Minutes
Introduction
<Mike Edwards>
Link to previous meeting minutes:
Resolution: Minutes of 2009-09-01 accepted w/o
Administrivia
Some outstanding comments on Test suite the need to be addressed
Note that todays meeting will recess and continue tomorrow at 12:00PM EDT / 9:00AM PDT
Existing Issues
MikeE:
Describes new proposal
...autowire is addressed by other issue so left out of this proposal
...all changes in section 5.5.1
SimonN:
Many bullets of form "if (this) then ..." but does not say if condition is not true
...better to be explicit about alternatives
SimonN:
Interfaces may be derived form component type and may be no explicit XML - wording should make clear that it is the "logical"
interface that is being referred to
AnishK:
Concerns regarding intents from component types
MikeE:
Have attempted to refer values of requires, etc attributes to Policy spec to get correct inheritance
...agrees that wording "value of attribute" should be changed to use "value applied" to indicate value needs computing as per
Policy spec inheritance rules, etc
SimonN:
Situation where there's an implicit wired by impl that needs to be covered
...multiplicity 1..1, internal target, non-overridable
MikeE:
Think this is a new issue
MikeE:
Inheritance of callbacks not cover in existing spec - new issue ASSEMBLY-181
SimonN:
Bindings for composite references need to be stripped out when promoted
Proposal needs more work - Simon N will E-mail his use case for discussion
Test Suite Issues
MikeE:
Proposal (in JIRA) to change wording
SimonN:
ASM_8018 seems redundant
...Same test but ASM_8009 is Service or Reference - ASM_8018 is Service only
MikeE:
Need to look at more carefully - there should not be any redundant tests
DaveB:
Wording for ASM_8009 seems awkward/wrong
New Issues
DaveB:
AI from ASSEMBLY-136
DaveB:
Summarizes concerns regarding multiplicity
Martin C:
Is this issue specific enough if there's so many problems?
DaveB:
Just to capture all the previous discussions
Motion: m=Dave B s=Simon N Open issue ASSEMBLY-178
Resolution: ASSEMBLY-178 opened w/o
MikeE:
From comments during internal testing
...describes issue - no proposal the works for bindings but needs addressing
Motion: Open ASSEMBLY-177 m=Danny V s=Mike E
Resolution: Issue ASSEMBLY-177 opened w/o
MikeE:
Summarizes issue regarding Test suite layout - problems with using tools
Motion: Open issue ASSEMBLY-179 m=Mike E s=Simon N
Resolution: Issue ASSEMBLY-179 opened w/o
AOB
Issues summary: Opened: 3 Closed: 0
Meeting RECESSED at 9:01AM PDT
Opening
Resumption of session from 2009-09-08
Roll call - Additional attendees
Roll - (inc Tues session) 25/28 = 89% Quorate
Pub/sub discussion
Why is existing model not sufficient?
Martin C:
Anyone think that NO CHANGES AT ALL are needed to cover eventing or pub/sub
SimonN:
Just use existing extension points?
...e.g. implementation.channel - create "standard" extensions
<clemensutschig>
a channel is NOT a component that one implements ..
<clemensutschig>
as the current proposal suggests - it's just there
ScottV:
OSOA considered that model but value in adding specifc model
...look at SCA from message exchange POV c.f. wiring between components, etc
...interface concept in WSDL could be used
AnishK:
Don't need to change existing model - need to add/extend model
<Martin C>
the trouble is wsdl has certain assupmtions about it, and if we change the assumptions we are in effect changing wsdl
<clemensutschig>
correct ..
<clemensutschig>
a wsdl implies 2 thigns
<clemensutschig>
a semantics of a SERVICE, that is an invocation (with an action ./ op), some binding and some shape
<clemensutschig>
first and second are certainly not events .. and EVENT is a state change, that only has a shape, and that's it.
AnishK:
Problems when some additional features needed - e.g. filters - can only be handled as special cases
AnishK:
Detailed summary of OSOA considerations as to why existing model insufficient
...COULD be used but a bit like "fitting square peg in round hole" after all necessary tweaks & mods
PeterN:
Agrees and notes MxN connectivity problems if using wires to connect pub/sub
...also notes that event handling and pub/sub are not the same
<Mike Edwards1>
I'd like to squeak for attendance once Scott finishes...
ScottV:
Some metadata needed for assembler to assemble components - but thinks WSDL could be tweaked
<Peter Niblett>
I also suggested we look at the additional capabilities that we need to be able to express in the assembly model, before getting
too far into the detail of how it's expressed
AnishK:
Want to avoid developer having to itterate over array of subscribers
ScottV:
Agrees and notes that Tibco never uses multiplicity of 0..n or 1..n
ClemensU:
What is the role of an Assembler in a pub/sub model?
...obviously needed for service/reference wiring but events are not "wired"
<Peter Niblett>
Ansih's point is the one I tried to make.. the grouping of event types into an "interface" has the opposite meaning in events
than it does for services
Martin C:
Many assuptions in WSDL that don't apply to eventing and also have to cover other things like Java
<Peter Niblett>
in standard service invocation the requestor is interested in a subset of the message types supported by the server (and the
server has to support them all). In events the consumer might only be interested in a subset of the message types supported
by the producer (but the producer supports them all)
...using WSDL (or other) in a manner different from common usage is something we need to avoid
<Peter Niblett>
so this causes a problem if you equate consumer to service and producer to requestor
SimonN:
Current OSOA spec implies that there will be a channel between pub and sub
...therefor only need multiplicity 0..1 or 1..1 from pub -> channel and channel -> sub
<anish>
wsdl 2.0 is much better from a model POV compared to wsdl 1.1, but there is almost zero uptake of wsdl 2.0
ScottV:
Wants to isolate Assembler and Developer views of things
...Assembler should not have to think in terms of development model (Java, etc)
...BUT Developer should not have to think in terms of Assemblers view (WSDL, etc).
JimM:
Use a special binding as opposed to implementation?
AnishK:
More problems if using binding than if using implementations
...Also any sort of wiring requires components to be aware of each other (to some extent) - should be avoided
<Mike Edwards>
I think that this discussion is the central debate about Event Processing
<Mike Edwards>
it was the debate that took place in the original OSOA discussions
<anish>
it took a f2f to hammer things out and get agreement
<Mike Edwards>
and it is good to ensure that everyone hears them and contributes to them here in OASI
<anish>
i'm wondering if this needs a f2f. i know times are difficult
<anish>
so, perhaps maybe extended calls as done in other TCs
<Martin C>
eric its the same roll url as yesterday
<Mike Edwards>
You might like to suggest that on the mailing list Anish
<Mike Edwards>
let's gauge how restricted people are wrt travelling
Action: Provide summary of discussion and tentative agreements owner=Martin C
AOB
<Martin C>
and thanks eric for scribing past two days
<Mike Edwards>
Eric - big thanks
Schreiber diagnostics output
[Delete this section before publishing the minutes]
final validation: Title not specified, default title 'OASIS SCA-Assembly TC...' was assumed
final validation: Chair not specified, default chair was assumed
statistics: Schreiber found 129 input lines
edits: Schreiber found the following text-edit commands:
edits: Line 16: s/Andmin/Admin/
edits: Line 52: s/form/from/
command-scribe: Line 3: Eric Wells is recognized
command-scribe: Line 3: Eric Wells's nick EricW has been selected
edit-substitute: command on line 16 succeeded, changed line 14 from 'Andmin' to 'Admin'
edit-delete: Line 16 was deleted
citation-detection-scribed: Line 17: Check for possible unrecognized nick 'Note that todays meeting will recess and continue
tomorrow at 12'
citation-detection-scribed: Line 19: Check for possible unrecognized nick 'ASSEMBLY 147'
citation-detection-scribed: Line 42: Check for possible unrecognized nick 'ASSEMBLY 165'
citation-detection-scribed: Line 50: Check for possible unrecognized nick 'ASSEMBLY 178'
edit-substitute: command on line 52 succeeded, changed line 51 from 'form' to 'from'
edit-delete: Line 52 was deleted
citation-detection-scribed: Line 60: Check for possible unrecognized nick 'ASSEMBLY 177'
citation-detection-scribed: Line 65: Check for possible unrecognized nick 'ASSEMBLY 179'
citation-detection-scribed: Line 72: Check for possible unrecognized nick 'Straggler Role'
citation-detection-scribed: Line 73: Check for possible unrecognized nick 'Issues summary'
citation-detection-scribed: Line 76: Check for possible unrecognized nick 'Meeting RECESSED at 9'
command-scribe: Line 79: Eric Wells is recognized
command-scribe: Line 79: Eric Wells's nick EricW has been selected
citation-detection-scribed: Line 180: Check for possible unrecognized nick 'Adjorned 10'
command-autoroll/oasis: Line 183: Attempting to fetch roll from http://www.oasis-open.org/apps/org/workgroup/sca-assembly/event.php?event_id=16108
command-autoroll/oasis: Line 183: Successfully fetched roll from http://www.oasis-open.org/apps/org/workgroup/sca-assembly/event.php?event_id=16108
system: Transformer: SAXON 9.1.0.7
[End of Schreiber diagnostic output]