Re: CR20: proposal 4a (crisper version)

The way I see it is, the whole problem arises as the default ('anon) is 
specified in the Core, but the Core does not define what 'anon' means or 
restricts it in any way.

Another issue I have is, if default is maintained in the Core, but 
soap/http binding spec says that 'anon' value is disallowed in the case 
of a request message (in a req-res MEP), then every request-message has 
to serialize a non-anon value (otherwise it would default to 'anon'). 
This takes away the advantage that a response message has wrt defaults 
(in a req-res MEP) where it does not have to serialize the value for 
wsa:To (zero-sum game).

-Anish
--

Jonathan Marsh wrote:
> AIUI the rationale for removing a default value for wsa:To is that in 
> more than half the cases (requests, especially over HTTP), the anonymous 
> value doesn’t have much utility, as it’s generally undefined or banned.
> 
>  
> 
> But in practice, no single default value has much utility as a default 
> in request messages.  But the default value of anon in HTTP responses 
> (and probably other protocol’s responses) does clearly have utility.
> 
>  
> 
> To disallow a default in the request case seems entirely reasonable, but 
> throws out the baby with the bathwater by removing the default 
> generally, in particular from the case where it actually has utility.  
> On the other hand, the harm caused by having a default that can’t be 
> used in some cases isn’t any worse than the cure (not have a default at 
> all).
> 
>  
> 
> So, -1 to losing the default value of wsa:To.
> 
>  
> 
> ------------------------------------------------------------------------
> 
> *From:* public-ws-addressing-request@w3.org 
> [mailto:public-ws-addressing-request@w3.org] *On Behalf Of *David Hull
> *Sent:* Monday, February 13, 2006 7:05 AM
> *To:* Anish Karmarkar
> *Cc:* public-ws-addressing@w3.org
> *Subject:* Re: CR20: proposal 4a (crisper version)
> 
>  
> 
> Anish Karmarkar wrote:
> 
> /[snip]/
> Alternately, if we don't want to change the cardinality of [destination] 
> property then I would suggest making the wsa:To infoset representation 
> REQUIRED.
> 
> FWIW, that's choice 2 in "Choices for CR 20" [1]
> 
> http://lists.w3.org/Archives/Public/public-ws-addressing/2006Feb/0058.html
>

Received on Monday, 13 February 2006 20:13:22 UTC