Tuscany IRC Log (April-18-2006)

[Jean-Sebastien Delfino <js...@apache.org>]

(12:01:32 PM) *jsdelfino:* hi all
(12:02:26 PM) ant_: hi
(12:02:41 PM) *jsdelfino:* it's 12:01 pt, should we continue the 
discussion we started yesterday?
(12:03:06 PM) jmarino: sure
(12:03:45 PM) *jsdelfino:* ok, I updated the wiki page with input from 
yesterday, what we think is in/out of the release, who volunteers and a 
tentative date
(12:04:21 PM) *jsdelfino:* how do you guys want to proceed? go down the 
list on the wiki? or jump to specific sections?
(12:05:22 PM) ant_: a quick question, where it has "tentative date week 
of April 24th", that really means 28th or maybe 30th right?
(12:06:18 PM) *jsdelfino:* good question, the answer yesterday was "next 
week", jim/jeremy do u have a more precise estimate?
(12:06:47 PM) jmarino: next week :) We still need to size it
(12:07:20 PM) *haleh [/n=hmahbod@c-24-6-102-194.hsd1.ca.comcast.net/] 
entered the room.*
(12:07:42 PM) ant_: another question, is there going to be much 
discussion on ""simplify the APIs and mechanisms to contribute 
extensions", proposals posted to the mailing list etc?
(12:08:00 PM) jmarino: yes I think there should be
(12:08:40 PM) *jsdelfino:* maybe we can cut this in 2 pieces, 1. define 
stable apis, 2. complete the runtime implementation, I think most people 
in the group are more interested in the apis and when they're going to 
be stable
(12:09:42 PM) jmarino: I think there is going to be some iteration so it 
may not be a clean break
(12:10:16 PM) *jsdelfino:* trying to help precise the "next week" 
estimate... if the apis are still changing on 28th, we won't be in a 
very good shape to start our test/freeze phase...
(12:10:54 PM) jmarino: so I don't think you are going to get it more 
precise from me right now until I have had a chance to be more thorough. 
Jeremy may be able to give you his side more
(12:11:25 PM) jboynes: not really - it's going to depend on how much 
discussion there is
(12:11:39 PM) haleh: is there a set of APIs that we need sooner than 
then rest? Can those be finalized first?
(12:11:42 PM) jboynes: but as jim says I think we will iterate closer
(12:12:21 PM) haleh: Jboynes, what do you mean by iterate? You mean bug 
fixes or you mean re-definition of APIs?
(12:12:51 PM) jboynes: iterative closer to what we agree on
(12:13:05 PM) jmarino: when I used that word I meant potential API changes
(12:13:24 PM) jboynes: yes - otherwise we're in a define up front waterfall
(12:13:35 PM) jmarino: yea I don't code that way :)
(12:14:24 PM) *jsdelfino:* ok, I guess it really depends on what other 
people in the group need and when... so until we go through the other 
categories, maybe we don't have enough context
(12:14:47 PM) *jsdelfino:* do u guys want to go over the other 
categories first, we decide what's in/out and then we'll have a more 
complete picture?
(12:15:00 PM) jboynes: sure
(12:15:16 PM) jboynes: data binding?
(12:16:37 PM) jboynes: I think we have bugs there e.g. from the issue 
cr22rc_away is seeing with the bigbank samples
(12:17:14 PM) jboynes: caused by a conflict between data binding define 
in wsdl and data binding defined by the application classes
(12:17:19 PM) *jsdelfino:* I think there's 2 aspects, fix the bugs we're 
seeing now, and coming up with a more architected extensible data 
binding api
(12:17:47 PM) *jsdelfino:* allowing other data bindings than sdo to be 
plugged in
(12:17:47 PM) jboynes: I think the fix for the bug should be to define 
the extension point and use it
(12:17:56 PM) *rfeng [/n=Raymond@bi01p1.co.us.ibm.com/] entered the room.*
(12:18:10 PM) jboynes: we have this bug just using SDO
(12:19:07 PM) jboynes: anyway, I think that should be [IN]
(12:19:33 PM) *jsdelfino:* ok, is this part of the same bucket that jim 
and you are working on?
(12:19:57 PM) ant_: I'm not clear...an API for plugging databindings 
than SDO, which surely means having another databinding working?
(12:20:38 PM) jboynes: ideally, although I'd be happy with a first-cut 
API just for defining bindings using SDO
(12:21:35 PM) *jsdelfino:* dan what do you need there? do u have any 
requirements in terms of databinding?
(12:21:40 PM) jboynes: we already have a bit of a hotch-potch with SDO 
specifics in our WSDL importer plus the JAXB stuff I think dkulp is 
doing with wsdl extensions
(12:22:22 PM) dkulp: I don't think we need much here from a celtix 
standpoint.
(12:22:28 PM) *jsdelfino:* reminder: we have 4 aspects here, 
registration of schema metadata, property config, serialization ,adjust 
the bindings to use that new api...
(12:22:40 PM) dkulp: Celtix already has a databinding API that I'm using 
to read/write the SDO objects.
(12:23:02 PM) *jsdelfino:* can we live with a subset of all this? given 
that rick is just bumping into (1), ie. the registration of metadata
(12:23:36 PM) jboynes: I think that's a good place to start
(12:23:38 PM) *jsdelfino:* frank has some related items in his SDO work
(12:24:02 PM) *jsdelfino:* on the wiki page: provide an SDO metadata 
configuration model so that static and dynamic models can be preregistered
(12:25:33 PM) *jsdelfino:* so are we ok with this SDO work item + 
integration of this in the SCA runtime? or do we need more?
(12:25:55 PM) *jsdelfino:* any opinions?
(12:26:18 PM) rfeng: this should be the key to get basics working
(12:26:32 PM) *fbudinsky 
[/n=fbudinsk@CPE0004e2fcf7ef-CM023459906283.cpe.net.cable.rogers.com/] 
entered the room.*
(12:26:52 PM) *jsdelfino:* fbudinsky, hi
(12:27:08 PM) fbudinsky: hi
(12:27:34 PM) *jsdelfino:* I'm proposing that for our release we include 
your work item: provide an SDO metadata configuration model so that 
static and dynamic models can be preregistered, and integrate it in the 
SCA runtime
(12:28:22 PM) fbudinsky: sounds good ... i was thinking about a schema 
that defines the model and can be included where you want
(12:28:32 PM) *jsdelfino:* and that would be it for now... since apr 
28th is approaching quickly
(12:28:57 PM) fbudinsky: the model would be generated and would have 
methods like register() that registers the meta data in the instance
(12:29:29 PM) *jsdelfino:* not sure about the tech. details, generation 
or handwritten, etc. but can we vote on the overall goal / work item? 
-1? +1? and move to the next one?
(12:30:00 PM) jboynes: I don't understand what the item is enough to 
vote on it
(12:30:01 PM) fbudinsky: sounds good +1 for me
(12:30:40 PM) fbudinsky: i think that the item is simple ... just a 
model to config what should be registered
(12:30:40 PM) *jsdelfino:* jboynes, ask your questions then
(12:30:48 PM) fbudinsky: it can evolve from there over time
(12:32:31 PM) *jsdelfino:* comments, questions, opinions anyone?
(12:33:06 PM) rfeng: so the SDO model registry is scoped by classloader?
(12:33:35 PM) fbudinsky: i think the register method will take a 
TypeHelper as scope argument
(12:33:42 PM) rfeng: ok
(12:33:57 PM) fbudinsky: maybe we can think about classloader versions 
as well ... i'm easy :-)
(12:34:35 PM) rfeng: I think it's fine for now to explicitly use 
TypeHelper instances as scopes
(12:35:00 PM) fbudinsky: i want to start simple with this ... so i like 
using TypeHelper as the scope for now
(12:35:21 PM) rfeng: later on we may need a pluggable scoping schemes to 
make the TypeHelper hierarchy flexible
(12:35:42 PM) rfeng: let's start with a fixed, simple scheme
(12:35:56 PM) haleh: OK. so, are we closed on this item. Is it in?
(12:36:34 PM) fbudinsky: i think it's something we want in SDO anyway 
... the only real question is whether SCA uses it for the May release
(12:37:05 PM) jboynes: we need something as the current mechanisms don't 
seem to be working
(12:37:31 PM) jboynes: e.g, cr22rc_away's problems
(12:38:18 PM) *jsdelfino:* so I guess the question is: are we +1 on 
solving this problem? and we have any volunteers to do it for this release?
(12:38:22 PM) fbudinsky: there are probably reasonable workarounds that 
could be used temporarily, but i agree that we need something better
(12:39:17 PM) fbudinsky: i'll provide the model ... in the next couple 
of days ...
(12:39:42 PM) jboynes: what do you mean by "model" ?
(12:39:58 PM) ant_: guys, we're spending a looong time on this one 
issue. how long do we have scheduled for this chat?
(12:40:23 PM) fbudinsky: a schema that defines the config ... generated 
classes that you can load an instance with, and a method to register the 
contents of the instance
(12:40:41 PM) *jsdelfino:* I had a lock on the wiki page but it expired 
already :)
(12:41:29 PM) *jsdelfino:* I was thinking about an hour, but we've 
already spent 41 mins...
(12:42:18 PM) haleh: Did fbudinsky answer everyone's question? Can we 
move on unless there are further questions?
(12:43:05 PM) fbudinsky: move on by all means ... i'm not sure why we 
even needed to discuss this one
(12:43:52 PM) *jsdelfino:* is the metadata registration mechanism 
proposed by Frank [in] or [out]?
(12:44:02 PM) *jsdelfino:* or we move on without consensus?
(12:44:09 PM) *jsdelfino:* I'm +1 for it
(12:44:33 PM) ant_: +1
(12:45:05 PM) jmarino: I'm not sure I understand it enough but I don't 
think we need to discuss right now
(12:45:41 PM) rfeng: +1
(12:45:46 PM) jboynes: I still don't understand it either but so long as 
we agree to solve the problem let's move on
(12:46:02 PM) fbudinsky: i'll post the design to the mailing list when i 
have a first pass ... and people can comment on it and decide to use it 
or not then .... ok?
(12:46:20 PM) ant_: sounds good. whats next?
(12:46:31 PM) *jsdelfino:* ok, so conclusion is: +1 to resolve the 
problem, no consensus on how to resolve it, if u don't agree with the 
summary please speak up, otherwise this is what I put on the wiki
(12:46:44 PM) *jsdelfino:* next:
(12:46:51 PM) *jsdelfino:* policy extensions?
(12:47:29 PM) jboynes: given everything else I think this needs to be out
(12:47:41 PM) *jsdelfino:* I agree, +1 to move it out
(12:47:49 PM) ant_: +1
(12:48:05 PM) jmarino: +1 out
(12:48:19 PM) *jsdelfino:* good, looks like we're making progress, 
easier to move out than agree of the content of what's in :)
(12:48:43 PM) *jsdelfino:* I've marked wiring extensions and host 
integration api as [in], do u guys agree?
(12:49:16 PM) jboynes: we could move wiring extension out
(12:49:25 PM) jboynes: the main user would be policy and that's out
(12:49:37 PM) *jsdelfino:* async depends on that as well, do we want to 
move async out?
(12:50:24 PM) *jsdelfino:* I'm -1 on moving async out
(12:50:44 PM) jmarino: how bout if we just get a first cut in for async 
but not expose it?
(12:51:20 PM) *jsdelfino:* for me that means it's in minus docs, the 
mechanism needs to be in anyway
(12:51:30 PM) *jsdelfino:* I can mark it as [in] without docs
(12:52:17 PM) jmarino: actually I don't think it means -docs..
(12:52:28 PM) *jsdelfino:* what does it mean then?
(12:52:43 PM) jmarino: I think it is a first cut of an API that may 
change (kind of like alpha) so don't rely on it being the same in the future
(12:52:44 PM) dkulp: That's going to have an impact on all the bindings, 
correct?
(12:52:47 PM) jmarino: does that work for you?
(12:53:30 PM) *jsdelfino:* I think this your statement my hold true for 
other apis as well, but I'm ok with that, it's [in] with a statement 
that it may change later
(12:53:44 PM) jmarino: Hopefully the bindings won't be impacted
(12:54:06 PM) dkulp: OK, then I'm fine with docs out
(12:54:07 PM) *jsdelfino:* sorry can't spell... I think your statement 
may be true for other apis as well...
(12:54:19 PM) *jsdelfino:* and I think the bindings won't be impacted
(12:54:24 PM) jmarino: yea but hopefully not as true ;-)
(12:54:54 PM) *jsdelfino:* so what about this: wiring extensions [in], 
experimental/alpha, no or little doc
(12:55:27 PM) *jsdelfino:* does that work?
(12:55:35 PM) jmarino: sure
(12:55:56 PM) ant_: +1
(12:56:03 PM) *jsdelfino:* ok, if there's no objection, let's move on to 
the next one: runtime configuration
(12:56:15 PM) *jsdelfino:* it's something dan I think will need right?
(12:56:46 PM) *jsdelfino:* actually not just dan, it's the ability to 
configure a runtime with a selection of features / component types / 
bindings
(12:56:59 PM) dkulp: Depends, short term the config can be "modify the 
classpath"
(12:57:00 PM) *jsdelfino:* makes sense?
(12:57:41 PM) jmarino: I believe we need this with the following caveat:
(12:57:51 PM) dkulp: That's just not a valid config scenario for the 
long term. For java-one, the classpath thing is OK.
(12:58:04 PM) jmarino: the SCA spec is likely going to change and we 
should use one of the mechanisms being proposed there for extensions
(12:58:23 PM) ant_: sounds like this ones out
(12:58:35 PM) jmarino: wait maybe I don't understand this
(12:58:58 PM) jmarino: why do we need to modify the classpath? for scdl?
(12:59:14 PM) *jsdelfino:* the solution discussed on the dev list is 
nice but would require support for composites I guess, so it may not 
make it.... so I'd be ok with a classpath based solution too, remember, 
apr 28th is approaching :)
(12:59:40 PM) *jsdelfino:* to select which bindings and containers and 
system components are available in a particular runtime
(12:59:41 PM) jmarino: to add system components we shouldn't need the 
classpath
(12:59:53 PM) jmarino: that should be configurable from a directory
(01:00:09 PM) *jsdelfino:* I think most of us agree with you jim, the 
question is can we get it done now?
(01:00:30 PM) jmarino: so my thing would be we should not allow any 
configuration of the runtime until we solve it
(01:00:46 PM) jmarino: claspath stuff frightens me
(01:01:06 PM) dkulp: Jim, how do I replace bindings.axis2 with 
bindings.celtix?
(01:01:18 PM) jmarino: you can't ;) so I think we should fix it
(01:01:30 PM) dkulp: That's what I'm saying, modifying the classpath may 
be OK for java-one.
(01:01:55 PM) jmarino: would it be that difficult to have a directory 
based thing?
(01:01:59 PM) ant_: is anyone actually volunteering to do this work item 
now?
(01:02:08 PM) jboynes: we also need to break the dependencies
(01:02:14 PM) jboynes: e.g. in tomcat you can
(01:02:18 PM) jmarino: I think it would be good to raise the priority 
and get someone on it
(01:02:20 PM) jboynes: e.g. in tomcat you can't remove axis2
(01:02:46 PM) jboynes: if we fix that then maybe classpath is ok
(01:03:03 PM) jboynes: but breaking the coupling is the bigger thing
(01:03:40 PM) *jsdelfino:* I agree, this is the host api stuff I guess, 
we have so much to do now... we have to be realistic, so if we get the 
host api right, then a classpath based solution may be ok for javaone
(01:03:56 PM) *jsdelfino:* and then we all agree we need a better 
solution, for sometime in june
(01:04:22 PM) *jsdelfino:* looks like dan, who is the main user for 
this, would be ok with that?
(01:05:01 PM) dkulp: Yea, that's fine.
(01:05:24 PM) *jsdelfino:* ok can we vote now?
(01:05:46 PM) ant_: [out]
(01:05:51 PM) *rfen1 [/n=Raymond@bi01p1.co.us.ibm.com/] entered the room.*
(01:06:01 PM) *jsdelfino:* [out] from me as well
(01:06:01 PM) *rfen1 left the room (quit: Read error: 104 (Connection 
reset by peer)).*
(01:06:11 PM) *cctrieloff left the room.*
(01:06:36 PM) dkulp: [out]
(01:07:30 PM) *jsdelfino:* jboynes, jmarino any opinion?
(01:08:01 PM) jmarino: I think classpaths are going to bite us somewhere 
but if there is too much for people to do then we have no choice
(01:08:06 PM) *rfen1 [/n=Raymond@bi01p1.co.us.ibm.com/] entered the room.*
(01:09:08 PM) *jsdelfino:* ok, so that means its out, if there's any 
strong objection please speak up, I'll update the wiki
(01:09:54 PM) *jsdelfino:* ok, time check...
(01:10:19 PM) *jsdelfino:* it's 1:09pm, can people stay until 1:30pm? or 
do we need another chat tomorrow maybe?
(01:10:19 PM) jboynes: I'm not comfortable with this esp in a non-tomcat 
environment but if no-one is going to work on it then we're not going to 
get it done
(01:10:37 PM) jmarino: I can stay
(01:10:44 PM) ant_: i can stay
(01:10:48 PM) rfen1: me too
(01:11:01 PM) haleh: me too
(01:11:26 PM) *jsdelfino:* ok so let's try to complete the next 2 sections
(01:11:51 PM) *jsdelfino:* I guess we agreed last time that java, 
ws/axis2 celtix are in
(01:12:12 PM) *jsdelfino:* a question was raised on the javascript 
container, we need to decide if it's in or not
(01:12:33 PM) jmarino: I think it should be in
(01:12:41 PM) *jsdelfino:* me too
(01:12:41 PM) jmarino: we need to show something other than Java
(01:12:44 PM) *jsdelfino:* yes
(01:12:52 PM) *jsdelfino:* +1
(01:13:13 PM) *jsdelfino:* any other votes?
(01:13:32 PM) ant_: that was a fast one :-) +1 for keeping it from me
(01:13:36 PM) *jsdelfino:* :)
(01:13:55 PM) *jsdelfino:* ok [in] will be then, again speak up if you 
disagree
(01:13:59 PM) *jsdelfino:* next
(01:14:04 PM) *jsdelfino:* jsonrpc binding
(01:14:39 PM) jmarino: are there volunteers to work on that?
(01:15:02 PM) ant_: what is the suggestion here if its out, SVN delete?
(01:15:07 PM) jmarino: the reason I ask is by default I would say put 
everything in that can get enough work coverage
(01:15:12 PM) jmarino: no not svn delete
(01:15:24 PM) jmarino: just not "enabled"
(01:15:35 PM) *jsdelfino:* my vote: [in], but we mark it as experimental 
if we don't have time to put more effort into it
(01:15:46 PM) jmarino: I think we then have to balance work coverage 
with priorities
(01:15:52 PM) *jsdelfino:* i think it's important to show that we 
integrate well with Ajax stuff
(01:16:03 PM) *jsdelfino:* and may get some attention at JavaOne
(01:16:06 PM) ant_: i'd like it to be in, but there is work to do and i 
expect I'm aboutto be landed with a lot of WS items
(01:16:12 PM) jmarino: yes we need to be careful about experimental too
(01:16:30 PM) jmarino: I think we should tentatively keep it in but 
decide later based on progress
(01:16:44 PM) haleh: +1
(01:16:48 PM) jmarino: for example, I personally would rather see the 
classpath issue fixed than JSON
(01:17:00 PM) jmarino: that just my preference though
(01:17:27 PM) *jsdelfino:* axis2 for example had some experimental stuff 
on their release, and they're way out of incubation, so I think it's 
fine to have experimental projects, it's another way to attract people, 
there are things remaining to be fixed
(01:17:52 PM) jmarino: let's not go there :)
(01:18:14 PM) jmarino: yes experimental good as long as there is base 
coverage on tests and functionality
(01:19:17 PM) *jsdelfino:* jboynes, any opinion?
(01:19:48 PM) jmarino: I think we should move on with the understanding 
that all features will be subject to this review later
(01:20:52 PM) *jsdelfino:* before we move on, I'd like to record where 
we are on this particular one, we didn't get everybody's opinion here
(01:21:14 PM) *jsdelfino:* my opinion: [in] as experimental
(01:21:22 PM) *jsdelfino:* can others give their thoughts? clearly?
(01:21:38 PM) ant_: +1 to in as experimental
(01:21:42 PM) rfen1: +1
(01:21:46 PM) haleh: On Wiki you ask the question 'what else is needed'. 
Do we know?
(01:22:48 PM) *rfeng left the room (quit: Read error: 110 (Connection 
timed out)).*
(01:22:56 PM) *jsdelfino:* I don't know, ant can you help fill in the 
todo here?
(01:22:57 PM) ant_: yes, i'll drop a note to the mailing list. come the 
end of april we can revist this and chuck it out if its no good
(01:23:30 PM) haleh: OK
(01:23:38 PM) *jsdelfino:* jim, your opinion, you can say -1 if you 
want, I'll just record that, just trying to gather clear answers from all
(01:23:58 PM) jmarino: well let me clarify...
(01:24:12 PM) jmarino: it's not that I don't think it is good. It's a 
question of completeness
(01:24:49 PM) jmarino: second, I just want to say that I think all 
features should be subject to review and we should be prepared to remove 
at that point things
(01:25:03 PM) *jsdelfino:* would a "[in] experimental and incomplete" 
make you more happy for now
(01:25:20 PM) *jsdelfino:* and I agree with your statement that we 
should be prepared to remove things later
(01:25:32 PM) *jsdelfino:* we're gonna need more chats as we get closer 
to JavaOne...
(01:26:11 PM) jmarino: I think we should move on...I would have liked to 
have seen other things worked on that I believe are higher priority
(01:26:40 PM) *jsdelfino:* so let's move on, I'll record what we've said 
here on the wiki page
(01:26:47 PM) *jsdelfino:* next section is WS binding
(01:26:56 PM) ant_: [out]
(01:27:00 PM) ant_: joking
(01:27:00 PM) *jsdelfino:* haha
(01:27:03 PM) jmarino: +1
(01:27:41 PM) jmarino: I would prefer to see SMTP ;-)
(01:27:49 PM) *jsdelfino:* first item in the list is a battery of tests 
against well known interop test suites
(01:28:10 PM) *jsdelfino:* the goal is not just to test interop, but 
test that basic functions actually work...
(01:28:34 PM) jmarino: are those able to be accessed off network?
(01:28:43 PM) *jsdelfino:* no
(01:28:54 PM) jmarino: then we need something else for build tests
(01:29:04 PM) jmarino: we can have those two but not in the build
(01:29:08 PM) jmarino: too
(01:29:20 PM) ant_: ok so firstly, is anyone other than me volunteering 
to help with any of the items in the WS binding section?
(01:30:03 PM) *jsdelfino:* I am going to volunteer to help on some of 
these, but I will not have much much time
(01:30:48 PM) dkulp: Unforunately, everything in that section is x2 
(axis and celtix)
(01:30:48 PM) jmarino: I'm maxed on core
(01:30:58 PM) *jsdelfino:* anybody else?
(01:31:36 PM) dkulp: I'll be doing some of that on the Celtix side.
(01:31:49 PM) ant_: ok so given that, how high a priority is WS-RM support?
(01:32:06 PM) jmarino: where would we get it from?
(01:32:12 PM) *jsdelfino:* I think it's low, but it's just me, what do 
others think?
(01:32:18 PM) jmarino: yea low
(01:32:22 PM) dkulp: [out]
(01:32:25 PM) rfen1: I can help on defects on WS
(01:32:30 PM) ant_: Apache Sandesha2 has axis2 integration
(01:33:01 PM) ant_: ok sounds like wsrm is out then
(01:33:12 PM) jmarino: I'd be happy to movee it out given evrything else
(01:33:22 PM) *jsdelfino:* +1, I think we need to get basic stuff 
working before we get to ws-rm
(01:33:47 PM) dkulp: besides, isn't ws-rm a "policy" thing we already 
pushed out?
(01:34:10 PM) *jsdelfino:* yes good point
(01:34:10 PM) ant_: there was the patch received for java2wsdl which i 
committed today, theres still work to do but that guy says he'll do it
(01:34:39 PM) ant_: will cr22rc_away be getting bigbank to work with that?
(01:34:45 PM) *jsdelfino:* I think that should be in, it's very 
important to be able to build an app with just java...
(01:36:00 PM) *jsdelfino:* what about I say on the wiki [tentative in 
need if cr22rc volunteers]?
