Mikel, I might be misunderstanding what you meant, but in my opinion
conformity is required for this type of project, and I do hope iD/JOSM/...
help us achieve that. To clarify:
* features with the same meaning (type) should be mapped the same way,
otherwise each consumer must understand all of
I second Jóhannes -- every dataset, including OSM itself (hehe) has errors.
Consuming each additional dataset is a complex task -- each dataset has its
own structure and conventions, thus the fewer datasets one has to work
with, the better. The fundamental problem with 99.9% of the datasets
It is very strange that we, on one hand, allow anyone to create any kind of
tags (just type it in), and on the other we create so many hurdles to
document it (we refuse to allow a wiki page about an item, but instead
demand that each key page go through a proposals process, approve it,
Christian, I would like to add torrent support to the download-osm tool
. While I could try to scrape
https://osm.cquest.org/torrents/ , I would obviously rather use the
structured xml file (or if you could provide a JSON file, even better).
* get the catalog file (xml/json)
I am in favor of this or similar language. I think for a more vote-like
discussion it might be better to use the wiki talk page (easier to add +1s
and short comments).
On Sat, Feb 8, 2020 at 2:59 PM stevea wrote:
> I don't know if here or https://wiki.osm.org/wiki/Talk:Good_practice is a
something we need to document in OSM, but it highlights that language
is not homogeneous with naming and just makes a fun story :)
On Sat, Feb 8, 2020 at 6:22 AM Rory McCann wrote:
> On 07.02.20 20:22, Yuri Astrakhan wrote:
> > (e.g. two fairly large groups of people could refer to
Thanks Stevea, I really liked your examples. And thank you Mikel - I agree.
OSM already has substantial amount of non-physical but relevant information
(e.g. many IDs pointing to external registries), and as Stevea points out -
even naming for something local could be contradictory (e.g. two
Andy, I agree that being frugal with bandwidth is important. Yet, there is
a significant operations cost involved here, that I suspect very few will
actually be willing to spend, unless it is made trivial -- the cost of
setting up an independent planet file update pipeline - i.e. a docker image
Andy, two major reasons:
* Anyone working on an evolving project like OpenMapTiles would attest that
the import schema constantly changes. Every time schema changes, one needs
to download newest planet, import it based on the new schema, and run diffs
from that point.
* Automation / easy
The news mentions that downloads from Planet OSM are currently rate limited
to 400 kB/s and suggest to use mirrors, but does not mention the related
announcement about the new tool to simplify such downloads. I think it will
help anyone downloading, and it might be worth including in the next
For those who download OSM data regularly, there is now a simple way to
reduce the load on the primary OSM servers, while also making download much
faster and ensure the data is correct.
OpenMapTiles new tool downloads the planet from all mirrors in parallel. It
usually takes just a few minutes,
check your OSM settings. AFAIK, iD is the default editor.
On Mon, Dec 23, 2019 at 2:10 AM Sören Reinecke via talk <
> so far I know currently Postlatch is the default editor on osm.org .
> Since it needs Flash to run and most users do not have Flash
I would love to fully transition to data items, as it would make
translation and cross-language tag consistency far easier.
We now have a data item editor you can enable by going to preferences /
gadgets, and enabling it.
Stevea, I think this discussion mixes two topics, as Martin pointed out:
* I want to be credited for my work (i.e. you couldn't have done it without
me, just say so)
* I want to control what you do with the results of my work (i.e. you must
not kill baby seals using the map I created)
stevea, I would not be exactly the same person without OSM. Does it mean
ODbL applies to me? A hammer was used to build a house, but the house does
not have hammer's copyright. Just because some data was used in the process
does not necessarily mean that whoever saw that data taints everything
Let me get this straight:
* I create a dataset from public data sources, e.g. a list of roads, and
publish it under the Public Domain dedication (i.e. CC0). (I agree that
MIT is weird here).
* Afterwards, I make a subset of my original data by removing any roads I
found elsewhere, e.g. in a
I agree with Christoph -- every tag used in OSM data must be documented --
otherwise it has near-zero value.. Actually negative value because it
confuses people -- some might want to delete it, but they don't know if it
is useful, so they just leave it there almost indefinitely.
