Notes from the
Higgins Developers
call on Thursday, Sep 11
Attendees
=========
* Duane Buss - Novell
*
Tom Doman - Novell
* Andy
Hodgkinson - Novell
* David
Kuehr-Mclaren - IBM
*
Drummond
Reed -
Cordance
* Bruce
Rich - IBM
*
Mary
Ruddy
- Meristic/SocialPhysics
* Paul
Trevithick -
Parity/SocialPhysics
*
Brian
Walker -
Parity
* Tom
Carroll -
Parity
Time:
noon
EDT (1700
London; 1800
Vienna,
Paris,
Berlin)
Dial-in:
1-866-362-7064 /
892048#
Agenda
1. [Brian] We are now
working on 1.1M4.
- Build is
currently scheduled for September 19th.
- See http://wiki.eclipse.org/Higgins_1.1M4
- {Brian] Still
tracking for the 19th.
I need to check with Mike on how he is doing with the WS-Trust and
SOAP
updates. Wanted to give folks
time to run through regression tests.
Based on progress, will decide if we stick with the 19th or
need to push it out. There are still 57 items open on the M4 candidate list,
so I will start pinging folks on those items.
2. [Brian &
David] Internationalization
- [Brian] No
changes for now. In arm twisting
mode to pin other component owners down.
3.
[David] Automated testing
- [David] Wasn’t asking for an update. Know Mike was more about the
build. You can put tests into
that but I don’t think it proscribes it.
My team had put together automated testing for context providers. Don’t
know if anyone was looking at this from a wider framework. Didn’t know if people were looking at
things like TPTP.
- [Brian] From past conversation, the general gist was we need to get to
that point. The idea was to at least agree on the common build platform, so
that we can leverage on top of that.
So that is why we were looking at Buckminster. The Buckminster team is interested in
moving in this direction.
Certainly the Buckminster team doesn’t have this capability today, but
is interested in developing it.
We would like to leverage that.
I think there were some questions about
TPTP…
- [David] We ran into the issue as well. Needed to go to TPTP or not.
Couldn’t go independently. Would the team be interested in seeing
the automation framework? We’re done now.
- [Brian] Great. I thought
we had a central wiki page for this.
I think it would be very valuable to share
insight.
- [David] Right. Then we
could share our comments.
4.
[David] Documentation
- [David] We were looking at the IdAS
documentation. That this is the API to create a
customer context provider, and there is no real users guide. And I’m wondering what the process
is.
- [Jim] Do you think it is possible to do that with Javadoc? Or do we
need wiki pages?
- [David] I think we need both.
There are holes in the Javadoc.
But then we also need a programmer’s guide written in English. Internally, we have a document written
in word, I need to check before donating intellectual property. Would that be
useful from us?
- [Paul] We don’t have any formal process. The wiki is what we try to
update. The wiki itself can’t
support Word attachments. But the
Higgins website can do this. From
a process point of view there is nothing stopping us. So for we have wiki pages. I’m interested in what people think.
I find wiki fast and easy to
update, but that is not necessarily what is best for the
project.
- [David] From a consumer point of view, how does that work with the
wiki?
- [Paul] It is just MediaWiki with a few twists. There is nothing automated about
it.
- [Paul] That is another thing we haven’t discussed. Like Jim and others, what do people
think? We are uncharacteristic of
Eclipse projects, using the wiki so much. I sometimes feel wikis are a way to
get things started, but you reach a point where you want words documents
somewhere, and start removing the wiki when have Word
doc.
- [Jim] For
now, wiki pages would be the best way to go unless it becomes unmanageable for
technical issues. I don’t know how possible it is to do internationalization.
If we are thinking about a dozen or so wiki pages, all you need is a bugzilla
account to update the wiki. You
don’t have to be a committer. So
to me that would be good. The short coming of the Javadoc can be pointed out
and bugs written. That can be
written a little at a time. I
probably don’t have a week to just do that, but if people throw changes into a
bug report, anyone can just throw that in. Javadoc doesn’t lend itself to writing
really good doc. You can’t put
enough in the code that turns into overview and detailed steps. All you get is package overview and
doc overview pages, one page per package. To me, it seems like using the wiki to
do the more human readable parts, then beef-up the Javadoc on the
API’s, seems
like the best route to go. Can we
attach pdfs to the wiki?
- [Paul] I could be wrong, my recollection, is just image files. I don’t believe you can attach a pdf
file to the wiki, but I should double check.
- [David] I recall attaching…
- [Jim ] Some wiki allow that.
- [Paul] It maybe an IP issue.
- [Paul] You can put a web link to the website. Some of the committers have the
permissions to check in updated documentation.
- [Jim] Or they can just be checked into the project. Because it is in subversion, you can
just link to the view SVN version.
- [Paul] Right. That is an easier way to do
that.
- [??] We can post to the dev list.
It may just be quicker to just give you the Word
doc.
- [Paul] So definitely, would love to have it. Any committer can put it into one of
the projects. And put it into the wiki page.
