[talk-pr] Is there a negative consequence of removing POI nodes when a building polygon has been tagged with all relevant information

[Agustin Graterole]

Found Sources and Documentation for the 2009-10 orthoimagery.


   - NOAA's base portal where you search for datasets:
   http://www.csc.noaa.gov/digitalcoast/dataregistry/#/



   - NOAA's web map app where you can download datasets (haven't tried it):
   http://www.csc.noaa.gov/dataviewer/index.html?action=advsearch&qType=in&qFld=id&datareg=1&qVal=1421#app=f8ce&8069-selectedIndex=0



   - Metadata!!:
   http://csc.noaa.gov/dataviewer/webfiles/metadata/2009_usace_pr.html#2





On Fri, Jan 17, 2014 at 12:33 AM, Jeff Haack <jeff.haack at gmail.com> wrote:

> Jose, if we host with Tilecache, I think we would need the size of the
> geotiffs (~50GB according to other thread) + maybe 10-20 GB more for the
> tiles.  The more the better I suppose, but I'm pretty sure Tilecache will
> expire and remove tiles as space is needed, or at worst you could set up a
> cron job to do it.
>
>
> On Fri, Jan 17, 2014 at 12:28 AM, Jeff Haack <jeff.haack at gmail.com> wrote:
>
>> If all the tiles are prerendered I would expect a minimum of 2TB of data.
>>  You can try it out with the GDAL library - you'll want gdal2tiles.py I
>> believe.  I rendered the country of Georgia once, black and white imagery
>> covering roughly 50,000 sq. km to zoom 18, and that was around 1.7TB in
>> tiles.  Not sure but thinking Tilecache would be better option.
>>
>>
>> On Thursday, January 16, 2014, Victor Ramirez <vramirez122000 at yahoo.com>
>> wrote:
>>
>>> Just for reference, in Amazon S3, 200 GB of storage would cost around
>>> $20 a month. How much data are we talking about?
>>>
>>> On 01/16/2014 05:33 PM, JOSE L CUEVAS wrote:
>>> > I can try to setup a test-bed infrastructure at the university.  Was
>>> thinking of MapServer with titlecache.  I have no experience with none of
>>> them but will try to get something started. Any idea of how much space we
>>> will need for the WMS data?
>>> >
>>> > On Jan 16, 2014, at 1:00 PM, Jeff Haack <jeff.haack at gmail.com> wrote:
>>> >
>>> >> Search will work the same with points and polygons.  I believe it is
>>> incorrect to have a point with identical information to a polygon, better
>>> to use a polygon if you can draw it.  It seems common though to draw a
>>> building and then put POIs as points within it when there are numerous
>>> shops in the same building.
>>> >>
>>> >> Re: Marshall's, I don't think there is a common accepted standard on
>>> this.  What I typically do is map the whole building as something if the
>>> vast majority of it is taken up by one place, as is the case with
>>> Marshall's.  But arguably it could be done with a POI too.
>>> >>
>>> >>
>>> >> On Wed, Jan 15, 2014 at 9:46 PM, Victor Ramirez <
>>> vramirez122000 at yahoo.com> wrote:
>>> >> For example, would it affect the OSM search database?
>>> >>
>>> >> If a shop happens to be in a building, should the building name be
>>> the name of the most prominent shop (e.g. mashalls old san juan)?
>>> >>
>>> >> _______________________________________________
>>> >> talk-pr mailing list
>>> >> talk-pr at openstreetmap.org
>>> >> https://lists.openstreetmap.org/listinfo/talk-pr
>>> >>
>>> >>
>>> >> _______________________________________________
>>> >> talk-pr mailing list
>>> >> talk-pr at openstreetmap.org
>>> >> https://lists.openstreetmap.org/listinfo/talk-pr
>>> >
>>> > _______________________________________________
>>> > talk-pr mailing list
>>> > talk-pr at openstreetmap.org
>>> > https://lists.openstreetmap.org/listinfo/talk-pr
>>>
>>>
>>> _______________________________________________
>>> talk-pr mailing list
>>> talk-pr at openstreetmap.org
>>> https://lists.openstreetmap.org/listinfo/talk-pr
>>>
>>
>
> _______________________________________________
> talk-pr mailing list
> talk-pr at openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk-pr
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-pr/attachments/20140117/ef55b253/attachment.html>