Re: [VCARDDAV] How to encode X-Properties in XML?

[Peter Saint-Andre <[e-mail]>]

On 3/9/10 8:39 AM, Cyrus Daboo wrote:
> Hi Simon,
> 
> --On March 8, 2010 10:10:13 AM -0500 Simon Perreault
> <[e-mail]> wrote:
> 
>> The main open issue we have for vCard XML is that of encoding
>> X-Properties in XML. vCard uses X- for unregistered extensions, while
>> vCard XML uses XML namespaces. It's not clear how to automatically
>> convert from one to another.
>>
>> A proposal: Create an XML namespace for X- properties. Put the content
>> into a <text> element.
>>
>> Example:
>>
>> X-FOO:bar
>>
>> would translate into:
>>
>> <foo
>> xmlns="urn:ietf:params:xml:ns:vcard-4.0:xname"><text>bar</text></foo>
>>
>> We don't need to add anything to the schema since we already specify that
>> an XML vCard parser MUST ignore elements in namespaces it doesn't
>> recognize. (Maybe we can add text to force it for this namespace.)
> 
> The choice we made for icalendar-in-xml was to simply allow the X-
> properties to appear as <x-...> elements in the default "icalendar"
> namespace. 

IIRC we had some discussion about this around the time of the Stockholm
meeting, starting here:

http://www.ietf.org/mail-archive/web/vcarddav/current/msg01126.html

Joe Hildebrand suggested:

X-FOO:test
VND-BAR:hi

Turning into:

<x name='FOO'><text>test</text></x>
<vnd name='BAR'><text>hi</text></vnd>

Simon said we could instead make use of XML namespaces.

<foo xmlns="urn:ietf:params:xml:ns:vcard-4.0:x"><text>test</text></x>
<bar xmlns="urn:ietf:params:xml:ns:vcard-4.0:vnd:bar"><text>hi</text></bar>

Joe replied that "pulling random elements into your own namespace is
somewhat odd from an XML standpoint."

I tend to agree. If we're going to be pulling in random elements, I'd
rather do so in a namespace other than the core one (let's have a
namespace that we know is going to be polluted from the start so that we
contain the ugliness).

> It did not seem worthwhile to have a special namespace for
> those as in reality they are processed in exactly the same way as
> regular properties. The key thing is that these elements/properties
> cannot be ignored. They are there for a purpose (be it only
> "experimental").

There might be value in doing the same thing as icalendar. Is it really
true that X- and VND data cannot be ignored? Or do we say that a
vCard-xml processor MUST support the urn:ietf:params:xml:ns:vcard-4.0:x
and urn:ietf:params:xml:ns:vcard-4.0:vnd namespaces (but perhaps MAY
treat data qualified by those namespaces differently somehow)?

Peter

-- 
Peter Saint-Andre
https://stpeter.im/