Paul,
OK - got it. I was aware of that situation
too. Too many
choices! Hopefully some CAM templates can
help
smooth out the rough edges here. For instance
a
sub-assembly could contain rules for all three
formats
and then a context parameter would drive creation
of
the right format based on use.
Cheers, DW.
----- Original Message -----
From:
Paul Spencer
To: David RR Webber ; Peter Kacandes ;
regrep
Sent: Sunday, January 18, 2004 2:05
PM
Subject: RE: [regrep] UK Government
Registry Requirements
David,
I
would just re-iterate that addresses are used for more than delivering mail.
If you have Royal Mail there, you have one aspect of addressing in the UK.
However, this address format is totally different from that used by, for
example, local authorities in the UK. If you look at the Election Markup
Language, you will see that we use "structured" and "unstructured" addresses
to cater for just this problem. In the International version, both map to the
CIQ formats. In the UK version, one maps to addresses based on some ISO
standard, and the other to the British Standard (BS7666).
Regards
Paul
-----Original Message-----
From: David RR Webber
[mailto:]
Sent: 17 January 2004 23:02
To:
Paul Spencer; Peter Kacandes; regrep
Subject: Re: [regrep] UK
Government Registry Requirements
Paul,
I'm certain the way forward here is a blend of
the UPU work, CIQ and
then using CAM as the processing
engine.
The Royal Mail rep' to the UPU has ensured that
the UPU has the
right semantics and rules for UK postal
addresses. I agree some
of it is surprisingly gnarly. When they
changed to the post code
system back twenty+ years ago it was a great
step forward in
automation - but I was amazed to find out how
many times the
seamingly obvious system is actually broken by
local rules.
What we are looking at is taking those formal
UPU rules and
encoding them in CAM, and then using the
content reference
system to equate to the CIQ dictionary and UPU
dictionary.
Once we have that - we can also look at storing
the
individual element noun semantics in the registry.
I believe we have the technology means to do
this now -
all we need is resource and time to actually
put it in
place.
Thanks, DW
----- Original Message -----
From:
Paul Spencer
To: David RR Webber ; Peter
Kacandes ; regrep
Sent: Friday, January 16, 2004 9:32
AM
Subject: RE: [regrep] UK Government
Registry Requirements
I know that John Borras is already heavily involved in this, so I
don't intend to start unless invited by him. I think the difficulty is
mapping the British Standard onto the CIQ format. Even if you can, and I
am sure there are ways, there are local semantics to incorporate (such as
for the Unique Property Reference Number) that require a means of
localizing the semantics. Hence my earlier comment about getting you
excited (since I am sure CAM can help here).
Regards
Paul
-----Original Message-----
From: David RR Webber
[mailto:]
Sent: 15 January 2004
21:04
To: Paul Spencer; Peter Kacandes;
regrep
Subject: Re: [regrep] UK Government Registry
Requirements
Paul,
Returning to the addressing stuff - the
OASIS CIQ work and
related to that the UPU work is extremely fine-grained.
The level of detail of not just the XML
elements - but the semantics,
rules and formats are covered. There
are roughly five address formats
for each postal service - as you point out
- its not just about
delivering bits of mail. The cultural
rules on address formatting are
also very important - as is being able to
handle addresses of
say - waterpumping stations and power
substations for utilities.
So - you will find that the UPU
particularly has a great deal of
semantics on all this.
The CAM has been upgraded to cover off what
the UPU needs,
and so the ultimate goal from the enduser
experience is to
have an address object - that they input
context into -
such as locale and use - and have the XML
autoconfigure
to that.
We have been busy looking at this in
significant detail.
Thanks,
DW.
