes (again 
"SUZA Indusrtial training Resillence Academy" but this suspect data
will be more detectable as shop=yes - see say
https://www.openstreetmap.org/node/6771699918)
shop = misc → shop = yes
shop = commercial → shop = yes
shop = Generic shop → shop = yes
shop = true → shop = yes
shop = Retail Shop → shop = yes  
shop = miscellaneous → shop = yes 
(second_hand / variety_store / catalogue / department_store etc may fit)

shop = miscelanea → shop = yes
shop = bazaar → shop = yes
shop = samoobsługowy → shop = yes (Polish translation)
shop = fixme → shop = yes 
shop = egg → shop = eggs 
(both undocumented for now, but 

[OSM-talk-fr] Le forum et la liste

[Vincent de Château-Thierry]

Bonjour,

- Mail transféré -
> De: "Marc_marc" 
> À: "Discussions sur OSM en français" 
> Envoyé: Lundi 17 Avril 2023 12:54:43
> Objet: Re: [OSM-talk-fr]  traduction de "tag" en français dans JOSM

> Marc qui a voté sur le forum https://localhost, ben quoi
> faut fragmenter...

Je rebondis sur cette remarque (pique ?) de Marc au sujet du forum et de la 
fragmentation, pour donner quelques éléments factuels sur le forum 
https://forum.openstreetmap.fr/

En 1 année glissante (20 avril 2022 -> 20 avril 2023) le forum a vu :
- la création de 4400 sujets, soit environ 12/jour
- l'arrivée de 600 nouveaux inscrits (pour un total de 3000, soit 4x plus que 
la liste talk-fr)
- la publication de 11000 posts soit environ 30/jour

Sur la même période on compte 1062 messages sur cette liste, à peine 3/jour.

Évidemment chacune, chacun a ses habitudes de lecture, ses préférences en 
termes d'ergonomie, et je ne suis pas là pour en juger. Mais poster sur le 
forum, en 2023, n'est pas fragmenter, au contraire : c'est aller à la rencontre 
du plus grand nombre, se donner le plus de chances de recevoir avis, arguments, 
conseils.

Voilà, sans polémique, mais pour mettre un peu à jour le paysage de nos canaux 
de communication, notamment pour celles et ceux qui ne consultent que la liste.

merci
vincent

___
Talk-fr mailing list
Talk-fr@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-fr

Re: [Talk-ee] SOTM Baltics 2023

[Rihards]

Heya, as we discovered some time ago, we're all such map geeks that we
didn't know there's Ice Hockey World Championship happening in Riga in May.
To make travel and accommodation possible, the conference was moved to June
8-9.
The location is also back to the premises of the University of Latvia,
https://www.openstreetmap.org/way/361765048 (SOTM-B page updates pending).

We'd be happy to meet you all in Riga this summer, and eager to hear about
your experience with OSM - please don't hesitate to submit talk proposals
at https://2023.sotm-baltics.org/ .

On Tue, 14 Feb 2023 at 12:34, Rihards  wrote:

> Something wonderful is going to happen - the Baltic GIT conference returns
> to Riga, and so does SOTM Baltics.
>
> These events will take place on May 18th and 19th.
>
> We'd be very happy to see you all in Riga - OSM community, GIT community,
> students, teachers and everybody else.
>
> If there's something you'd like to share, we'd be very happy to see you as
> a speaker - please do submit a proposal.
>
> Please see https://2023.sotm-baltics.org/ for more information.
> --
>  Rihards
>


-- 
 Rihards
___
Talk-ee mailing list
Talk-ee@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk-ee