Gadgets Minutes 2008-1-30

From MemberWiki

Jump to: navigation, search

URL: http://www.openajax.org/member/wiki/Gadgets_Minutes_2008-1-30

Attendees

  • Stewart Nickolas, IBM
  • Jon Ferraiolo, IBM
  • Frederik De Keukelaere, IBM
  • Rama Gurram <rama.gurram(at)sap.com>

Agenda

  • Review last week's minutes and action items
  • Review Open Items located here: Gadget Open Issues

Minutes

Jon: Invited all members of the task for to attend in Interopability Call.

Stew: Cool, speaking of interop...

Stew: From last week we said we were going to implement IFRAME loading in the gadget loader implementation and host the code on openajax.org. We have completed both of those tasks. Jon: How is the hub managed in the IFRAME case?

Stew: For now, until Hub 1.1 is ready we simply use a same domain IFRAME for defining a 'gadget site' which provides a handle (globally) to the OpenAjax Hub. This approach effectively treats the OpenAjax.hub as a singleton across all frames (trusted model).

Jon: What about the case where a gadget uses the hub to communicate with subelements within the gadget and doesn't want to expose those events to external consumers (that is, intra vs. inter gadget communication).

Fred: Have you integrated Smash into the model yet?

Stew: Many gadgets models define their own internal event models, which I anticipate will still be used at the lower level. A gadget writer will make explicit decisions about higher level events to publish to the OpenAjax bus.

Stew: Again the implementation of available currently simply uses Hub 1.0 for these higher level event in a singleton model, we will integrate Hub 1.1 when available. Jon: Ok, so the current implementation available in opensource does not preclude intra?

Stew: True, we will integrate Hub 1.1 when available into the existing model.

Stew: Regarding the IFRAME implementation, we found that Toolkits such as Ext and Dojo (at least the base CSS setting) tend to assume they own the document, they change the class name associated with html and body for example.

Stew: This leads to unexpected behaviors on the different browser as well as loss of control of your pages styles. Currently, we simply sandbox the gadgets in their own IFRAME which solves the immediate problem, but has longer term disadvantages (e.g. performance)

Stew: We considered some 'tricks' merging stylesheets, but did not pursue because we felt they would introduce more problems than they would solve and introduce unneeded complexity.

Stew: OK, lets have a look at the agenda...

ED: Minutes work-in-process, need to finish

Personal tools