Notes from the Higgins Developers
call on Thursday, May 8. (note there was no
call on May 15 due IIW and the Data Sharing summit)
Attendees
=========
Charles
Andres
Paula Austel -
IBM
Jeff Broberg
CA
Duane Buss -
Novell
Anthony Bussani -
IBM Zurich
* Greg Byrd -
NCSU/IBM
Brian Carrol -
Serena
* Tom Doman -
Novell
Andy Hodgkinson -
Novell
*Valery
Kokhan
- Parity Ukraine
* David Kuehr-Mclaren -
IBM
* Mike McIntosh -
IBM
* Tony Nadalin -
IBM
*Dale Olds -
Novell
Ernst Plassmann -
IBM
Uppili Srinivasan -
Oracle
* Drummond
Reed -
Cordance
* Bruce Rich -
IBM
* Mary
Ruddy
- Meristic/SocialPhysics
* Markus Sabedello -
Parity
* Jim Sermersheim -
Novell
George Stanchev -
Serena
Daniel
Sanders
* Paul
Trevithick -
Parity/SocialPhysics
* Brian Walker -
Parity
* Jeesmon
Jacob -
Parity
* Carl Binding -
IBM
Tom Caroll -
Parity
* Ernst Plassmann -
IBM
* Alex ???
*Mohamad - Oracle
* Attendees
Meeting
Notes
=====================_
Proposed
Agenda
===============
1) [Paul] Higgins
Positioning (for IIW, etc.)
--------------------------------------------
- Higgins Context Data
Model (see [6])
- Higgins and
UDI (Universal Data Identifiers) (see
[4])
- Higgins and the Global
Graph and Data Portability
- Higgins and our
industry's vision of an "Identity Bus"
"Identity Metasystem",
etc.
-
Higgins and
-
[Mary] Paul is going to
start by talking about some new positioning.
[Paul] There are a
number of interrelated issues here. In the last couple of weeks, I have
changed the wiki so the part that was called GG is now just called
context data model.
[Paul] Higgins is used in a
lot of different ways. Some people are interested in a way to enable a
global graph, with global addressability. There are other people who aren’t
interested in that. So we backed off from using the GG name for those for whom
that is not relevant. Like entities, entities have attributes and they live in
context data sources. Now there are a bunch of us at Parity,
Drummond, Markus and I who did a lot of work to enable universal
identifiers.
[Paul] We have written up
the specs for these optional identifiers called the
UDI which Higgins developers can use if
they want to correlate identities across context. If you click on the
links below, you will see an early version of the spec. It is a first cut
on that and we welcome feedback.
[Jim] Haven’t read
that. Are you proposing that if I've opened a context and get
entities id. Would this be in addition to that?
[Paul] Yes, in item number
4, this is one of the last few issues holding this up.
[Paul] The attempt here is to allow both.
Any identifier that fits in a string. However if you want cross context
addressability you can use the UDI spec. You won’t find anything
new in there. The reason for writing this up as a spec is to make this
easier to understand Higgins. For people new to Higgins - who just want to
look at the spec. So we can come bearing gifts to the other
ecosystems. Some of the DP guys have a need for this. So it
builds on existing technology - XRI and OpenID's and XRDS are available for
storing documents. Also along the way the Semantic web guys have been working on
URI's (cool
URI's ) and they have two competing
proposals. The UDI spec says we will support those as
well. So with Higgins you can support new green field Semantic Web way and
existing stuff.
[Paul] Another issue is the ebb and flow of the
identity bus meme. From a positioning it is an industry wide vision.
[Dale] There is nothing new
that we haven’t been working on for a long while.
[Paul] With Higgins we have
done a lot with some protocols and less with Liberty. Liberty believes they will be part of this.
Also there was the Microsoft announcement about RSS feed. So Higgins is
contributing to that industry wide trend.
[Paul] Mary and I and
others are working to rebuild the Higgins website. So we are going to build a
web page that is very open and doesn’t use abstractions like a framework or bus,
but starts at the beginning. We are
often perceived as open source information cards. This is not bad - it is moving into more
neutral territory.
[Paul] Any objections or
violent disagreements?
(Silence)
[Jim] I have a
question. Did people get the link I
sent for the next topic?
2) [Jim] IdAS Access
Management Call: See [5] Also, for #2, there is the wiki page at http://wiki.eclipse.org/Access_Control_in_IdAS
[Tony] Why are we adding
UDI? Not sure how it came about
or why
[Paul] It is shipping in Higgins 1.0. It just has a new name. It is the
way that Higgins implements the global graph, softened so that that isn't the
only way it is implemented.
[Tony] This is part of the
problem. It changes too frequently and too much. We need to be consistent and
there are too many different messages.
[Paul] You are right. But while the positioning is moving
around, the code is not moving around..
[Tony] People are viewing
it as moving around. This is very
bad. People are getting different messages. (I.e. introducing and getting rid of
Global Graph.) We need to be very
careful.
[Paul] I can appreciate
that. I’m trying to fix problem of
Higgins getting mis-perceived.
[Tony] Don’t we need to
come together as a group.
[Pal] – We’ve tried to
engage people. Maybe we can do a better job of having people interested in
Higgins position if we show up and do stuff and move
quickly.
[Tony] Higgins’ position
comes from participation in the group.
Any can participate. The members can do whatever they want.
We hope we all come to the same position.
We haven't and therefore there are problems there.
[Paul] The site is part of
the problem. It is confusing. We have a group of people make it easier for new
people to understand.
[Jim] The other problem
with positioning any change is that we need a way of saying for Higgins 1.0 site
vs. Higgins 1.1 site. So that that we can be free to make changes, and to
temper the need to appear stable with doing the right thing. We all agreed to be
down with meta data in one day.
