Setting the VIEWER ID when user is not logged in

[Kaan Bingol]

Hi,

Following a brief discussion with Arne (attached below), it seems that
defining a standard response for PERSON object when user is not logged-
in (like anonymous or guest users) is helpful for app. developers.

When it is important for an app. to know or differentiate anonymous
users, it would need to handle various responses from different
containers. If spec standardizes this response, cross container app.
development would be easier and there would be no need to learn/know
various behaviours of different containers.

Are there any suggestions for a standard response for anonymous users?
Arne commented that some containers return nothing while others use a
specific ID in this case.

Regards,
Kaan


---------- Forwarded message ----------
From: Arne Roomann-Kurrik <api.kur...@google.com>
Date: Nov 16, 10:11 pm
Subject: What is the VIEWER ID when user is not logged in?
To: OpenSocial Application Development


Hi Kaan,

Currently, I believe that the two types of response for this are to
return nothing at all (like orkut) or return a Person object with ID
of -1 (like MySpace). It sounds like this would be a great thing for
the spec to define. Can you post a thread
tohttp://groups.google.com/group/opensocial-and-gadgets-spec/topics
asking for standardization here?

~Arne

On Nov 15, 12:55 am, Kaan Bingol <bing...@gmail.com> wrote:

> Hi,

> Is there a common practice among different containers to set VIEWER
> for anonymous or guest users. I know some containers do not allow
> anonymous users to reach open social apps but since there are
> increasingly different containers and uses of open social gadgets, how
> a container sets VIEWER if user is not logged-in becomes important.
> Should it be null or a specific like Guest?

> A gadget may need to know this notation in order to handle anonymous
> users. Any information on how various containers handle this, is
> useful to build a common practice.

> Thanks,
> Kaan

[Brian Eaton]

I agree this would be very useful. We ran into this problem with
iGoogle OAuth support, and handled it by arranging for
gadgets.io.makeRequest to return an "oauthError" parameter of
"UNAUTHENTICATED" when an anonymous user attempts to use OAuth.

[Bess Ho]

+1

Add my vote on supporting guest and anonymous users.

 

We are currently using -1 to handle guest and anonymous users. Let's make it a standard to make our lives easier for cross-platform migration.

[Scott Seely]

+1

We should support anonymous users.

 

FWIW, MySpace returns a -1 to handle anonymous users.

 

 

<br

[Scott Seely]

Kaan:

Can you write a proposed patch to the spec? Feel free post to http://wiki.opensocial.org/index.php?title=Anonymous_Viewer.

 

Thanks!

[Kaan Bingol]

I have posted on the wiki page as a proposal.

http://wiki.opensocial.org/index.php?title=Anonymous_Viewer

Among alternatives I have proposed to use -1 as the ID of anonymous
user, since this was already used in containers and somehow easier to
handle than Null. Hope this does not interfere with ID numbering of
some containers. Comments are welcome on the proposal of course.

Regards,
Kaan

On Nov 19, 12:39 am, Scott Seely <sSe...@myspace.com> wrote:
> Kaan:
>

> Can you write a proposed patch to the spec? Feel free post tohttp://wiki.opensocial.org/index.php?title=Anonymous_Viewer.

[Scott Seely]

+1

[Arne Roomann-Kurrik]

+1, thanks for your work Kaan!

~Arne

--
OpenSocial IRC - irc://irc.freenode.net/opensocial

[Chad Russell]

+1

[Lane LiaBraaten]

+1

[Maxwell Newbould]

+1

---

From: opensocial-an...@googlegroups.com
To: opensocial-an...@googlegroups.com
Sent: Thu Nov 20 09:11:04 2008

[Paul Lindner]

yes, please +1

Of course I expect app developers will now need a new API --  opensocial.requestLoginNowPrettyPlease()

:-)

Paul Lindner

plin...@hi5.com

[Scott Seely]

I think this issue is accepted. J

[Scott Seely]

Updated http://sites.google.com/site/opensocialdraft/Home/opensocial-javascript-api-reference/person,
setting the values for opensocial.Person.Field.ID,
opensocial.Person.Field.NICKNAME, opensocial.Person.Field.NAME

Updated http://sites.google.com/site/opensocialdraft/Home/restful-protocol-specification,
section 6.1 (People), adding final paragraph to the spec.

Evan-- please reflect these changes in the simplified JS API docs.

On Nov 20, 9:50 am, Scott Seely <sSe...@myspace.com> wrote:
> I think this issue is accepted. J
>
> From: opensocial-an...@googlegroups.com [mailto:opensocial-an...@googlegroups.com] On Behalf Of Paul Lindner
> Sent: Thursday, November 20, 2008 9:40 AM
> To: opensocial-an...@googlegroups.com
> Subject: [opensocial-and-gadgets-spec] Re: Setting the VIEWER ID when user is not logged in
>
> yes, please +1
>
> Of course I expect app developers will now need a new API --  opensocial.requestLoginNowPrettyPlease()
>
> :-)
>
> On Nov 20, 2008, at 9:36 AM, Maxwell Newbould wrote:
>
> +1
>
> ________________________________
>
> From: opensocial-an...@googlegroups.com
> To: opensocial-an...@googlegroups.com
> Sent: Thu Nov 20 09:11:04 2008
> Subject: [opensocial-and-gadgets-spec] Re: Setting the VIEWER ID when user is not logged in
>
> +1
>

> plind...@hi5.com

[Bess Ho]

+1

--
Bess Ho

[Scott Seely]

We appreciate the enthusiasm!

 

At this point, we aren’t voting on whether or not the edits are correct. This means that your +1 is only being interpreted as a vote of confidence in the edit (which is nice to see!). These messages are intended to let people know that the proposal has been placed into the spec. This allows the group to perform a review soon after edits are made.

 

The vote to officially accept the edits happens when the spec editor (me in v0.9) officially announces that the draft is ready for review. This announcement will happen on December 9 if we keep making progress.