In an ideal world
You can use Sophox to pull this type of information, possibly just not all
Here's a query that will give you all first level country sub-divisions
(Canada in this case), as well as each province's Captial, Flag image, and
(this query was modified
Also, fully accurate data is a myth, even if we only have 1%
completeness. Once data is beyond a certain size, it is guaranteed to be
wrong, simply because humans always make mistakes and things always become
outdated. We can only discuss how close we are to the ideal "perfect
accuracy", and what
On Mon, Jul 29, 2019 at 6:19 AM Martin Koppenhoefer
> speaking about risks, having an incomplete network of verified, correct
> roads is probably more useful and less troublesome than an "overcomplete"
> one which also contains non-existent roads (e.g. waterways interpreted as
> roads) or
On Tue, Jun 11, 2019 at 12:23 AM Mateusz Konieczny
> Pull requests work and that is exactly scheme used now - by iD, JOSM,
> Vespucci, StreetComplete etc.
Ahem, actually no, they do not -- as demonstrated by the recent scandal
with iD presets. Community is disconnected from
On Mon, Jun 10, 2019 at 8:08 PM Mateusz Konieczny
> Also, "edit on Wiki automatically and silently starts equivalent of PR" is
> also not going to work.
Why a pull request won't work?
talk mailing list
On Mon, Jun 10, 2019 at 8:03 PM mmd wrote:
> This idea was recently discussed on Slack US #id channel. To quote Bryan
> on this: "We will never use the wikibase for our presets, sorry".
mmd, this was recently discussed in the iD sync-up call, and Bryan was much
more welcoming to the idea :)
> Count on my support to go further with them.
> Le lun. 10 juin 2019 à 18:38, Yuri Astrakhan a
> écrit :
>> (please fix them in the wiki pages in the corresponding languages, the
>> bot will automatically update the data items)
> What if we
Another big advantage is that various tools can use this information
directly from the wiki, without any 3rd party sites - i.e. iD editor gets
new localized descriptions and images the moment they are updated in the
wiki - and we could store other info like validation rules and even solve
> I think we need to discuss if this is desired, before any more time is
> spent on adding all of those data items.
Joseph, I agree that ideally we should not have duplicates anywhere.
Currently we have thousands of mismatches between languages - e.g. in
statuses and what key/tag should be
On Fri, May 10, 2019 at 3:39 PM Yves wrote:
> Some validation tools, like Osmose, make great efforts to maintain a
> 'false positive' database.
If the same validation is done by multiple tools, they need to share the
"false positive" data, otherwise only one tool would know not to change
If the whole issue is optimizing search results, lets just create an
"Archive" namespace that is not included in search by default. Moving to
archive is different from deleting because only admins can see deleted
On Mon, Apr 22, 2019 at 4:11 AM Lester Caine wrote:
> On 22/04/2019
Interestingly enough, FOSS4G San Digeo where I was just attending has had
several talks about wikidata IDs usfulness in OSM... Also, AFAIK, mapbox
is using WD in brands to do their searches for POIs, as that is much easier
to consume than names. OpenMapTiles have discussed wikidata ids at length,
Andy, you could also use one of the example queries in sophox.org to find
all objects that are usually tagged with brand:wikidata=* (e.g. have more
than 10 instances of that), and find all other objects that use that same
value for wikidata tag. The query currently finds over 15,000 such
Thx Mark, makes sense. Data items are wiki pages too, and we can protect
them the same way. Still, I think it's a moot point -- at least initially
preset data will go via GitHub as regular pull requests.
On Thu, Apr 11, 2019, 05:34 Mark Wagner wrote:
> On Thu, 11 Apr 2019 03:17:27 -0400
On Thu, Apr 11, 2019 at 2:10 AM Mateusz Konieczny
> * easy to edit by community
> I am dubious whatever "anybody can
> edit any preset stored as wikidata
> items" will be considered as benefit
One could also doubt that allowing direct OSM and Wikipedia edits by anyone
On Wed, Apr 10, 2019 at 5:04 PM Rory McCann wrote:
> In JOSM, people, or groups, can make their own tagging presets. AFAIK iD
> unfortunately doesn't have this feature. If it did, the iD version on
> openstreetmap.org could be configured to something special, people could
> have their personal
anto or Irish or something.
