OpenStreetMap
Idea: export the entire data storage of openstreetmap osm into wikipedia on git. Use cached rendered pages checked in. all layers generated by local layers from json data.
h4ck3rm1k3님이 English로 2009년 10월 19일에 게시함.Idea: export the entire data storage of openstreetmap osm into wikipedia on git. Use cached rendered pages checked in. all layers generated by local layers from json data.
check my new video on svn-git
full video 1:5h introduction to git-svn for Subversion/SVK users and deserters http://www.archive.org/details/SvnGitVideo
토론
2009년 10월 19일 12:39에 TomH님의 의견
Are you completely insane?
2009년 10월 19일 19:45에 wallclimber21님의 의견
Maybe you should first explain what problem you're trying to solve?
2009년 10월 19일 23:26에 eAi님의 의견
I'd almost say this was auto-generated spam....
2009년 10월 20일 06:36에 h4ck3rm1k3님의 의견
No spam, listen.
let me explain myself. I am thinking about a distributed git repository for openstreetmap.
Imagine you would get all your data needed via git, and would communicate with others via commits to git.
this would be distributed to other people.
even tiles would be rendered and checked in, maybe to a different area but with the same naming scheme.
you wont need a central database, you could add the names to a central index file that is also in git. So basically you would get away from a central database and import and export.
The other tools could be made to run off of these static pages by embedding code/data in the page that would be executed/evaluated when needed.
this would work via richer clients and allow mapping to be done by a peer to peer sharing load sharing system.
mike
2009년 10월 20일 08:37에 Gnonthgol님의 의견
We have a central database and a CVS. Why use git? There are no problem with how things are done today.
And there are far to much data to use a distributed repository. You might use a DHT if you insist on distributing the data but the lookup time would be to long.
The best way to handle the huge amount of data is with a central database.
2009년 10월 20일 08:41에 h4ck3rm1k3님의 의견
I am thinking of local servers, layers of them working together to map the world.
each local team would have its own central server that is pulled from bigger and bigger servers until we get the central one.
the local servers could run a database and gui instance, and have the rendering tools.
git allows the raw data to be stored, and stores only unique ones.
if you put the small enough files of data on there, only the unique ones will be stored, nodes and ways as files for example.
2009년 10월 20일 19:44에 Gnonthgol님의 의견
But still why? What advantage does we get?