- From: Mark Little <mark.little@arjuna.com>
- Date: Wed, 19 Jan 2005 22:41:53 -0000
- To: "Hugo Haas" <hugo@w3.org>, "Francisco Curbera" <curbera@us.ibm.com>
- Cc: "Mark Baker" <distobj@acm.org>, "David Orchard" <dorchard@bea.com>, <public-ws-addressing@w3.org>, <public-ws-addressing-request@w3.org>
I think the pragmatic view on RefProps/RefParams has to be that they will stay (rightly or wrongly, there are implementations and specifications out there that now rely on them). I agree that the term "identifier" can be contentious. However, so can the term "state". How about just calling it/them "additional information that referencing specifications [aka using specifications] or implementations need in order to ultimately address the endpoint service"? That way we're not saying *what* goes in there, only *why*. Just my 2 cents worth. Mark. ---- Mark Little, Chief Architect, Arjuna Technologies Ltd. www.arjuna.com > Hi Hugo, > > I am glad you see this as a step towards successfully closing this issue; > reading your last main on this topic I realized this proposal was pretty > close to your view also. Two comments: > > 1. You are right that a proposal compatible with this thinking would > certainly require removing from the text everything that would seem to > imply that EPRs and/or ref. props. are identifiers (I made a proposal for > this some time ago). > > 2. I think the ref. properties vs. ref. parameters discussion is not > central to issue 1 and could be cleanly separated. If we assume that the > only identifier in an EPR is the Address field, and characterize both > properties and parameters as "state" associated to the interaction, then > the difference between the two boils down to whether a change in one of > them should be interpreted by the consumer of the EPR as implying a > possible change in the policies that affect the interaction. Note that a > change of policies is in no way associated to a change in "identity" or > anything like that, as policies can change very dynamically during the > course of the interaction (depending of the date or time of the day) w/o > any implication that the "resources" or parties involved have changed. > > If people are concerned about that issue (props vs. params) I suggest we > separate it into a different issues. > > Paco > > > > |---------+-----------------------------------> > | | Hugo Haas <hugo@w3.org> | > | | Sent by: | > | | public-ws-addressing-req| > | | uest@w3.org | > | | | > | | | > | | 01/17/2005 11:14 AM | > |---------+-----------------------------------> > >--------------------------------------------------------------------------- ----------------------------------------------------------| > | | > | To: David Orchard <dorchard@bea.com>, Mark Baker <distobj@acm.org> | > | cc: public-ws-addressing@w3.org | > | Subject: Re: Issue #1 proposed resolution | > >--------------------------------------------------------------------------- ----------------------------------------------------------| > > > > > Hi Dave and Mark. > > This is indeed very encouraging. > > * David Orchard <dorchard@bea.com> [2005-01-16 14:16-0800] > [..] > > 1. Consider a stateful interaction model where ref. properties are > > > > interaction data and the full state of the interaction (in the form of > > > > actual business data: a shopping cart, for example); this usage follows > > > > the canonical stateless implementation that http cookies allow. In fact, > > > > the semantics of ref props in this usage case are not very different > > > > from cookie semantics except for the fact that WS-Addressing does not > > > > define an EPR lifecycle. The EPR in this situation acts as a mere > > > > packaging mechanism for transmitting to the service requester the URI > > > > address + state that is characteristic of a particular interaction. In > > > > particular, there is only one identifier involved here, the URI > > > > providing the network address. The state represented by the ref. props. > > > > has no identification function in this case. The EPR is nothing but an > > > > XML envelope for the two. To reiterate, the EPR is not an identifier > > > > but a container used to specify echoed state information. > > Aren't you describing a use for reference parameters here, and not > reference properties, then? > > > 2. The use of EPRs and ref props as a server side "identifier" > > > > (as opposed to client side identifiers) is another possible and probable > > > > use case. Here, the combination of URI and ref. props. is used by the > > > > infrastructure to retrieve stateful artifacts associated with the > > > > interactions. Typically, in these cases the internal implementation of > > > > the server side provides for mechanisms to compare two EPRs for > > > > "sameness" and can operate with proper id semantics internally. > > > > Reference properties carry server side specific data (not business data > > > > as before). > [..] > > * David Orchard <dorchard@bea.com> [2005-01-16 22:17-0800] > > Yeah, we should remove that wording as it's true yet implies the client > > will make use of the EPRs for identification purposes. I have no > > problems with RefPs being used for identification, but it's by no means > > the only case. > > So, if only the server-side is concerned with what RefP's are used > for, and that the wording about the identifying role of [reference > properties] should be removed, then doesn't that it means that > [reference properties] and [reference parameters] are equally opaque > bags of XML data from the requester's perspective? > > * Mark Baker <distobj@acm.org> [2005-01-16 23:48-0500] > > On Sun, Jan 16, 2005 at 06:51:59PM -0800, David Orchard wrote: > > > EPRs aren't defined to be identifiers. An EPR may contain identifiers, > > > but it may contain state. Therefore, there's no conflict with EPRs (as > > > a class) as they stand with advice to use URIs for identifiers. > > > > This is very encouraging, and if that were the case I agree that there > > would be no issue. But for this; > > > > "A reference may contain a number of individual properties that are > > required to identify the entity or resource being conveyed." > > -- > http://www.w3.org/Submission/2004/SUBM-ws-addressing-20040810/#_Toc77464317 > > > > Your RefProp example in your initial post didn't use them as identifiers > > though; it used them as RefParams in fact. So I'll go out on a limb and > > ask; would you be content with removing RefProps? Alternately, would > > you be content with saying that RefProps shouldn't be used for > > identification? (though I guess you'd then have to distinguish them > > from RefParams, but perhaps you have some ideas about that). > > I think that your alternative only holds true if [reference > properties] are not considered by the client side for metadata > comparison purposes. If it were not the case, then [reference > properties] would be identifying this metadata about the endpoint. > > If we agree on this, then I think that we are saying [reference > properties]' use is essentially similar to [reference parameters]', > and that they are basically the same. Are we? > > Cheers, > > Hugo > > -- > Hugo Haas - W3C > mailto:hugo@w3.org - http://www.w3.org/People/Hugo/ > > > #### signature.asc has been removed from this note on January 17, 2005 by > Francisco Curbera >
Received on Wednesday, 19 January 2005 22:43:35 UTC