> If you want real fun, just talk about "tabling an agenda item"
> On 08/04/2019 18:32, Yuri Astrakhan wrote:
> > Martin, thanks for explanation, but my point still stands -- in tags, we
> > treat words not at their own meaning, b
at 3:50 AM Martin Koppenhoefer
> sent from a phone
> > On 7. Apr 2019, at 22:23, Yuri Astrakhan
> > A good example is "denomination=evangelical" -- German speakers should
> not use it for "evangelisch" which sta
On Sun, Apr 7, 2019 at 1:48 PM john whelan wrote:
> Tagging is not always easy, highways in Africa are an example. See a dirt
> track in Europe and its probably a highway=track. In Africa it probably
A "road is a highway" is confusing because the word "highway" has a
Tobias, thank you for clarification. The lang parameter was very often
incorrect because people copy/pasted it without changing. I did a minor
clean up with a manual inspection of every edit - similar to many of my
previous edits (by "yurik" account), but with the bot+minor flag to avoid
Tag:oneway=reversible was using image Baustellenampel.jpg because English
page had no image, and the Polish translation used it, so it got copied
form PL to the Item:Q5736 and became the default for all languages. If
that image is incorrect, it should probably be removed from the Polish
Thanks to a kind donation by Elastic (of Elasticsearch fame), we now have a
much cleaner and more stable Sophox service.
Sophox is an RDF database that contains:
* All of OSM data without geometries
* All of OSM Metadata (keys, well known tags, etc) as defined in OSM Wiki
items (with all
this whole thing.
On Fri, Nov 23, 2018 at 2:24 AM Frederik Ramm wrote:
> On 23.11.2018 01:42, Yuri Astrakhan wrote:
> > One idea (perhaps this should go into a separete thread):
> There already is a separate thread over on the tagging list started just
Victor, thank you for a very thorough and accurate analysis. I especially
appreciate 4.2 -- the fact that this decision is very bad for the data
users -- I would not be able to use this data at all, simply because most
of the time one needs to render the map from the perspective of the
Stefano, thanks! One important aspect - it seems a single generic request
with guidance gets done much better than multiple smaller items. I highly
recommend uniting behind the general "improve maps" request to get WMF to
commit to maps full time, rather than "throwing us a bone" (creating a
On Mon, Oct 22, 2018 at 8:22 AM Mateusz Konieczny
> I think a country relation should describe how the specific country think
> of its borders. So if two countries claim the same territory, those two
> relations will overlap.
> That is absurd and conflict with OSM rule to map what
I think a country relation should describe how the specific country think
of its borders. So if two countries claim the same territory, those two
relations will overlap.
While not ideal, this is preferable for many data consumers - when
generating a map, one always has to consider whom it is
Michael, thanks for your comments!
The original searchbox functionality has been restored, sorry for the
delay. See also some related requests to make it even more useful. 
I have made some changes to the OpenStreetMap:Wikibase wiki page - could
you take a look? I would love your help in
On Thu, Sep 27, 2018 at 9:50 AM Mateusz Konieczny < matkoni...@tutanota.com
> Main point of separate presets is that creator of an editor has control
> over it.
Mateusz, who should control an app's behavior - the developer or the
community? Can app make certain editing choices
Mateusz, I think we have a chance of success this time because unlike other
rulesets that are stored in the specific project's GIT repos, this project
* wiki based and centralized, just like the tag documentation, and
accessible to a much wider community. In the last week, we had 18 users
Yep, that's probably because Tom has merged the settings change :)
Thanks for reporting it!
On Mon, Sep 24, 2018 at 1:06 PM OSM Volunteer stevea <
> Well, I'm no longer seeing the Lua errors I saw, so "caches cleared" (all
> the way down) and the problem seems
On Sun, Sep 23, 2018 at 6:34 PM Christoph Hormann wrote:
> > The content stored in Wikibase is editable by wiki users using a
> > regular wiki account, same as with other wiki pages.
> This i have no doubts about - but this is not the question. The
> question is who is de facto in control of
On Sun, Sep 23, 2018, 14:58 Christoph Hormann wrote:
> For clarification: The main purpose of the OSM wiki is to allow mappers
> to document tags they use and allow them to coordinate and communicate
> about this use of tags with other mappers.
