2) Branch options
[Paul]: Will soon create a 1.0 branch, until then working on doc and minor fixes. It is my understanding that the
way Eclipse does versioning, the next one could be 1.1. We may decide that 1.1 is incremental improvement
over 1.0, and then we might need another
verison a 1.2 pershaps for more
major things.
[Paul]: It will be interesting in the Face-to-Face to discuss the roadmap: polishing 1.0 vs. larger changes that go
beyond just polishing.
3) Documentation
[Paul]: I’ve
checked in the architecture PPTs (and some others) into this folder:
http://dev.eclipse.org/viewsvn/index.cgi/org.eclipse.higgins/trunk/doc/org.eclipse.higgins.doc/?root=Technology_SVN . We'll
put links to some of these on the Higgins website.
There were a number of PowerPoints and I have checked them in. Most
important is the architecture PowerPoint
that is used to generate
the diagrams in the wiki.
Mary, I want to talk to
you about pulling some of these
out. Maybe with
a link from the website. We don't want to have two copies in
SVN
[Paul]: Another topic not on the list is more clean-up of the Eclipse website. I'm creating a list of folders to delete. We can work on
this so there are no extra
folders..
4) Promoting Higgins
- Eclipse site: still waiting for improvements from
Eclipse
- Perhaps we need a "marketing" site
(higgins-project.org,
higgins.eclipse.org?)
+ no sense of the project itself
+ our brand is overwhelmed by Eclipse's brand
+ apache approach is better http://maven.apache.org http://cocoon.apache.org, http://cocoon.apache.org ,
etc.
+ like what many projects do "overlaying" sourceforge
- BTW, there's a new book on CardSpace out. Written by MSFT
folks. Pretty much a CardSpace sales tool. No mention of Higgins: Understanding
Windows
CardSpace: An Introduction to the Concepts and Challenges of
Digital Identities, By Vittorio Bertocci, Garrett Serack, Caleb Baker, Copyright
2008, Addison Wesley
- http://www.dataportability.org/
--getting lots of buzz, 655 list members, corporations like Google are
"joining"; Should Higgins be raising its profile?
- IBM demo at CES with Android
[Paul]: We talked
to Eclipse about the navigation fix. No new news. Even if they
do make the change, the big question is if
it will be enougth. A number
of folks have off list provided
input. On the Higgins site, our
brand overwelmed by the Eclipse brand. Maybe we should go back to the way we did long a go with a separate marketing
site for Higgins. We need to
maintain some stuff
on the Eclipse.org/Higgins site
that is required. But have a front door that
is more accessible. I wish Eclipse organized its website like Apache does, so
visitors wouldn't get lost.
Mary you and I can talk to Eclipse about a subdomain.
[Paul]: The context of another website
overlaid over the developer site. For example many source forge sites are organized this way. We would be
interested about how others feel. Let
Mary and I know.
[Drummond]: I definitely would support
this. Higgins is growing as a brand with 1.0 coming out
it really needs to be set up so
that people can get a clearer picture.
[Paul]: I was looking at a project that has 3 different pdf
colored introductions.
[Jim]: I agree a nice launching page would help.
[Paul]: Anyone think we should not do this?
Silence.
[Paul]: There is a new CardSpace book that is much better. It is a pure marketing document. It is not just about the technology. Clearly a book Microsoft will use to evangelize. It
is very readable. There is
no mention of Higgins. There is a mention of DigitalMe. Therefore there is an oblique
reference. OSIS is also mentioned.
[Brian]:
What is the title?
[Paul]: The title is in the agenda.
[Paul]: I was on the phone last
night with Microsoft for an hour
about terminology. I will mention at the Face-to-Face how that conversation went and next steps.
[Paul]: Another thing that is now on my radar is Portability.org. They
are phenomenal marketers.
Some people I really respect like Arron publically blogged about why he is leaving
the list huge puff of
smoke), but boy it is generating buzz.
[Paul]: It seems like Higgins has a great deal to contribute to
conversations, but it also takes marketing cycles. I
would encourage folks to take a
look. Randy at Yahoo asked what is Higgins doing about it .
[Daniel?]:
It looked like the
sponsors were ignoring enterprise
stuff.
[Paul]: We should all try to get our own sense of it. List is about 23 messages a day. To me everything looks like it is missing context.
These technologies are appropriate for
shallow, global, context free use cases.
[Paul]: The next topic is CES. I understand that IBM demonstrated a prototype Higgins card selector running on Google android . We are also working with android. We can discuss at the Face-to-Face starting an andriod work area and coordinating our efforts for a unified story. Maybe others are working on this as well.
In the specific case of Android,
Google hasn't fully open sourced it
yet.
[ ??] They haven't released the SDK
yet.
[Paul]: At the Face-to-face in NY, Dave of Google
was really excited. Though he thought the
selector belonged in Android, not
Higgins. He was personally very enthusiastic. He can't imagine typing passwords on a mobile phone. There is a prize for android applications.
[David?]: That has a March deadline. That is not
impossible. We need to
look into IP.
[Dale]: One other comment on CES. Someone
built and demoed a selector selector at CES. Need to find out what was actually
demoed.
[David]: I was working on that
team. What was shown was a demo Google app: buy media for a phone and a card selector popped up ...and get customized content based on
info in card.
[Dale]: Cool.
[David]: I
think it was one selector not a selector selector.
[Paul]: Andriod is not only a mobile platform. It happens to have Java in
it.
[Dale]: Sounds similar to what mono
does.
[Paul]:
The general topics is
as Higgins moves through the
release process to look more carefully at the documentation.
[Dale]: I sent an email to Andy about the documentation
of his soluton,
DigitalMe. We will get it done.
[Paul]: One of topics for the
Face-to-Face is: is there a way to bring Digital Me and Higgns a little bit
closer...
[Dale] I try to say DigitalMe is based on the Higgins identity selector. DigitalMe is just too memorable. I'm open to
ideas and discussion. It is a
trademarked Novell name.
5) Results of Eclipse survey [Mary]
- 3% said they were using or planning to use Higgins (up from 0%
last year)
[Mary]: Every year Eclipse
commissions a survey about Eclipse usage. In the 2006
survey no one mentioned that their organizations were using Higgins or
planning to use it in the next 12 months. In the 2007 survey, even though
Higgins is not released yet, 3% said their organization was using or
planning to use Higgins in the next 12 months.
6) IPR/Release review status (1 item is checkintocvs) [Mary].
Submitted the 1.0 IP Log to Eclipse legal last week. Finishing up the committer
and contributor contact list. Starting the review presentation.
[Mary]: We still have one
item that is checkintocvs (SVN). I submitted the 1.0 IP log to Eclipse
legal last week after the call. We are finishing up the committer and
contributor contact list and I've started work on the review
presentation.
7) Planned Higgins regular F2F January 29-31 in
Provo:
http://wiki.eclipse.org/Jan_29-31_Provo_F2F_Agenda
[Mary]: If you are planning
to attend, please add you name to the attendees section of the agenda or let Jim
know so that he can order enough food.
[Paul]: Please add additional
topics to the agenda.
[Jim]: I did have one question on the agenda. Can we delay a few topics to Wednesday or Thursday so Daniel can attend? Mainly he wants
to be around for the STS and also the configuration component. Can
you keep the STS on Wednesday?
I'm hoping to leave early on Thursday.
[Jim?]: I'd like to do a presentation on our identity provider. It is on the
agenda.
[Paul]:
Great.
[Jim]: I have lots of Higgins data model questions.
[Mary/Paul]: Yes.
[Paul]: We have talked about the first half of the call being general topics and the
second half a special topic. People can leave if they
are not interested in the special topic.
[Jim]: I have 5 data model topics. The first topic...
[Paul]: In my
perspective, first off we are trying not to reinvent things. I selected owl, then
thought lets see how it works.
Fairly soon I found owl
couldn't express certain concepts. Then
thought was to layer on owl to evolve a
data model that meets our needs. I don't think of the IdAS API as being definitial.
[Jim]: I think so. API is just reactive to what we decide.
In my mind owl has moved back and forth: Decide what want, represent it in
owl, then implement.
[Drummond]:
Where is it authoritatively
specified how what you just said
works?
[Paul]:
I agree. I will take notes
and go back and try to make it clear.
[Jim]: It is important to make that clear. Someone will say
your API does something I don't expect. The owl restricts me from doing x therfore your API should similarly behave. Need to have data model on wiki and
explain in human language what the intention is.
[David]:
One of the things I hope to get out of the data model is the
ability to work with it
programatically. If the model can be expressed, it make it easier to work
with.
[Paul]: We used to use EMF
extensively and the IBM
symantic tool kit trying to bridge
those worlds. There are symatic mismatatches between EMF and owl. They are close. But it wasn't
mature enough. The jena API was very established. Personally I'm a real EMF fan. It just didn' t work 2 years ago when we tried it.
[David]: Are we still commited to owl as the expression?
[Paul]: The short answer is yes. We are learning what the limits are
gradually. I am currently
working on cleaning up Higgins
owl.
[Jim or David]:
It would be interesting to know if there was a way for an IdAS to emit its model.
[Paul]: That was another problem. In the early days of the project we used to have
dependencies. EMF is only expressed
in Java. This was a problem for
Novell.
[Jim]:
Could something be emitted
that EMF could consume?
[David]: That is exactly what I'm saying. Make it consumable by popular tools. I
will look into it.
[Jim]:
Item 2: data model
expressiveness. There is
disagreement as to what the types
are. What kind of types are
allowed to describe an attibute value. Tom's understanding is
that any URI could be there (for simple types). I thought we were restricted to 20 basic xml data types. It effects IdAS, at least the doc.
[Paul]:One question I have.
Are people chafing against that?
[Jim]:
Tom can speak to
that.
[Tom]:
There are two things involved. Supposed to be created in types, not just value types derived
from a Higgins type. As far as value types go, in
the LDAP stuff, these are
different.
[Jim]: Are there LDAP syntaxes that can't be rep resented by the 20 something types?
[Tom ]: No. It is
dooable.
[Paul]: On
that one, I'm working on an anwer, but
am not done yet.
[Jim]: Should we allow an attribute to contain values of two different data types? Concensus would say no, they are the same value type. Didn't get a lot of conversation after Daniel 's comments.
To be continued.