- [Davie] The other half is if we open a bug, then it may get a fix. If we also included the updated
version of the Javadoc, that would be best.
[Jim] If someone has gone to
the trouble of putting a suggested fix in there, I would feel more obligated
to do it.
- [David] I’m done with that topic.
- [Paul] I noticed a note about a build script. That just came out of the
blue. I didn’t have time to look
at it.
5. [Mary] Next
Higgins F2F
- Doodle for
Boston as the next
location
- http://www.doodle.ch/participation.html?pollId=gxzpfiuzdg4kveha - Please
indicated your availability. Still looking for room
possibilities.
- [Mary] We’re waiting to hear back from Mike about dates when he is
available and has a room available.
- [Paul] If
IBM couldn’t
find a room on a convenient date, Parity could volunteer a room at its
lawyer’s offices.
- [Mary] Thanks for your
offer of a room. Please update the doodle with the
times you could attend the meeting if you have not already done
so.
6.
[Mary/Paul/Dale] DIDW/Interop update.
- [Mary] The
next item is the OSIS Interop.
Paul was physically in the room for more of the
time.
- [Paul] It turned into rolling presentations
and tutorials. Unlike other
interops, there was not much testing going on. The room was mostly like a
seminar.
IBM wasn’t there
so there was no testing of their RCP selector. Tom was there from Novell and I was
there. It was kind of a non event
from the point of view of testing.
Turned into more of an education and outreach. Mike Jones of Microsoft
was there. One of the take-aways
in the I4 matrix wiki was there have been a lot of new test entries. There are a lot of tests that are not
done because the test themselves are new. One of the reasons that Microsoft has
been putting a lot of energy into this is because the new TC has as part of
its charter taking as input the test matrix of OSIS I4. So Microsoft has been trying to make
sure that it has some content to put into the OASIS standards process. That is
an overview.
- [Paul] I,
with Parity, was there testing a new Parity selector based on the Adobe
AIR selector
with enhanced packaging of the installer. We made the trade off we could test
the new one, we call it Azigo, or the other. We didn’t have enough resources to
test both. Needed to make a trade
off similar to what Novell did with DigitalMe.
7. [Paul]
Entity ID not an attribute discussion.
- [Paul] Someone put this on
the agenda. That is great. Tony
this morning sent an email.
Before I respond to Tony and Raj, wanted to state the current
proposal. Usually the issue is
terminology. However, I would
love to know what Tony means.
- [Paul] Is there anyone on the call that speaks Tony English?
- …..
[Drummond} I’ve actually been composting a response during the
call. I completely agree
with you Paul. Based on my XRI experience, unless you are very, very precise,
and get your terminology spot on, these things can just spin forever. So I’m trying to get this down to a
set of precise questions about the model and the interface for interacting
with the model.
- [Paul] Great.
- [Jim] I know in the past, there has been confusion about unique in
context vs. globally unique. I never remember which side Tony came
on. I thought it was unique in
context.
- [Drummond] Yes.
- [Jim] Maybe he is thinking there may be multiple unique identifiers,
but the entityID may be something else.
- [Dr] Exactly. I’ve been
down this tree before. If we
break this down into the individual decision about the model and the
interface, we can get this down to yes and no, and it can be documented on
that page.
- [Paul] We do have wiki pages for all these things. If as you document
this Drummond, you look at the wiki pages and see any inconsistencies, you can
update the pages.
- [Jim] Is there is a section of the wiki just for open
items.
- [Paul] Context data model 1.1 page open issues. That would be great. I think we are all for clarity. If the
way we are defining it doesn’t work.
We need more precision.
- [Drummond] I’m fairly confident with more precision, then whatever the
decision is, it won’t need to be revisited. We need people saying why we did
this. Probably need a page
devoted to this issue on what would be the resolution of these
issues.
- [Paul] Actually, on the open issue page, there is a resolved issue.
That is where we were capturing
it.
- [Drummond] I will take the action item to write up this stuff, put it
on the page and sent a response to the mailing list and reference that and
work toward closing it.
- [Paul] Great. You should also look at the open
issues list. Need update item
from open to resolved.
- [Drummond] I will do
item.
- [Paul] That would be great.
- [Paul] Guess we’re done with that item,
- [Mary] Any other items?
- [Drummond] A relative schedule for travel [For the Face-to-Face
meeting] by next call.
- [Mary] I will ping Mike.
- [Drummond]If
it makes any difference there will be a VRM architecture meeting in
Boston. Was going to be in
Seattle the week of
15th of October.
- I would be in favor of it being that week. Don’t need to be in all of it. Trying to kill both birds with one
stone.
- [Mary] So update the doodle as a comment.
- [Mary] That would be great.
- [Jim] If Mike
is not able to get an IBM office, we
have an office in Cambridge. I haven’t asked, but seems
doable.
Cambridge is the
preferred location.
- [Mary] Great.
- [Paul] Anything else?
- [Paul] All
set.
- End