Title: Implementation question
Yes, nav state could be used to hold long
lived session data that persists across consecutive user logins. I believe a
lot of portlets could utilize a “as last viewed” display e.g. my
local weather, with view state set directly via the UI, as I fly from hot and impossibly
humid, hurricane endangered Florida to grey but safe England.
Regards,
Andre
From: Michael Freedman [mailto:]
Sent: 06 September 2005 17:52
To:
Subject: Re: [wsrp-interfaces]
Implementation question
We [the Oracle Portal] view navigational state as
unrelated to user/session state. Navigational state is encoded in the
URLs on the encapsulating page. Currently, this is the only
attempt/context for maintaining these URLs -- thus when you navigate to another
page and return navigational state is lost. Users of course always have
an option of bookmarking the page with its current navigational context.
What is the prupose of your "hint"? Is it to allow the consumer
to better manage a navigational state cache if/when it decided to manage such
data internally vs on the URL? If so, in what circumstances would a
portlet indicate anything other then "true" as this gives the closest
semantics of bookmarked navigatinal state? Are you thinking there are
situations where the portlet uses navigational state to hold portlet session
data? If so why do they do this vs using portlet sessions as they can
transfer the management of this opaque state to the consumer as well?
-Mike-
Andre Kramer wrote:
If I may somewhat abuse this forum for a quick poll on how navigational state is handled in WSRP consumer implementations, wsrp-interfaces being
the most active / vocal one:
Do your Portals facilitate saving a WSRP Portlet’s
navigational state between end user login sessions? I.e.
if I navigate to a remote
portlet’s XYZ screen and then log
out from the consumer will I actually be able to log in later (go to the page for the WSRP Portlet) and see XYZ if the portlet has saved/encoded “show
XYZ” in its navigational
state? I know the spec supports
this but is it widely implemented / selectable?
If not, do people think this should be supported by a new boolean flag in a portlet’s metadata description:
requestingNavStatePersists=true/false?
Also, I will not be able to make the SC call
next week so, again, my apologies.
Regards,
Andre
