[Talk-us] OSM US Server Infrastructure

stevea steveaOSM at softworkers.com
Sun Feb 24 00:58:20 UTC 2013

>A us-only taginfo. Frequently the tag usage distribution varies 
>between regions which is why there local taginfo instances for a few 
>The taginfo sqlite databases total about 4GB for the planet so this 
>shouldn't take much space, but I'm not sure how much CPU and disk is 
>involved in creating the db.

Yes, I agree, Paul:  us-only taginfo hosting would be excellent.

Especially done the right way, perhaps post-data analysis turns into 
wiki pages showing us-only structure. (Examples of tagging for 
Interstate highways, national park boundaries...us-only things which 
have a de facto tagging scheme).  That plus a growth management 
attitude and tools that allow discussion, changes to existing 
structure (syntax, documentation...) and growth to occur organically 
are "just what we could use about now."  Those basic tenets are in 
place now (thank you community), we have an idea now (thank you Mr. 
Norman), we have an offer of hardware for the hosting (thank you Mr. 
Dees).  Project management, anybody?  (I am not volunteering, rather 
pointing out that projects are led by leadership people who perform 
activities on a usually tracked plan to experience and usually 
measure positive results).

The index usage might allow searching based on geography (e.g. a 
bounds_rect, similar searches) and/or interpret colon and semicolon 
conventions that are well-established and becoming better-established 
(i.e. certain highway networks, plus other tags that use such a 
syntax, usually hierarchical, always structured).

These data are warehoused in various places (wiki pages, tag-info 
pages, 4 GB of actual tags...) but I'm not sure how "made sense of 
with a parochial us-only attitude" they have been taken from there, 
and if so, how digestible a comestible that is for someone looking to 
so consume.  In other words, hungry OSM contributors looking for fun 
us-only mapping to do should enjoy finding methods to examine 
well-structured tag data we can call attention to (Interstates, other 
conventions that are us-local).  Then we will naturally reach 
conclusions of "oh, here's what I can do/is my level, applies well to 
my city, is something I understand, is something that needs doing, 
and looks fun:  let me jump right in."

Thought about like that and done like that, that works.  Ian and Paul 
have stepped right up.  Anyone else?

Yes, it could become a model project for other nations/communities, 
too:  very OSM.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20130223/9140673b/attachment.html>

More information about the Talk-us mailing list