RE: Civic Address for V2

For reference, my address in the UK was that didn't work in a two line
address format was:

Waterfront, 2nd Floor
Hammersmith Embankment
Chancellors Road
London 
W6 9RU
United Kingdom


I think either Andrei's proposal or Richard's proposal would work (they
are essentially the same), though I don't like the A5 mapping to
Premises; it should be to CAType 25 in RFC4775 (building).  I don't
think either proposal would satisfy Japanese addresses without the
addition of at least one more field (A4) for chou, but I'm not an
expert.  Even with that addition, both proposals achieve relatively few
fields but still allow semantics.  The trade off is separation of street
details and the semantic designation of "extra" information like Floor
and Room.  I haven't seen agreement in the group about whether that
information is necessary for all the use cases or not.

Here is another address for Google's office in Amsterdam, which is a
good (hopefully obvious) example of why semantics matter:

Claude Debussylaan 34
Vinoly Mahler 4
Toren B, 15th Floor
Amsterdam, Netherlands
1082, Netherlands 

-----Original Message-----
From: public-geolocation-request@w3.org
[mailto:public-geolocation-request@w3.org] On Behalf Of Ian Hickson
Sent: Tuesday, March 03, 2009 2:24 PM
To: Richard Barnes
Cc: Marc Linsner; Doug Turner; Alec Berntson; public-geolocation@w3.org
Subject: Re: Civic Address for V2

On Tue, 3 Mar 2009, Richard Barnes wrote:
>
> It's not really the number of fields that's important, right?  If you 
> don't care about the semantics of the fields, then you can just use 
> one fields where everything's smashed together.
> 
> The important thing about the discussion in the Japanese document is 
> that Japanese addresses require substantially different *semantics* 
> than western addresses.

How is the following mapping incorrect?




> you may as well just use a single field.

That might not be a bad idea, actually. What's the use case for having
the information in multiple fields rather than just a multiline field?


> the whole utility of semantic tagging --

What is the utility in this case?


> If you need examples, there are more (from Austria) in this document:
>
<http://tools.ietf.org/html/draft-ietf-geopriv-civic-address-recommendat
ions-02>

Address: 1234 Musterstadt, Hauptstrasse 1a - 5a Block 1b Haus 2c Stiege
1

 Andrei's proposal  
 country            
 region             
 county             
 city               Musterstadt
 premises           5a Block 1b Haus 2c Stiege 1
 street             Hauptstrasse
 street number      1a
 details            
 postalcode         1234


Address: 6173 Oberperfuss, Riedl 3097 (Pfarrkirche)

 Andrei's proposal  
 country            
 region             
 county             
 city               Oberperfuss
 premises           (Pfarrkirche)
 street             Riedl
 street number      3097
 details            
 postalcode         6173


Address: 
   Karlsplatz 1/2/9
   Wien  A-1010
   Austria


 Andrei's proposal  
 country            Austria
 region             
 county             
 city               Wien
 premises           
 street             Karlsplatz
 street number      1/2/9
 details            
 postalcode         A-1010


How are these mappings incorrect?


> There's an example in there that uses 11 different fields in a single 
> address (in its original format) and 8 in the PIDF-LO format.  The
only 
> reason it uses fewer fields in PIDF-LO is that they defined an 
> unambiguous conversion to and from that format.  Which, as I've said, 
> could be an option here.

It's always possible to have a format that's even less ambiguous than
the 
last, so we're always going to have a conversion step at some point,
even 
if we reuse an address format. The question is, what is the simplest 
format that we can use that is still practical and addresses our use 
cases? Marc said nine fields wasn't enough; why not? Alex proposed a six

field format that is already shipping internationally, but Perry said 
that it didn't work for his addresses. What exactly what the problem
with 
it? Does the nine-field option work better?


Are there use cases that a one-field answer wouldn't solve?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Tuesday, 3 March 2009 23:13:44 UTC