If you have some data
On Sun, Sep 23, 2018 at 2:36 PM Richard wrote:
> > Key:bridge:movable:https://wiki.openstreetmap.org/wiki/Item:Q104
> > Tag:bridge:movable=bascule:
> how do I get at other language's versions, and how do I get easily at the
> original source
I think the Lua crashing is related to this discussion:
> You set $wgScribuntoEngineConf['luastandalone']['cpuLimit'] too low.
Either raise it or unset
Everyone, please ignore the "don't edit or translate the label"
restriction. There is now a dedicated ID P16 property.
We can change label to be more useful to the new editors, e.g. "name in
English" instead of "name:en"
On Thu, Sep 20, 2018, 18:01 Shu Higashi wrote:
> Hi Yuri,
as of today, OSM Wiki can store structured tag metadata similar to
Wikidata. In every possible language, cross-linked, with images,
validation rules, or anything else the community decides to store there.
I'm a big fan of plus codes, and even have a pending implementation of it
in the Elasticsearch (as an aggregation hashing function). I doubt there
are any legal restrictions on using this - the code is licensed under
Apache 2, and Google states "Plus codes are free. There are no licensing
> It generates an Overpass Query showing all the objects related to the
> Wikipedia entry via wikidata tags in the OSM data.
> Op zo 29 jul. 2018 om 15:01 schreef Yuri Astrakhan <
>> Hi everyone. Thanks to Tom Hughes, we now ha
> Hi everyone. Thanks to Tom Hughes, we now have Scribunto extension set up
> on OSM wiki, which allows Lua language in addition to the very slow and
> unreadable wiki template language.
> * https://www.mediawiki.org/wiki/E
Hi everyone. Thanks to Tom Hughes, we now have Scribunto extension set up
on OSM wiki, which allows Lua language in addition to the very slow and
unreadable wiki template language.
* The BIG part of the announcement is that OSM data is now being used to
create truly multilingual map with a very large exposure. I hope some
people here are not saying that such project should not have been built? If
you want to help, see at the end.
* I do not think names should be simply
Marc, Oleksiy, thanks for your insights! And Marc, I agree that it should
always be up to the local community to decide. The only thing I ask is to
please keep in mind that the "default_language" is simply a reflection of
what local OSM editors have already used for the name tag in the majority
t; to compare with on those objects.
> 2018-05-10 1:10 GMT+02:00 Martin Koppenhoefer <dieterdre...@gmail.com>:
>> sent from a phone
>> > On 10. May 2018, at 00:47, Yuri Astrakhan <yuriastrak...@gmail.com>
languages are not very helpful for the
truly multilingual maps.
On Thu, May 10, 2018 at 2:10 AM Martin Koppenhoefer <dieterdre...@gmail.com>
> sent from a phone
> > On 10. May 2018, at 00:47, Yuri Astrakhan <yuriastrak...@gmail.com>
countries the regions are not so clearly
> 2018-05-09 2:37 GMT+02:00 Yuri Astrakhan <yuriastrak...@gmail.com>:
>> Polyglot, thanks! I just ran the list of names for Belgium -
>> http://overpass-turbo.eu/s/yEj (takes a few minutes and 20MB download
> sent from a phone
> On 9. May 2018, at 02:37, Yuri Astrakhan <yuriastrak...@gmail.com> wrote:
> or to two languages that match "Bruxelles - Brussel" ("fr - nl" ?).
> The hyphen/dash is no
ench only and
>> multi-lingual ones: https://overpass-turbo.eu/s/yE5 (thx trigpoint &
>> FredrikLindseth on IRC!) Do you want to change it, or should I?
>> Also, there are still about 60 countries without a tag:
Also, there are still about 60 countries without a tag:
On Tue, May 8, 2018 at 10:59 PM Daniel Koć <daniel@koć.pl> wrote:
> W dniu 08.05.2018 o 21:31, Yuri Astrakhan pisze:
> > This query shows a list of regions that have the new default_language
> > t
page at the
P.S. same query, but with the last editing user included:
On Tue, May 8, 2018 at 12:02 AM Yuri Astrakhan <yuriastrak...@gmail.com>
> Hi, most countries now have a default_language  tag, specifying the
> most likely language of th
Hi, most countries now have a default_language  tag, specifying the most
likely language of the "name" tag in that region. Here's a list of 60+
countries that have multiple official languages. If you have local
knowledge, or can research it, could you add the proper default_language
at kind of data is provided).
