Interoperability Minutes 2009-06-01
From MemberWiki
URL: http://www.openajax.org/member/wiki/Interoperability_Minutes_2009-06-01
Attendees
- Jon Ferraiolo, IBM
- Howard Weingram, TIBCO
- Javier Pedemonte, IBM
- Adam Peller, IBM
Minutes
(walking through the Hub 2.0 spec chapters)
Appendix A: References
(no issues)
Chapter 10: Implementation Metadata
(no issues, unchanged since Hub 1.0 spec)
Chapter 9: Best Practices
Jon: Howard pointed out that this chapter is non-normative, and that MUSTs should be SHOULDs. I agree. We can simply make a global search/replace.
Javier/Howard: Fine.
Jon: I'll fix after the call.
Jon: I agree with changing section header from "subscriber callbacks" to just "callbacks". All of the text applies to all callback functions.
Jon: "false positives"
Howard: When do you a navigation, start unload timer, hide iframe, etc. That eliminates this as an issue.
Howard: When potential phishing, hide iframe, start timer. When innocent navigation, reconnect happens within the timer, so OK. When navigating away, page goes away before timer is done, so OK.
Javier: Do away with frame phishing alerts? But that gives frame time to mess around. Why not have MH put unload handler? At least MH knows if is being unloaded and can ignore frame phishing.
Howard: Might be difficult when used as a library.
Javier: I just don't like getting rid of frame phishing logic. I prefer unload at MH level.
Howard: Yes, that's what I'm doing now, but I don't like it.
Javier: We could do it ourselves
Howard: But someone could blow it away
Javier: Not if you use addEventListener
Javier: I'll look at it this afternoon
Howard: Then we can remove false positives section
Jon: "should not change"
Howard: I'll fix this.
Jon: Your writeup below talks about CA talking to MA via an API that sits about the Hub. Should we add "not part of this specification"? Same with polite removal section.
Howard: OK, I'll do that.
Jon: "Hub implementation security concerns"
Jon: I think it's OK to change MUSTs to SHOULDs and leave it alone
Howard: Wasn't there something earlier in the spec?
Jon: I don't think so, but I'll check and make sure to remove redundancies.
Chapter 8: Topic Names
(no issues)
Chapter 7: Containers
(no issues)
Chapter 6: Hub APIs
Howard: I fixed text per Jon's note about who owns the params to say callee owns it.
Jon: Makes sense.
Jon: Mix-in code?
Howard: I'll remove that. Not needed for spec.
Chapter 5: Managed Hub Overview
Jon: I see you added comments to the samples, but not checked into open source.
Howard: Waiting for people to review.
Jon: Looks fine. I'll check into open source.
Chapter 4/2: Unmanaged Hub and Libraries
Jon: Per last week, I updated the APIs to use JSDoc conventions.
Chapter 4: Unmanaged Hub
Howard: String object for subID?
Jon: Good point! When used via MH underneath, the object will be a String. I'll fix that.
Ready for Release Candidate?
Howard: I made one fix - hiding the tunnel iframe now.
Javier: Good thing you caught it.
Howard: I'm OK with RC, but should we have one more phone call in 2-3 days?
Javier: I want to take a look at unloading.
RESOLUTION: Special phone call at 9amUS-PT on Wed. June 3.
