Notes
for Higgins dev call – Sep 23, 2010
Attendees
·
Mike
McIntosh - Azigo
·
Paul
Trevithick – Azigo
·
Mary
Ruddy – Meristic
·
Alex
- Azigo
·
Valery
– Azigo
Logistics
•
Time: Thursday 10:00 Eastern (16:00 UTC) U.S.
•
Skype: +9900827047990866
•
Dial-in: 1-201-793-9022 passcode 7990866#
Agenda
1.
[Mary,
Mike, John Bradley] 1.1 Open IPR list, status update, OpenID4Java
- [Paul] The W3C CEO is very
interested in browser clients. Mary, in your role as chair of IC3 you
should meet with him and talk about collaboration. Yesterday I presented
the new
deck on Higgins, Active Clients and Personal Data Stores. I can send you the link
[5].
- [Mary] I’d be happy to if
you provide additional background after the call.
- [Paul] I will add a link and
sent to the list.
- [Mike] I want to leave design
items until all of the Europeans are on the line. Mary is there other
items we should do while we are waiting?
- [Mike] I got some email
feedback last night.
- [Mary] After last week’s
call I went through all the outstanding third party software programs that
haven’t yet had IPR reviews.
- [Mike] I haven’t had the
time to go through the list. I need to have Sergey look at the list and
let me know and then get back to you. He is working on a higher priority
items. I will talk to him about this.
- [Mary] Ok. Next is OpenID4Java.
John is there any news?
- [Mike] Based on the diff file,
there is so little change involved. They should have had him send a bug
report. The changes could have been made by anyone.
2.
[Mary,
Tom] What is the plan for the Eclipse/Higgins/Mydex press release?
- [Mary] Next is the press
release.
- [Mike] Tom is typically in
transit this time of day.
- [Paul] What needs to be done?
- [Mary] We are running out of
runway.
- [Paul] I will send an email to
Tom and I this AM.
- [Paul] Tom is not on the call.
Have a call with Tom after this call.
- [Mary] OK.
3.
[Paul]
New deck on Higgins, Active Clients and Personal Data Stores [5]
- [Paul] Next item, there is a
link. We could add this link to the webpage and label it Paul’s
ramblings about Higgins, so it isn’t an official position. This was
done in preparation for the WC3. They wanted info about Higgins and active
cards. Next week there is another briefing. They feel they can bring an
international flavor to this conversation and credibility and bring the
browser vendors to the table. I would like to combine IC3 and WC3 to see
how there is opportunity to collaborate. Some may be interested in that,
I’m guessing.
- [Link has been added to the
Higgins home page.]
4.
[Paul,
Alexander, Mike] App-card design review [2,3,4]
- [Paul] Now we are on to the
links about app-card and app design.
- [Mike] Those are left over from
last week.
- [Paul] Click on one of those
app card links. It takes you to the personal page; scroll to section 10
of the persona – template .owl – is an entirely new vocabulary.
I will review it…
- [Paul] I introduced a new kind
of context: it is like a Swiss army knife. The idea is that … I
made a change to the Higgins data model and I added to the context a
pointer to a context to control attributes, so a context can point to a
context that will control it (define it in some ways.) This controls what
a control context can have in it. The main class is the template class
context. I will walk through the several contexts below that. I am cross
posting in a Skype window.
- [Paul] Or should we use IRC?
- [Mike] I know I need to start
using that again.
- [Paul] The first one is an app
data attribute. Basically a change has been made, we had an app data type
context and we changed it in a small way. Click to the section on 14 (app
data, HOWL) in that overview, now an app data context has …. now
holds it in a data instance rather than in the context. Now we can go
back to section 10, and we can see why that was a convenient change to
make. Now a template context can have an app data attribute and point to
an instance of app data. A template context is a more general purpose
thing than an app data context or a mapping context. If you had an app
card, you would have a pointer to a template context and in the template
context would have an app data value pointing to an app data
instance….
- [Mike] Can I restate that? The
way an app card developer takes advantage of this is they create a
template context and populate it with a template for the mapping context
or exactly what the context is and this will be in the PDS with global
read permissions. When this is installed, something about the install
process will read this template and instantiate it. May get pointer to
the mapping context or some instantiation of same and pointer to app data
or instance of the app data, where template values are replaced by user
values. By processing this template, it knows about all the different
contexts within the users PDS that are needed. Is that how this is
working?
- [Paul] Yes.
- [Mike] Some of the questions
are: you can’t just copy the template, because any UDI’s are
put into a template version of the UDI, not the user specific copies.
- [Paul] Are there any
UDI’s like that …
- [Paul] The app data object
itself is the common stuff, then there is the per user stuff. We
don’t have a template for the persona node inside the app data.
- [Mike] One thing missing from
last week’s updates was the _javascript_ link to the import handler
- [Paul] I didn’t add it,
and I didn’t add the list of API’s to call.
- [Mike] Do you understand why we
need it? We need some way …
- [Mike] App event handler
URL…
- [Paul] So the main _javascript_
is not listed… the app rule, you get from Kynetx is determined by
app id.
- [Valery] They have separate ...
to make it generic need to have _javascript_ injected by….
- [Paul] So app service type can
be set to two values, Higgins (undefined) and Kynetx.
- [Mike] Idea is how to document
this. We could leave it up to the developer, or we host their _javascript_
so that we can make sure they are behaving. If we want to host it, we
need to define some rules for how this works for our benefit. If we want
to let them do it, we need to do a lot of work to specify this. If at some
point we want to define this service, maybe we specify an additional
service. For. now, having separate URL is more pragmatic. Easy for us
to doc what is going on and for developers.
- [Paul] Sure.
- [Paul] Someone just needs to
propose the changes.
- [Mike] I will rewrite the
proposed changes. And sent it to the Higgins dev list as a proposal.
- [Paul] Make a contract delegate
to the app data class. Should make logical renames.
- [Mike] There will be app card
specific processing that the developer has to do. Hopefully we can do
that inside the _javascript_.
- [Valery] To make sure we have code
for Oct 11, we need to create some sample files.
- [Paul] I have a complete
template file in there, but it is not a specific one. I don’t know
where to document the app card wiki extension. It will point in section
13… document this extension on the app card wiki page.
- [Paul] The next one is
authnmaterials. There are a number of attributes need to make a UDI
dynamically resolve. Look at UDI meta data class and click on the anchor 3
link for the document with Markus’ service endpoint. Right now the
only possible value for authnmaterials is user name and password. You
also need a context type.
- [Mike] Developers need real
concrete examples including the backend template stuff.…..
- [Paul] Valery to make
changes…
- [Mike] If make them look like
UDI’s might be confusing. If make them look like URL’s….
- [Mike] If make them
URL’s, the question is what we can do now to make transition to
UDI’s easier.
[1] http://www.eclipse.org/projects/project-plan.php?projectid=technology.higgins
[2] http://wiki.eclipse.org/Persona_Data_Model_2.0#AppCard
[3] http://wiki.eclipse.org/Persona_Data_Model_2.0#app-data.owl
[4] http://wiki.eclipse.org/images/0/0b/UnderstandingAppCards.pdf
[5] http://www.slideshare.net/ptrevithick/higgins-active-clients-and-personal-data-stores-v2