> 4. May 2018 10:14 by yuriastrak...@gmail.com:
> Btw, this is already possible - Wikipedia servers let you access both
> images and raw vector tiles, so if someone wants to do the client part, it
> shouldn't be too hard.
> On Fri, May 4, 2018, 1
On Sat, May 5, 2018 at 2:25 AM Janko Mihelić <jan...@gmail.com> wrote:
> sub, 5. svi 2018. u 00:18 Yuri Astrakhan <yuriastrak...@gmail.com>
> napisao je:
>> Tag description:
P.S. Janko, please take a look at the single vs multiple languages per
region in that wiki page. Does that make sense?
On Wed, May 2, 2018 at 1:01 PM Janko Mihelić <jan...@gmail.com> wrote:
> pon, 30. tra 2018. u 08:28 Yuri Astrakhan <yuriastrak...@gmail.c
Btw, this is already possible - Wikipedia servers let you access both
images and raw vector tiles, so if someone wants to do the client part, it
shouldn't be too hard.
On Fri, May 4, 2018, 11:11 Yuri Astrakhan <yuriastrak...@gmail.com> wrote:
> In reality, it is not impossible, or even
In reality, it is not impossible, or even that hard. If the vector tile is
sent to the client, than the client can decide which language to render
be used to decide the labeling. Caching would only improve, because instead
slower generation and rendering, and bigger cache fragmentation. See
various Kartotherian repos in github for more information.
-- Yuri / @nyuriks
On Fri, May 4, 2018 at 4:14 AM Daniel Koć <daniel@koć.pl> wrote:
> W dniu 04.05.2018 o 02:30, Yuri Astrakhan pisze:
> > Daniel, t
Daniel, the only real difference between serving every available language
and serving just one is cache fragmentation, and that's may be different in
The vector tiles get pre-generated the same way, and they simply contain
all languages instead of one. Disk space wise, the difference
I cannot speak for WMF, only about the actual Kartotherian stack behind it,
and the way they are currently using it:
On Fri, May 4, 2018 at 12:52 AM Daniel Koć wrote:
> 1. The localized maps lack fallback rules (I'm speaking of Polish
> language at least). I would ask for
Japanese will be transliterated as Chinese Pinyin. I couldn't find a better
way to distinguish Chinese Hanzi and Japanese Kanji. So if you would like
to romanize Japanese Kanji, please consider kuroshiro."
On Fri, May 4, 2018 at 12:
Christoph, I agree that this would be an awesome improvement, yet I think
there is a problem to implement it. Most languages have their own
transliteration rules, so transliterating "name" tag without the knowledge
of its language will produce a lot of incorrect names.
I have posted in another
Multiple semicolon-separated values do not solve the main problem -
figuring out the language of the "name" tag. If a region uses one value in
the name tag, "default_language" should be set to just one language.
If the whole region uses "xx - yy" convention in the name tag,
Adding a language=xx to each feature seems excessive, and will be forgotten
most of the time, unless there is some extensive tool support for it.
Adding it to admin regions seems like a better approach. Some utility
could then calculate a clean translation map, using admin_level number as
Pedro, thanks for all your efforts! I heard that Wikipedia just launched
its own 3D file storage, functioning in the similar fashion as their other
files (community curated images/sounds/videos, proper license and
attribution, and file version control). Do you see your site being in a
for the OSM community. How do we get to such a consensus about tagging
> patterns, that we can actually say "this is correct" and "this is wrong
> enough to warrant correction" without upsetting a large number of people?
> As soon as a discussion is about right vs wro
Declarative rules are usually not very good. Every tool must understand
every type of rule, and must be updated when new rule types are introduced.
