[OSM-talk] Xml design

[Andy Robinson]

Mike Collinson wrote:
>Sent: 10 January 2007 2:05 AM
>To: talk at openstreetmap.org
>Subject: Re: [OSM-talk] Xml design
>
>At 08:31 PM 9/01/2007, SteveC wrote:
>>* @ 08/01/07 01:41:47 AM immanuel.scholz at gmx.de wrote:
>> > > What was the original rationale for the current XML design of k=""
>> > > v=""  design?
>> >
>> > The idea was, that keys and values are choosen by user at need. They
>were
>> > thought more as loose tags, not as fixed attributes (in fact, the
>original
>> > design had just a text string where the user could enter anything).
>Think
>> > of "the flickr way". The hope was, that what works for flickr, will
>work
>> > for a map too.
>>
>>Almost...
>>
>>I sat down and tried to design an ontology and realised this was the
>>dumbest thing in the world to try and do within a few minutes.
>>
>>IMHO it's worked out rather well.
>
>
>It did/does indeed.
>
>Just a thought. Is it possible to have the best of both worlds? Keep
>the <tag k='highway' v='trunk'  /> space for folksonomy and slowly
>(grindingly slowly? :-) ) develop <highway='trunk'>-like OSM "Core
>Set" for renderers to rely on and for swapping data with other open
>source initiatives.
>

You can have a highway=trunk now simply by parsing the returned key/value
pair data. I don't see any need to be messing with the k/v of the database.
So it's perfectly logical to feed planet (or even live data) to another
format. It just needs an extra step from the data user.

Cheers

Andy

>Mike
>Manila
>
>
>
>
>_______________________________________________
>talk mailing list
>talk at openstreetmap.org
>http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk