[OSM-talk] Long tail challenges - was: Re: Bitcoin Spam
Jaakko Helleranta.com
jaakko at helleranta.com
Tue Dec 3 22:50:40 UTC 2013
Hi Fred and others interested,
I respectfully disagree that the disagree example items would be (blatant)
spam.
They are clearly all missing what I call a "primary (classifying) tag"
(shop=* or office=* most likely for all referenced). And there are some
tagging mistakes too on existing tags. Like using note for what seems to be
more a description (which is possibly a problem of the coin map
documentation?). And perhaps we should have better documentation / stated
policy on adjectives / tone of the description tag?
These hold true.
But spam? I don't think so.
As for mail (/Internet) order shops I'd think that eg
office=mail(/internet)_order_store would always be good - - for their
physical office, that is.
I would only consider it spam if it doesn't actually exist in reality or
perhaps if it is grossly misplaced (to the extent that it is clear that no
one even tried to place it right.
Overall, I think this is probably more an issue of how we should handle a
growing "low end" of a long tail in OSM contributions. There is no reason
that we could and should not see this as a positive problem /challenge.
Perhaps it could be possible to have a MapRoulette challenge for verifying
all lone nodes + only_name_no_main_tag-objects from contributors that have
only a few change sets in their edit history? (under 10-50?)
Some tools for verifying n00bs' change sets would be great in general (and
I still have no clue of coding so I'm useless for actually developing
them).
This issue also links to previously (from time to time) discussed topic of
how to curate beginners' edits, even have some sort of review system for
beginners' / other selected users' edits.
Anyways. I think that having more people contribute - including people
whose main (primary) interest is to add only their business is good (when
not done in bad faith, which doesn't seem to be the case in your examples).
We just need to figure out how to provide targeted support, quite possibly
better targeted documentation (in this case probably coin map's job) and
how to curate in a targeted manner.
Cheers,
-Jaakko
--
Sent from my Android device.
On Dec 3, 2013 5:11 PM, "Frederik Ramm" <frederik at remote.org> wrote:
> Hi,
>
> we're seeing a rising number of new ways and nodes which seem to be
> added by people who create an account for just one purpose, namely
> adding a business to the map.
>
> This could be great - if every business were to add themselves to the
> map, we'd have a nice collection of POIs.
>
> However, in the case at hand, it seems that the interest is not to
> improve OSM but instead we're just a vehicle for people to show up on
> the "coinmap", a business directory for bitcoin-accepting businesses.
>
> It seems that a name and "payment:bitcoin=yes" is sufficient for that
> site, with an optional advertising slug in the "note" tag. But for us,
> not so much. First of all because advertising has no room in OSM; second
> because many of these businesses seem to be not really "on the ground"
> (but just a mail-order place that wants to have some marker somewhere),
> third because they often don't contain even minimal information that
> would make them useful to us. I've collected these objects created in
> the past couple of days here
>
> http://www.remote.org/frederik/tmp/bitcoin.osc
>
> A few examples:
>
> <node id="2520077777" version="1" timestamp="2013-11-04T17:56:44Z"
> uid="1795331" user="The Tobacco Seed Company" changeset="18716456"
> lat="51.5442768" lon="0.7236584">
> <tag k="name" v="The Tobacco Seed Company"/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="website" v="http://www.tobaccoseed.co.uk"/>
> </node>
>
> This is blatant advertising for a web site. It doesn't even say what
> kind of shop this is supposed to be.
>
> <node id="2523904649" version="1" timestamp="2013-11-08T04:46:05Z"
> uid="1798995" user="mkondratov" changeset="18776505" lat="41.4183069"
> lon="-81.694649">
> <tag k="name" v="Noosphere Ltd, Computer Repair"/>
> <tag k="note" v="Computer Repair Services"/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="phone" v="1-216-459-1994"/>
> <tag k="website" v="http://www.noospherecomputers.com"/>
> </node>
>
> This, too, is little more than a name on our map. We don't usually
> include the field of business in the name - this should have been
> expressed through a proper "shop" tag.