Plus declarative grammar is either too limiting, or eventually starts
looking like a scripting language itself, and we end up building an
Permanent IDs has been brought up several times, especially as part of the
Wikidata ID discussion. I started a wiki page to outline the requirements
and goals, but it might be incomplete, feel free to add / correct / comment.
Once we reach the
Christoph, unregenerate implies I should apologize for doing a wrong thing.
In the discussion, the only thing I **actually did** was I wrote a new tool
and posted about it. Was I wrong to write a tool? Was I wrong to discuss
it with the community?
I patiently sifted through all the negative
Richard, in both languages, the main issue is the same. It says that the
discussion has restarted with the negative commentary, but skips the main
point - that the tool has been substantially reworked based on community
feedback. It's like saying some people got rich without mentioning the
Thanks Frederik. This is a good explanation. Can some of it perhaps be
added to the document to make it clearer?
On Wed, Nov 22, 2017 at 6:22 AM, Frederik Ramm <frede...@remote.org> wrote:
> On 22.11.2017 04:16, Yuri Astrakhan wrote:
> > Pierre, I suspect the num
ets would facilitate the follow-up.
> Actually, not all instances of the Tasking Manager add an hashtag to
> document the host and project no. For QA tools, specific projects /
> missions are not documented either.
> Le mardi 21 novembre 2017 21:21:55 HNE
While this might not have been the intention, the
> b) directed by a third party exactly what and how to contribute to
can be applied to any "challenge style" sites such as the MapRoulette or
Osmose. I think there should either be a clarification about this, an
Martijn, I think this is very similar to what Osmose does in its DB view
(and I think several other tools do in the map view) - they offer a choice
of a "world view" (unfiltered) or a "region views" - where users may choose
what region they are interested in (e.g. a dropdown).
I think it would be
> On 18 November 2017 at 13:26, Yuri Astrakhan <yuriastrak...@gmail.com>
>> John, not trusting a brand name and being unreasonable about new project
>> are two different things. One is a healthy caution. The other is a
>> baseless witch hunt
> How does that break "unwritten rules"?
> It relates to trust and politics with a small p. Your brand name is
> Cheerio John
> On 18 November 2017 at 13:11, Yuri Astrakhan <yuriastrak...@gmail.com>
James, this is not about hurt feelings. This is about misrepresentation.
Last week I re-wrote Sophox tool based on the community feedback. The new
tool uses the same approaches as existing tools. Yet, somehow I violated
some unwritten rule by creating a new tool? This is bogus.
There were many
Whataboutism at its best?
John Oliver: https://youtu.be/1ZAPwfrtAFY?t=6m2s
On Fri, Nov 17, 2017 at 7:10 PM, Andy Townsend wrote:
> On 17/11/2017 22:52, Clifford Snow wrote:
> I think we are all thankful for the newsletter. And believe they are free
One important aspect was missing in the announcement. The tool's new name
is a tiny part of a much bigger set of community suggested and requested
changes. Fully ignoring functionality changes that many community members
suggested is biased.
Mechanical edit claim was also never justified --
edia as I know
> there's a whole world of people there doing their own thing. It doesn't
> have to go both ways.
> In short, I have looked at your tool and don't think it is currently
> beneficial to the OSM ecosystem. The discussions ongoing here suggest it
> won't ever be.
I have noticed that the proposed fixes were not marked with vote=1. I fixed
I'm not sure if vote=1 is needed for the multiple-choice challenges. They
were originally copied from the officially deprecated tags, so
On Mon, Nov 13, 2017 at 6:02 PM, Michael Reichert <osm...@michreichert.de>
> Hi Yuri,
> Am 13.11.2017 um 22:58 schrieb Yuri Astrakhan:
> > Andy, I can only assume you agree with the rest of my argument. As for
> > case -- this is not a mechanical
On Mon, Nov 13, 2017 at 6:19 PM, Frederik Ramm <frede...@remote.org> wrote:
> On 11/13/2017 10:58 PM, Yuri Astrakhan wrote:
> > Andy, I can only assume you agr
While it is easy to throw tons of accusations and be less civil, I will try
maintain my level of decency. I have forwarded you a snippet of one of the
emails I received (without the sender name). Also, you are welcome to
organize some independent person you trust in NYC to stop by and examine it
1 - 100 of 225 matches
Mail list logo