[OSM-talk] Redacting 75, 000 street names contributed by user chdr (Spain)

Frederik Ramm frederik at remote.org
Wed Aug 30 12:07:07 UTC 2017


Hi,

On 30.08.2017 09:29, alan_gr wrote:
> However there is confusion about the process. Several mappers on talk-es
> think that if the name of the affected streets can be verified using the
> cadastre ahead of time, this will avoid them being deleted. Some mappers
> have already started checking streets and tagging them as "source=Cadastro
> España". 

This is not what I had in mind. What I had in mind was people suggesting
a refinement to my selection process, for example telling me that "if
the name change is simply expanding Av to Avenida then you can ignore that".

I can cater for such edits as you describe and refrain from removing
names where a source tag has been added after chdr, but I don't have a
good feeling about it. I think it would be the cleanest way to delete
all names and then have them re-added, perhaps with the help of
Maproulette or similar.

> - exactly what will happen on "redaction day", including the impact on ways
> that have been edited by other mappers AFTER chdr

The current plan is to re-set the name tag to whatever it was before
chdr has edited the name (including "empty" if it was a name added by
chdr), EXCEPT where the name has meanwhile been changed by someone else,
and also EXCEPT where chdr has made primitive edits like a name expansion.

I have noticed that chdr has sometimes also added name:xx attributes
that would have to meet a similar fate.

> - approximately when "redaction day" is likely to happen

I'd say not before 15th September, but in September.

> - what mappers should do (or not do) now.

Ideally, look at a few samples from my list and check if they do indeed
represent a name that was added by chdr, and shout if they find that my
script has identified a case where chdr didn't indeed add meaningful
information.

Mikel has already identified one issue where the script didn't properly
recognize a simple "N"->"North" expansion, and there might be more.

I have uploaded a slightly reduced file that now shows the following counts:

  16331 "Mexico"
  15108 "Brazil"
  11944 "United States of America"
   6763 "RSA"
   2802 "Spain"
   2608 "Australia"
   1922 "Argentina"
   1672 "Nigeria"
   1535 "India"
   1428 "Canada"
    952 "Malaysia"
    744 "Botswana"
    701 "Philippines"
    619 "Indonesia"
    553 "Italy"
    414 "Turkey"
    290 "Hungary"
    283 "Chile"
    245 "Kenya"
    127 "Saudi Arabia"
    107 "Paraguay"
    106 "Panama"
    100 "Morocco"

The counts have changed mainly for the US, and the per-state numbers
look like this:

   5150 "Arizona"
   5102 "Texas"
    592 "New York"
    132 "New Jersey"
    129 "New Mexico"
    105 "Illinois"
     96 "California"
     90 "Hawaii"
     81 "District of Columbia"
     79 "Colorado"
     61 "Iowa"
     49 "Rhode Island"
     46 "Michigan"
     27 "Nebraska"
     27 "Missouri"
     26 "Kansas"
     25 "Pennsylvania"
     20 "West Virginia"
     19 "Maryland"
     16 "Georgia"
     14 "Ohio"
     14 "Nevada"
     13 "Minnesota"
      9 "South Dakota"
      8 "Virginia"
      8 "Indiana"
      3 "Connecticut"
      1 "North Carolina"
      1 "Maine"

Bye
Frederik

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



More information about the talk mailing list