>
> <node id="2526590372" version="1" timestamp="2013-11-10T15:46:58Z"
> uid="1801179" user="79s VOF" changeset="18818705" lat="52.372218"
> lon="4.8653634">
> <tag k="name" v="79s VOF"/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="website" v="https://store.79s.co"/>
> </node>
>
> Spam.
>
> <node id="2537387222" version="1" timestamp="2013-11-18T01:07:54Z"
> uid="1809524" user="webhostpl" changeset="18964238" lat="50.0727563"
> lon="19.8938861">
> <tag k="domeny" v=""/>
> <tag k="hosting" v=""/>
> <tag k="name" v="Webhost.pl"/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="strony internetowe" v=""/>
> <tag k="website" v="http://www.webhost.pl"/>
> </node>
>
> Broken tagging (quite frequent).
>
> <node id="2540057545" version="1" timestamp="2013-11-19T18:09:14Z"
> uid="1651798" user="oldenburg69" changeset="18998877" lat="36.2026532"
> lon="-115.0597195">
> <tag k="addr:city" v="Las Vegas"/>
> <tag k="addr:housenumber" v="5216"/>
> <tag k="addr:postcode" v="89156"/>
> <tag k="addr:street" v="Glendale Ave."/>
> <tag k="name" v="Hannig Fab Works"/>
> <tag k="note" v="Hannig Fab Works LLC is a custom metal
> fabrication shop specializing in creating high quality metal fixtures,
> custom fabrication and metal art to customers in the Southern Nevada and
> abroad via the internet."/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="website" v="http://www.hannigfabworks.com/"/>
> </node>
>
> Tagging is ok as far as the address node is concerned, but without a
> shop tag the rest is kind of useless, and the "note" tag is not for your
> marketing tagline.
>
> <node id="2548748273" version="1" timestamp="2013-11-24T14:38:32Z"
> uid="1817212" user="l337 PLace" changeset="19091714" lat="60.1491622"
> lon="24.6551426">
> <tag k="addr:city" v="Espoo"/>
> <tag k="addr:housename" v="1337Place"/>
> <tag k="addr:housenumber" v="4"/>
> <tag k="addr:postcode" v="02320"/>
> <tag k="addr:street" v="Espoonlahdenkatu"/>
> <tag k="name" v="1337place.com (Logistics only)"/>
> <tag k="note" v="Quality products shipping worldwide starting
> @5EUR. BeagleBone Black and much more. U can pay with Bitcoin! #BTC
> #Bitcoin"/>
> <tag k="payment:bitcoin" v="yes"/>
> <tag k="phone" v="+358466401678"/>
> <tag k="website" v="http://www.1337place.com"/>
> </node>
>
> Whatever BeagleBone Black is, this house is certeinly not called
> "1337Place"...
>
> <node id="2563617422" visible="true" version="1" changeset="19261838"
> timestamp="2013-12-03T21:42:21Z" user="EcoBox" uid="1828695"
> lat="29.4561384" lon="-98.4193203">
> <tag k="moving boxes" v="moving boxes"/>
> <tag k="name" v="EcoBox"/></node>
> </node>
>
> What shall I say. The changeset comment contained something about
> accepting bitcoin.
>
> This is all rather undesirable - people adding their business to OSM
> would be great, but advertising isn't, and we'd prefer if they actually
> used tagging that is used in OSM, rather than simply rushing in a name
> tag and a location because that's enough to get listed on some third
> party web site.
>
> I delete the ones I encounter when they're outright spam but I'm shying
> away from suggesting some kind of automated cleaning job because I'm not
> clear on what the minimum tagging should be on any node. We don't
> currently have any such rules but seeing people dumping things into our
> database that we can't use just so they're shown on coinmap seems a bit
> strange.
>
> Bye
> Frederik
>
> --
> Frederik Ramm ## eMail frederik at remote.org ## N49°00'09" E008°23'33"
>
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk/attachments/20131203/5ede8547/attachment-0001.html>
More information about the talk
mailing list