W3C home > Mailing lists > Public > www-ws-desc@w3.org > February 2004

Re: Marking operations safe

From: Michael Champion <mc@xegesis.org>
Date: Fri, 27 Feb 2004 11:12:13 -0500
Message-Id: <B4053229-693F-11D8-891A-000A95CCC59E@xegesis.org>
To: www-ws-desc@w3.org

> The SOAP 1.2 Web method feature [1] implies both safe and idempotent 
> when
> the property value is GET.
>
> Has WSDWG come up with a way to use the web method feature in WSDL?
>
>
> I suppose there is probably a need to split "safe" and "idempotent"
> apart.  You can then assume web method feature (WMF) is GET if both 
> safe
> and idempotent (but perhaps not).

I'm concerned that WSDL not assume too much based on the WebArch 
document or the assumptions of SOAP 1.2 about the true safeness of GET. 
  There are plenty of wireless systems out there that will charge you 
real money for every GET; I can see how the TAG wording gets around 
this, and I realize that it is a property of the transport and not the 
service, but still -- there is a more subtle problem here.  Likewise, 
there are web services such as Google's that restrict the total number 
of accesses in a given time period, so repeats of a safe,idempotent 
operation really do matter to the user even if they don't affect the 
logical state of the service itself.  I don't know what you can do 
about this besides put in some weasel words to the effect that "safe" 
doesn't necessarily mean really and truly no obligation incurred.

I definitely agree with Paul that safeness and idempotency need to be 
separated.
Received on Friday, 27 February 2004 11:12:06 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:15:02 UTC