Minutes, Wed Afternoon 12 June 2002 FTF

Scribe: Jeff Mischinsky (irc stripping by Jonathan Marsh, who apologized
for the delay.)

Server types is the topic of the moment 

Resolved: issue-portType-extensibility as duplicate of general
extensibility 
Proposal url:
http://lists.w3.org/Archives/Public/www-ws-desc/2002Jun/att-0046/01-port
Types-2002-06-09.html 

Sanjiva: wants to define specific semantics of a service element 

[jeffsch] +1 to named collections of port types 

Sanjiva: mulitple services in a WSDL file 

[jjmParis] +1 

[jeffsch] ?? limit to <= 1 service EII in a WSDL file (document)? 

sanjiva: review of proposal 

jeffsch: consider scenario with say a pip on a tv 

jeffsch: so it has two tuners 

[jeffsch] From WSDL 1.1: If a service has several ports that share a
port type, but employ different bindings or addresses, the ports are
alternatives. Each port provides semantically equivalent behavior
(within the transport and message format limitations imposed by each
binding). This allows a consumer of a WSDL document to choose particular
port(s) to communicate with based on some criteria (protocol, distance,
etc.).

Three cases:
- same port type, diff bindings, one instance 
- same port type, diff bindings, two instances 
- same port type, same binding, two instances 

Spec is somewhere between ambiguous and silent on which of these options
are allowed, required, and/or prohibited. 

separate discussion into 2 parts: service type, multiple port type
instances 

sanjiva: single instance of a port type per service type in current
proposal 

Consensus: having notion of a service type is a "good thing" 

Should WSDL doc be allowed to define at most one service type? 

Sanjiva: yes - tooling issue (easier for tools to generate), and
"philosophical" -- WSDL is for defining one service 

gudge: doesn't see why a tool can't spit out multiple outputs for
whatever it is interested in 

attempt to picture diagram on board: 
  S-->P-->B-->PT and S-->ST-->PT making a triangle, i.e. two ways to
from a Service to a Porttype 

Claim is this captures essence of Sanjiva's proposal 

Note: proposal requires a service to implement service type 

Summary:
[sanjiva] 1. multiple instances of a portType within a single service
type 
[sanjiva] 2. how many service types per document 
[sanjiva] 3. semantics of multiple ports within a service 
[sanjiva] 4. service type aggregation 
[sanjiva] Proposal: accept basic proposal subject to (A) add an issue
for (1) above 
[sanjiva] (B) Close (2) in favor of having multiple service types per
doc 

[Gudge] +1
[jeffsch] +1
[sgg] + 1 
[jjm] +1
 
[sanjiva] (C) Leave (3) open as an issue 
[jeffsch] multiple service elements and/or multiple service types within
a single WSDL document (as per WSDL 1.1) 

vote: Anyone who cannot live with having a single service type within a
single WSDL document (including imports) 
- Kevin objects 
- Jeffm objects 

New Proposal: Introduce service type a la Sanjiva's proposal into our
document along with the 3 issues to be worked on in the future. 

NO OBJECTION 

steve: add service type aggregation 
[Gudge] Gudge also objects to limiting # of serviceTypes per WSDL 

gudge: should there be a "base" property when doing ST aggregation or
should a ST be a collection of port types (syntactic sugar) 

Decison: Add this as an issue so that we can consider the implications
and develop a concrete proposal 

Note: there is general agreement that this is a "good thing" but we
should be careful about adding it at 4:30 at the end of a 3 day meeting 

[sanjiva]
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/part1/part1.html 

Note: issue to be discussed is first issue (yellow) in section 2.7 
[Gudge] And for those of you watching in black and white... 
[Marsh] Issue (issue-intra-port-relationship): 

Straw poll: remove first bullet in 2.7 and close issue saying "bullet
removed" 

bullet: None of the ports communicate with each other (e.g. the output
of one port is not the input of another) 

NO OBJECTIONS - make it so 

Next - Issue (issue-toplevel-element-name-uniqueness): (last in doc) 

[dbooth] gudge, it can't be related to what JeffM said, because all 6
may exist in multiple files. 

Straw poll: Should all top-level names be unique across the target
namespace? 

yes = 4, no - 5 or 6 abstain 5 or 6 

argument in favor: simplicity 

more discussion - back and forth 

straw poll again - yes - 3 no - 7 abstain - 6 

AI: take issue up in upcoming telecon 

In particular, group wanted to understand why Eric posed the issue and
what relationship this might have to the RDF mapping work 

gudge has the action item 

Round of applause for our host *********** 

no telecon tomorrow

Received on Wednesday, 26 June 2002 16:42:09 UTC