(01:36:10 PM) *jsdelfino:* unless somebody else volunteers
(01:36:33 PM) ant_: so the rest is just writing a few tests right? so i 
can do that so all the rest is [in]
(01:37:05 PM) *jsdelfino:* ok... the various ports to the changing apis 
are [in] was well right?
(01:37:23 PM) *jsdelfino:* [in] as well sorry
(01:37:36 PM) ant_: yeah, but the new APIs are going to be so simple 
that shouldn't take long
(01:37:56 PM) jmarino: what APIs?
(01:38:10 PM) *jsdelfino:* the APIs you're working on jim :)
(01:38:20 PM) ant_: the new Extensibility API you're doing from the 1st 
section :-)
(01:38:29 PM) jmarino: what do you mean "so simple"?
(01:38:32 PM) *jsdelfino:* the apis for next week :)
(01:38:41 PM) ant_: This is about improving our extensibility story. We 
need to simplify the APIs and mechanisms allowing people to contribute 
container, binding, and policy extensions to Tuscany
(01:38:49 PM) ant_: (quoting from the wiki)
(01:39:13 PM) jmarino: yes but the "so simple" it shouldn't take long 
put me off. What do you mean?
(01:39:43 PM) ant_: it was a joke. sorry should have put a smiley. no 
one really expects them to be simple :-)
(01:40:08 PM) haleh: forward then?
(01:40:14 PM) jmarino: k I didn't catch the joke. I thought we were 
still on the swdl2java stuff
(01:40:17 PM) jmarino: wsdl
(01:40:32 PM) *jsdelfino:* ok, so I'm going to mark all these items 
[in], except for ws-rm, and a question on the bigbank work
(01:40:38 PM) *jsdelfino:* does that work for everybody?
(01:41:27 PM) *jsdelfino:* volunteers for this work, ant main 
contributor, myself part time
(01:41:40 PM) *jsdelfino:* and raymond volunteering as well right?
(01:41:45 PM) ant_: actually, think most of the things like Google, 
eBay, Amazon are rpc/enc which we don't support but I guess the idea is 
some real WS
(01:42:04 PM) *jsdelfino:* yes some real WS
(01:42:15 PM) rfen1: yes
(01:42:38 PM) rfen1: Do we target to support doc-lit-wrapped and doc-lit?
(01:43:10 PM) *jsdelfino:* I think we need at least doc-lit, wrapped is 
an option?
(01:43:49 PM) rfen1: some interop suites and real-world ws can be used 
for testing purpose
(01:43:58 PM) ant_: (confused) we have to do wrapped don't we?
(01:44:00 PM) rfen1: doc-lit-wrapped must, doc-lit optional?
(01:44:06 PM) ant_: yes
(01:44:18 PM) ant_: but i think we should try quite hard for doc-lit
(01:44:58 PM) *jsdelfino:* yes me too, if we support doc-lit, I'm 
actually not so sure we need doc-lit-wrapped, but this is something we 
should discuss on the dev list
(01:45:36 PM) *jsdelfino:* do u guys agree? I'll put on the wiki a 
statement recording the question
(01:45:47 PM) ant_: ok
(01:46:28 PM) *jsdelfino:* ok, we're running out of time, can we cover 
the next section real quick? do we still have enough people on the chat 
for that?
(01:46:43 PM) ant_: i still have time
(01:47:09 PM) *jsdelfino:* ok, so do we think async is in or out?
(01:47:38 PM) *jsdelfino:* opinions anyone?
(01:47:43 PM) ant_: didn't we have it in using the incomplete API from 
section 1
(01:47:56 PM) *jsdelfino:* yes, just making sure
(01:48:42 PM) *jsdelfino:* ok then, next one: subsystem level wiring, 
in/out? who? when?
(01:49:03 PM) rfen1: I'm volunteering to look into this
(01:49:16 PM) *jsdelfino:* k good
(01:49:33 PM) *jsdelfino:* do the other people think it's important to 
show wiring at this level too?
(01:49:46 PM) *jsdelfino:* any opinions from our core team?
(01:50:41 PM) ant_: if rfen1 has time, but i hope it doesn't mean 
there's no time to help with WS testing stuff
(01:51:00 PM) *jsdelfino:* my opinion: we need to show that level of 
wiring, it's a very important part of the SCA PM
(01:51:34 PM) *jsdelfino:* rfen1, do you think you can spend 80/20 or 
70/30 of your time on WS / subsystem wiring?
