[Talk-us] Proposal: delete census-designated place polygons

Serge Wroclawski emacsen at gmail.com
Mon Nov 15 11:41:12 GMT 2010

On Mon, Nov 15, 2010 at 12:05 AM, Nathan Edgars II <neroute2 at gmail.com> wrote:
> On Sun, Nov 14, 2010 at 9:20 PM, Bill Ricker <bill.n1vux at gmail.com> wrote:
>> On Sun, Nov 14, 2010 at 4:38 PM, Nathan Edgars II <neroute2 at gmail.com>
>> wrote:
>>> place=suburb doesn't work for inner-city neighborhoods.
>> nor for truly rural named crossroad settlements of no legal standing, of
>> which some yet remain
> How about place=community as a catch-all for unincorporated places?

I'm becoming frustrated because I don't feel heard.

Nathan, please do not do any mass data substitutions or modifications
on the map for things you can't verify well.

The real answer, as Phil alluded to, is that CDPs conceputally map to
different things depending on the geographic area. For example, here
in Maryland, CDPs map roughly to towns, and where Phil lives they map
roughly to neighborhoods. Neither is wrong, or invalid.

Some general principles:

1) We don't need a single unified tagging schema that works everywhere
all the time. That's just not reasonable. Never mind the entire world,
the US itself is so geographically and politically distinct that
making trying to map a single system on top isn't going to work.

1.5) Just because something doesn't fit into your local area's
pattern, doesn't make it invalid.

2) To get a better map, subtraction should be a last resort. If
there's potentially useful data in there, don't remove it.

3) We are not a repository of other datasets. We are our own dataset.
I'm for preserving the links back to the other data sets in the case
of a dataset import/merge, but in the theres nothing wrong with
modifying the data so it makes our dataset better.

- Serge

More information about the Talk-us mailing list