[Jim] From 1.0 data model
changed. Need a snapshot of the website.
[Paul] We introduced
versioning for wiki pages. It defaults to 1.0. I definitely
agree.
[Jim] I don't mind making
changes as long as we agree that this is the right thing to work and do in a non
confusing way.
[Jim] We will probably do a
phone call at 10:00 EDT.. (8 mountain) I'll send out a
number for that.
[Jim] There are more
possibilities that we need to flesh out. Also we need some way of avoiding the
scenario that someone grants Bob permission and someone deletes Bob and re-adds
a new Bob and the new Bob inherits old Bobs' permission. We need to
address a number of issues here.
[Jim] For tomorrow I will
try to rework the wiki to bring focus and direction to the wiki. If there is
anything else people need to think
about before the phone call, let me know..
[Mike] I can join for the
first half hour.
[Paul] I’ll circulate a
pole to the list.
[David] I'd like to join as
well.
[Jim] I’ll put
out a phone number to the list after the meeting..
3) [Brian] Stable build of
1.1M1 was done on May 6. Next stable build is 1.1M2, currently scheduled for
June 13. More items are on 1.0.2 list, which we are continuing to schedule as
TBD.
[Brian] We are still
working on 1.1M1. There are some
builds still failing. Hopefully will get that rectified today. As soon as
we get the M1 build today, we will send out an email. The wiki page is now set
up for m2, targeted for June 13. Also building a list, right now only 1
bug is on it for a 1.0.2 patch release.
We will continue to keep that generally open. We don’t have a target date
for that.
4) [Paul] HOWL
update
---------------------
- I'm almost ready on a
HOWL1.1 update
- Last issue I'm working on
relates to allowing String Entities AND UDI
- Proposal that we add a
getUDI() call to IEntity (optional support)
[Paul] We already covered
this. Since we allow any string to be used as an entity id, it seems reasonable
to add something to differentiate between structure and a string. Optional to
context.
[Jim] So what does it
do?
[Paul] We need to decide if return relative or
fully qualified result. But at least you would know what you have and could pass
to IdAS registry and resolve. What about similar methods on the context
and for attribute identifiers?
[Jim] Get udi and context
and attribute?
[Paul] Absolutely. The intention is to
disambiguate.
[David] Will
UDI be the way to do attribute
references? Will there be a type for UDI so I know what it is?
[Paul] That hasn't been
resolved. David you are the person pushing for that. You need to push for
that and explain the use case... Do we want to support that in a first class
way?
[Jim] How do you separate
out the interrogation use cases from the management use
cases?
[Paul] We haven't resolved
that.
[Tony] So I'm not sure why
we need to have two methods here. That just adds to the negative
perception.
[Paul] In Higgins 1.0, you
can be completely flexible. Any string is allowed. The challenge is
that the application can’t tell the difference.
[Drummond] If you want
flexibility of strings and capability of structured identifiers, do you see any
other solution than to have two methods?
[Tony] This is adding
unneeded complexity.
[Jim] I think it is fairly
well understood that there is a need for global identifiers. There is a push
back to not use the existing entity id...
[Paul] That would reduce
most of the utility I want from Higgins.
(Talk of phone number
example. If we want to use phone numbers we need a
registry…)
[Tony] There is no globally
unique identifier without a registry. We shouldn’t be requiring another
registry….
[Drummond] We are using
URI’s. All the
UDI proposal is
URI's
[Paul] Some didn't want us
to be restricted to use phone number syntax. Today you can use a phone
number in a context….
[Tony] This is definitely a
broken model.
[Paul] My number can
uniquely identify an entity in two contexts…
(Disagreement on whether
the current model supports everything.)
[Tony] Globally unique
identifiers can’t be per context.
[Paul] We need to outline a
set of use cases and show Tony how it can support these use cases. There
is a distinction between context and use cases.
5) [Mary]
Opportunity to influence CardSpace 2.0 May
21-23. Possible Higgins Face-to-Face with CardSpace team on afternoon of
Wednesday the 21st or afternoon of Friday the 23rd. See [1] for link to wiki
page to collect suggestions. See [2] for persons who may attend. Microsoft did
not agree to run this as a Higgins meeting. One alternative is for individuals
to attend an open meeting. There would be no restrictions on attendance at this
half day session.
[Mary] So the meeting with
the CardSpace engineers can’t be under Eclipse rules.
[Mike] I’m still working on getting approval to
attend. If I go, I would also be
available for other meetings that we have that week.
[Tony] Just to be clear –
Higgins should not be participating unless all have agreed to Eclipse rules.
Therefore only individuals can participate.
[Mary] So now that meeting
is a separate, non-Higgins meeting and we will not discuss it further
here.
6) [Mary] We haven't gotten
much response [3] to having a Higgins F2F before the meetings in
Redmond. As Tony has pointed out, this
would make for a long week, especially after IIW the previous
week.
[Mary] So who can come and
to a general Higgins meeting before the CardSpace meeting?
[Mary] We have a few names
on Doodle.
[Mary] So the take away is
there will be a small Higgins F2F for a few people before the Microsoft SRD
meeting. Drummond and a few others will be there on Tuesday and
Wednesday.
[Mary] There will be no
call next week due to IIW. We will
resume in two weeks.
-end
Links
-----
[1] http://wiki.eclipse.org/CardSpace_wish_list
[2] http://www.doodle.ch/participation.html?pollId=6r5yw3wwdv3gtdp8
- doodle for meeting with CardSpace engineers
[3] http://www.doodle.ch/89kgw35swfgf29hw
- doodle for F2F before all this
[4] http://www.parity.com/udi.html
[5] http://doodle.ch/participation.html?pollId=szvwztqnd66uvh3a
- access mgmt call
[6] http://wiki.eclipse.org/Data_Model