(01:51:57 PM) rfen1: do you feel 60/40 is acceptable?
(01:52:11 PM) jmarino: sorry got pulled away for a sec
(01:52:14 PM) *jsdelfino:* what about 80/60?
(01:52:14 PM) jmarino: biology
(01:52:15 PM) *jsdelfino:* :)
(01:52:22 PM) *isilval left the room (quit: Read error: 110 (Connection 
timed out)).*
(01:52:22 PM) jmarino: wiring for subsystem?
(01:52:27 PM) rfen1: yes
(01:52:38 PM) jmarino: the spec is going to change in this area
(01:53:05 PM) *jsdelfino:* yes, the spec is going to change in may 
areas, component, composite, all this new recursive model
(01:53:15 PM) *jsdelfino:* many areas
(01:53:26 PM) rfen1: should we build some basic support so that it can 
evolve as the spec goes
(01:53:31 PM) ant_: is there any public spec other than the one from 6 
months ago, and whens th enext one planed for?
(01:53:35 PM) jmarino: so my point is we should keep that in mind, 
particularly if we are going to do throw away work
(01:53:48 PM) rfen1: or we have to hold on?
(01:54:12 PM) jmarino: so the new extensibility api will be based on a 
recursive model
(01:54:30 PM) jmarino: I'd like us to do remote wiring in line with that 
and not necessarily the old way
(01:54:32 PM) haleh: jboynes - what do you think?
(01:55:45 PM) *jsdelfino:* jmarino, in the new recursive model, we'll 
still be able to wire across composites right?
(01:56:13 PM) jmarino: yes when we add support for it remotely. We do 
that for autowire locally today
(01:56:24 PM) *jsdelfino:* we rename subsystem to composite or 
partialComposite, but wires are expressed in a similar way right?
(01:56:50 PM) jmarino: no there is no similar concept of subsystem in 
the recursive model
(01:57:11 PM) jmarino: we'd have the same context structure we have today
(01:57:20 PM) jmarino: AtomicContext and CompositeContext
(01:57:23 PM) *jsdelfino:* yes or no? looks like we're saying the same 
thing, but you said no....
(01:58:06 PM) rfen1: is subsystem replaced by partialComposite or it's 
completely removed?
(01:58:20 PM) jmarino: a composite isn't the same thing as a subsystem
(01:58:50 PM) *jsdelfino:* what about a partialComposite?
(01:59:02 PM) jmarino: i'm note sure we are going to have it
(01:59:11 PM) jmarino: we are still discussing a few things
(01:59:47 PM) jmarino: in the runtime, we could support remote wires 
with what we have today
(01:59:59 PM) *jsdelfino:* I have a bigger question actually, I am 
assuming that for apr 28th we are going to continue to expose to app 
developers the 0.9 spec scdl, not a brand new recursive programming 
model, is that assumption correct?
(02:00:03 PM) jmarino: there would be work for tomcat, etc to create the 
wires
(02:02:21 PM) *jsdelfino:* ok... looks like we're not going to close 
this one today
(02:02:22 PM) *jsdelfino:* time check
(02:02:32 PM) *jsdelfino:* 2pm, we've been going for two hours
(02:02:52 PM) *jsdelfino:* sounds reasonable to close the chat for 
today, set up another one tomorrow maybe?
(02:03:06 PM) rfen1: +1
(02:03:10 PM) ant_: i've still time but with jboynes not around maybe 
its best to stop
(02:03:22 PM) *jsdelfino:* yes, +1
(02:03:24 PM) jmarino: yes we should stop
(02:03:39 PM) *jsdelfino:* ok I'll try to record what we've said today 
on the wiki page
(02:03:49 PM) *jsdelfino:* remember... this page is open to all for editing
(02:04:00 PM) jmarino: btw I think the buid is broke
(02:04:14 PM) *jsdelfino:* so if I forget something or mis-understood 
something then please feel free to contribute to the page
(02:04:24 PM) *jsdelfino:* will send the IRC log to the list
(02:04:30 PM) *jsdelfino:* thank you all

-- 
Jean-Sebastien