[Tagging] Feature Proposal - Voting - relation type=person

Pieren pieren3 at gmail.com
Tue Oct 14 19:19:49 UTC 2014

On Tue, Oct 14, 2014 at 2:58 PM, Ilya Zverev <ilya at zverev.info> wrote:
> Hi! I've noticed today there is an epic voting battle on the fields of
> wiki. Hundreds of mappers came to state their disapproval of person
> relation type. I wonder how many of them actually mapped one, or even
> seen such relations. I found some of them in 2011, and posted an entry
> in SHTOSM about such relations. After at least three years, this
> proposal is not about new relation type, and you should understand
> this.

Well, they are plenty of tags in OSM that nobody has ever heard
excepted his creator ... It's not because we don't use a tag that we
cannot give an opinion. And I cannot remember any public discussion
about such 'person' relation type, at least on the international lists
which could explain why it is coming under the spotlight only today.

> There are 1648 relations of type=person already in the database.

But if I believe some feedbacks, most of them have been created by
only three contributors which is tempering the stats.

> The proposal is basically about documenting a relation type.
> It either can be found in the wiki, or not. Nobody here has the power
> of removing all such relations from the database.

The database is the result of our consensus. Of course anyone has the
power to add any thing but others have the same equal power to remove
them. (The question is more about making a mass insert or a mass
delete). If I find personal data on my own family in OSM, I will
delete them immediatly without any permission.

> Some say, you cannot make something appear in the database just by
> passing proposals in the wiki. You should do some mapping first.

You can add new tags directly in the database. And you can write
anything in the wiki. But it does not mean that the whole community
will accept your idea. Some might, some might not, most will ignore
you. This proposal is questionning the limits of OSM, it is not really
related to some physical element on the ground and is about privacy.

> Well,
> poles have been mapping for 4 years. Is it enough? Or should they have
> waited another 4 years, so there are 10k relations of this type? It is
> a part of OpenStreetMap, whether you like it or not. It won't go away
> if you vote "no" on the proposal, and there won't be less such
> relations added. So this is not about a new type. This is about
> documenting something that has been mapped for years.

But if the "vote" (I prefer "feedback") is showing that a significant
part of the community does not like the idea, then you cannot refuse
that they will delete one of such relations when they meet one of
them. (mass deletion has to be treated differently)

> And now, this far into mapping "persons" in OSM, you should not
> prevent documenting established relation types, but help document them
> better.

In one way, you say it is already widely used and in the other way,
you admit it could be better documented. I will show you where we have
problems in the proposal and why a "vote" process is sometimes better
than simply enforcing an idea.

I copy the description here:
"The main purpose of this type=person relation is to link (bind)
different objects describing a person in the data base (such as grave,
memorial place). "

Here clearly, the graves and memorials are mentionned as examples. The
relation could be used for everything related to a person and we find
now such relations linking e.g. a grave with all streets using that
person name. First mistake : it has not been strictly limited to
graves and memorials.

"That kind of relation is meant to describe dead persons. For a living
people that kind of relations should be created only if it is
reasonable. "

Second mistake : it has not been strictly limited to dead people. It's
really encouraging contributors to use this relation for everything
related to a person.

"Let's assume that good reason for creating that relations is
"encyclopedity", that means existing of article about a person on
Wikipedia. "

Third mistake : It is not strictly reserved for "notable" people and
can be used to name all graves in a cemetery (which might be forbiden
in some countries). Privacy is never mentionned. To solve this, you
could enforce a link to wikipedia because they are already an
"encyclopedia" and check people notability
(http://en.wikipedia.org/wiki/Wikipedia:Notability_%28people%29). And
once you create a link to wikipedia (or wikidata), you don't need the
relation anymore-


More information about the Tagging mailing list