- From: Sanjiva Weerawarana <sanjiva@watson.ibm.com>
- Date: Sun, 16 Jun 2002 16:31:01 +0600
- To: <Jochen.Ruetschlin@DaimlerChrysler.com>, <moreau@crf.canon.fr>
- Cc: <joyce.yang@oracle.com>, <www-ws-desc@w3.org>
If I recall correctly there was pretty good consensus to remove operator overloading and we were waiting for the rationale from Joyce (now I don't recall why). Are we re-discussing the issue? Jonathan: How do we close this issue (one way or the other)? Thanks, Sanjiva. ----- Original Message ----- From: <Jochen.Ruetschlin@DaimlerChrysler.com> To: <moreau@crf.canon.fr> Cc: <joyce.yang@oracle.com>; <www-ws-desc@w3.org> Sent: Friday, June 14, 2002 7:43 PM Subject: Re: Rationale to close the operation overloading issue > > > I don't think the point is about the least common denominator, but about not > tying > > a particular Web Service to a given implementation. Allow operation > overloading > > would, IMO, just do that. > > Seen from an implementational point of view, this is correct. But if you take > operation overloading as a kind of structuring mechanism for adding more > semantical information to the description, I think this is implementation > independent. For example: I have one (logical) operation "getAddress" which > returns the address of a certain person. With an overload mechanism I can > express the fact, that the following operations are "instances" (not in an > implementational way but in a logical) of an operation providing one certain > functionality: returning an address of a person which is identified in > different ways. > > getAddress(socialNo) > getAddress(name, surname) > getAddress(login) > ... > > Having no overload mechanism results to a more or less unstructured and maybe > missleading description (depending from the authors preferences), e.g. > > <operation name="getAddressFromSocialNo" .... > <operation name="getAddressWithNameSurname" ... > <operation name="getAddressFromLoginInfo" ... > > IMO it seems easier to map from WSDL to the PL then the other way round (for > the last case see also the comment from Russel [1]). > Finally some mapping between the "ASCII-based" operation name and the real > method name of the implenentation has to be done in any way. So, why it should > not be possible to do so with additionally considering the message format (i.e. > the input parameters)? > > WSDL with overloaded ops | PL without overload ops. > ====================================================================== > getAddress(socialNo) -> getAddressFromSocialNo(socialNo) > getAddress(name, surname) -> getAddressWithNameSurname(socialNo) > getAddress(login) -> getAddressFromLoginInfo(login) > > Now it could be stated, that this again results to unstructured (and maybe > missleading) method names as shown above in the oposite direction. But (1) > there are no other possibilities in this PL (because we have no overload > mechanism) and an implementation has to be done anyway in this way. And (2) I > think we have to decide, if we want to have complicated identifiers in the > description (i.e. the WSDL document), which is published, or in the > implementation, which is usually private. > > I'm not sure if additionally considering the input parameters for this mapping > is such a big deal, apart from the fact that --- as far is I understand our > activities --- we are focusing on the description of interfaces and not how to > implement the mapping in a efficient way (which are implementation details). > > > > So what I want to say is, that overloading of operations is not (only) an > implementational aspect. It is a usefull feature and enhances the > expressiveness of an interface description like WSDL . Yes, it's true that > there are actually (not always trivial) implementation aspects which has to be > considered when implementing the mapping, but IMO these seems solvable and are > out of the focus of our activity. > > JM2p > > jr. > > [1] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/0172.html > > > Jochen Rütschlin > DaimlerChrysler · Research and Technology > Data and Process Management (RIC/ED) > P.O. Box 2360 · D-89013 Ulm (Donau) · Germany > Visitor's address: Wilhelm-Runge-Straße 11 > Phone: +49.731.505-2830 > Telefax: +49.731.505-4401 > Internet E-Mail: jochen.ruetschlin@DaimlerChrysler.com > Internet: > http://www.informatik.uni-stuttgart.de/ipvr/as/personen/ruetschlin.html > > > > > moreau@crf.canon.fr > 14.06.2002 12:11 > Bitte antworten an moreau > > > > An: Jochen Ruetschlin/FT/DCAG/DCX@WK-EMEA2 > Kopie: www-ws-desc@w3.org, joyce.yang@oracle.com > Thema: Re: Rationale to close the operation overloading issue > > I don't think the point is about the least common denominator, but about not > tying > a particular Web Service to a given implementation. Allow operation overloading > would, IMO, just do that. > > Jean-Jacques. > > Jochen.Ruetschlin@DaimlerChrysler.com wrote: > > > As stated in 2.1 of our charter > (http://www.w3.org/2002/01/ws-desc-charter#prog> ) the WSDL framework "is not > geared towards any programming language". The > > other way round this could mean, that we should not exclude useful features > > only because some --- let me be more restrective and say: "exotic" in the > sense > > of not used in a broad way --- programming languages does not allow function > > overloading. > >
Received on Sunday, 16 June 2002 23:41:13 UTC