OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

wsrp message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: RE: [wsrp][interfaces] Portlet instance handles

Hi Mike,

The way I was looking at the handle was as an opaque generated by the
portlet and returned to it on consecutive calls by the portal.
The portal will probably have its internal handle for each portlet, but this
is outside our scope.
The idea here is that the portlet CAN generate a different handle for every
user, if it manages personalization data, or it can return the same handle
for all users.
The advantage of this over passing the user id is that the portlet know that
this is the same user that requested it before, but it doesn't know WHO this
user actually is. This gives the user an added layer of privacy (for
portlets that do not require profile or user information).


-----Original Message-----
From: Michael Freedman
Sent: 5/9/02 11:47 PM
Subject: [wsrp][interfaces] Portlet instance handles

   In a number of subcommittees I believe it has been expressed/assumed
        a) Portlet instances have unique persistent ids/handles that
will be retained both by the portal and the portlet service.
        b) A portlet instance (handle) corresponds to a given user of a
given portlet on a portal page.
        c) This handle will be generated by the portlet service.

   I have been surprised by (b) being part of the assumption and would
like to understand better why folks feel this the right way to model
things.  I am surprised because it seems to require the portal maintain
N instance references per portlet on a page where N corresponds to the
number of users (who have visited this page).  Representing a database
company I guess I should be happy that we want this data bloat ;-)
Seriously, however, I would like to understand why we think its
proper/necessary to manage this persistent id when a more lightweight
mechanism exists -- namely one persistent id that represents the portlet
instance on the page + the current user identity.  Basically, it seems
the downside of the assumed/proposed approach are large(r) data sets
being maintained in the portal that likely makes export operations,
upgrades, etc. more expensive.  In addition it requires the portal to
runtime map from an obvious key (the page's portlet instance and the
user id) to this handle.
   On the surface it seems that portlets need an id that represents a
portal's portlet page instance on a per user basis as a key to that
user's personalization data for that portlet instance.  Are there other
needs for this key?  If not why do we need to generate such an id if
there is no personalization data or if the personalization is being
managed by the portlet itself (in this case the portlet can choose to
generate its own id and managed the mapping itself)?  And even if the
portal is maintaining the personalization data why can't it choose the
key itself?

To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC