[ckan-dev] Map-building "IDE"
Adrià Mercader
amercadero at gmail.com
Mon Apr 18 19:52:29 UTC 2011
Seb:
>> If you get a chance, I would be interested in the top-level background
>> on the pros and cons. Of course, you can tell me to be less lazy and
>> find out for myself ;)
The most usual criticism of traditional OGC standards is that they are
not suitable for the modern web: WMS vs easyly cacheable tiles, XML vs
geojson. Also their generally long and sometimes cumbersome
specifications and interfaces (I have suffered several times the OGC
specifications, but even then I prefer a cumbersome specification to
none at all)
A nice example of WMS services criticism:
http://sproke.blogspot.com/2011/02/overcome-by-events-or-rearranging-deck.html
These two posts summarize the latest iteration of the neo-geo
movement, called NoGIS (a tongue-in-cheek reference to NoSQL). And the
comments show some of the opposing views:
http://mike.teczno.com/notes/nogis-slides.html
http://blog.geoiq.com/2011/03/29/what-does-nogis-mean/
My humble opinion is that "neogeographers" focus in a very specific
aspect of the spatial world, which is publishing nice and fast maps on
the web, and they do a brilliant job at it. (good examples: TileMill,
http://protectedplanet.net/,
http://afghanistanelectiondata.org/election/2009/data/). But there's a
lot more to it, and there are also desktop GIS applications that need
to consume services, SDI in different countries to connect, different
projections to support, etc.
>> OSM geocoder and reverse-geocoder that does not have the limitations
>> of the Google one.
>And what are those limitations?
You can only geocode a certain amount of items per day, and you must
show the results on Google Maps only.
David:
> I'd be interested to hear (in v. broad terms) what the CKAN
> geo-related features we're considering implementing. I'm guessing it
> could be some/all of these:
I don't know which are the geo plans for CKAN, but all the points you
mentioned are interesting and probably relatively to implement
> * Search filtered by bounding area (hence your mention of PostGIS)
We are actually doing this in ckanext-spatial, which adds a api query
like this one:
http://.../api/2/search/package/geo?bbox=-3.024175,53.950255,-3.224605,54.129025
I think it will be integrated in data.gov.uk soon
> * Search filtered/ordered by distance from a point or a place (hence PostGIS & geocoding)
> * Data preview that plots the data points onto a map (hence Mapnik)
That will be really cool (and more tricky than the other ones). We
would need to link to an actual geo dataset instead of a service
(which can be previewed too, like the WMS preview)
More information about the ckan-dev
mailing list